[{"data":1,"prerenderedAt":2592},["ShallowReactive",2],{"site-chrome-en":3,"resource-trust-your-ai-scorecard":902},{"header":4,"footer":78,"siteSettings":342,"fields":518},{"id":5,"logoHref":6,"nav":7,"signIn":65,"secondaryCta":68,"primaryCta":71,"menuLabel":74,"updatedAt":75,"createdAt":76,"globalType":77},1,"\u002F",[8,13,18,23,60],{"id":9,"label":10,"href":11,"children":12},"6ab3da06c46d8dcdda81f5eb","Why Trackian?","\u002F#features",[],{"id":14,"label":15,"href":16,"children":17},"6ab3da06c46d8dcdda81f5ec","Pricing","\u002Fpricing\u002F",[],{"id":19,"label":20,"href":21,"children":22},"6ab3da06c46d8dcdda81f5ed","FAQ","\u002Ffaq\u002F",[],{"id":24,"label":25,"href":26,"children":27},"6ab3da06c46d8dcdda81f5f2","Resources","#",[28,32,36,40,44,48,52,56],{"id":29,"label":30,"href":31},"6acf951e9e446caf0f92ef43","Free tools and guides","\u002Fresources\u002F",{"id":33,"label":34,"href":35},"6ad0d078c5d300c7eb1472f3","AI demo","\u002Fevents-demo\u002F",{"id":37,"label":38,"href":39},"6ac66c8c24b35e889d69ee3c","Platform Status Monitor","https:\u002F\u002Fstatus.trackian.com\u002F",{"id":41,"label":42,"href":43},"6ac66c8c24b35e889d69ee3d","Trackian Help Center","https:\u002F\u002Fhelp.trackian.com\u002Fhc\u002Fhelp-center",{"id":45,"label":46,"href":47},"6ab3da06c46d8dcdda81f5ee","Trackian Insights & News","\u002Finsights\u002F",{"id":49,"label":50,"href":51},"6ab3da06c46d8dcdda81f5ef","Client Case Studies","\u002Fmarketing-intelligence-client-case-studies\u002F",{"id":53,"label":54,"href":55},"6ab3da06c46d8dcdda81f5f0","About us","\u002Fabout\u002F",{"id":57,"label":58,"href":59},"6ab3da06c46d8dcdda81f5f1","Careers","\u002Fcareers\u002F",{"id":61,"label":62,"href":63,"children":64},"6ab3da06c46d8dcdda81f5f3","Contacts","\u002Fcontacts\u002F",[],{"label":66,"href":67},"Sign in","https:\u002F\u002Ftrackian.app\u002Flogin",{"label":69,"href":70},"Book a Demo","\u002Fbook-demo\u002F",{"label":72,"href":73},"Start for FREE","https:\u002F\u002Ftrackian.app\u002Fregister","Open menu","2026-10-15T13:09:12.962Z","2026-09-23T13:54:14.420Z","header",{"id":5,"about":79,"columns":80,"newsletter":251,"copyright":312,"legalLinks":313,"updatedAt":339,"createdAt":340,"globalType":341},"At Trackian, we unify and analyze marketing data, so businesses can navigate complexity with clarity.",[81,132,186],{"id":82,"groups":83},"6ab3da06c46d8dcdda81f5fe",[84,108],{"id":85,"title":86,"links":87},"6ab3da06c46d8dcdda81f5f8","Use Cases",[88,93,98,103],{"id":89,"link":90},"6ab3da06c46d8dcdda81f5f4",{"label":91,"href":92},"E-commerce Growth","\u002Fuse-cases\u002Fecommerce-growth\u002F",{"id":94,"link":95},"6ab3da06c46d8dcdda81f5f5",{"label":96,"href":97},"Marketing Agency","\u002Fuse-cases\u002Fmarketing-agencies\u002F",{"id":99,"link":100},"6ab3da06c46d8dcdda81f5f6",{"label":101,"href":102},"Brands Lead-gen & B2B","\u002Fuse-cases\u002Flead-gen-b2b\u002F",{"id":104,"link":105},"6ab3da06c46d8dcdda81f5f7",{"label":106,"href":107},"Enterprise Expansion","\u002Fuse-cases\u002Fenterprise-expansion\u002F",{"id":109,"title":110,"links":111},"6ab3da06c46d8dcdda81f5fd","For Roles like",[112,117,122,127],{"id":113,"link":114},"6ab3da06c46d8dcdda81f5f9",{"label":115,"href":116},"Brand Founders","\u002Froles\u002Fbrand-founders\u002F",{"id":118,"link":119},"6ab3da06c46d8dcdda81f5fa",{"label":120,"href":121},"Marketing Leaders","\u002Froles\u002Fmarketing-leaders\u002F",{"id":123,"link":124},"6ab3da06c46d8dcdda81f5fb",{"label":125,"href":126},"Performance Experts","\u002Froles\u002Fperformance-experts\u002F",{"id":128,"link":129},"6ab3da06c46d8dcdda81f5fc",{"label":130,"href":131},"Marketing Data Analysts","\u002Froles\u002Fmarketing-data-analysts\u002F",{"id":133,"groups":134},"6ab3da06c46d8dcdda81f60a",[135],{"id":136,"title":137,"links":138},"6ab3da06c46d8dcdda81f609","Trackian Features",[139,144,149,154,159,164,169,174,178,182],{"id":140,"link":141},"6ab3da06c46d8dcdda81f5ff",{"label":142,"href":143},"Data Integration Sources","\u002Fintegrations\u002F",{"id":145,"link":146},"6ab3da06c46d8dcdda81f600",{"label":147,"href":148},"Landing Page Health Monitor","\u002Ffeatures\u002Flanding-page-health-monitor\u002F",{"id":150,"link":151},"6ab3da06c46d8dcdda81f601",{"label":152,"href":153},"Data Anomalies Detection","\u002Ffeatures\u002Fanomaly-detection\u002F",{"id":155,"link":156},"6ab3da06c46d8dcdda81f602",{"label":157,"href":158},"Business Opportunity Finder","\u002Ffeatures\u002Fopportunity-finder\u002F",{"id":160,"link":161},"6ab3da06c46d8dcdda81f603",{"label":162,"href":163},"Marketing Daily Brief","\u002Ffeatures\u002Fmarketing-daily-brief\u002F",{"id":165,"link":166},"6ab3da06c46d8dcdda81f604",{"label":167,"href":168},"AI MCP tools & API access","\u002Ffeatures\u002Fai-mcp-api\u002F",{"id":170,"link":171},"6ab3da06c46d8dcdda81f605",{"label":172,"href":173},"Additional Services","\u002Fservices\u002F",{"id":175,"link":176},"6ab3da06c46d8dcdda81f606",{"label":177,"href":16},"Trackian Pricing",{"id":179,"link":180},"6ab3da06c46d8dcdda81f607",{"label":181,"href":21},"Frequently Asked Questions",{"id":183,"link":184},"6ab3da06c46d8dcdda81f608",{"label":185,"href":51},"Client Testimonials",{"id":187,"groups":188},"6ab3da06c46d8dcdda81f615",[189,208,232],{"id":190,"title":191,"links":192},"6ab3da06c46d8dcdda81f60f","Help & Support",[193,196,200,204],{"id":194,"link":195},"6ab3da06c46d8dcdda81f60b",{"label":38,"href":39},{"id":197,"link":198},"6ab3da06c46d8dcdda81f60c",{"label":199,"href":43},"Technical Support",{"id":201,"link":202},"6ab3da06c46d8dcdda81f60d",{"label":203,"href":43},"Knowledge Base",{"id":205,"link":206},"6ab3da06c46d8dcdda81f60e",{"label":207,"href":63},"Contact Us",{"id":209,"title":210,"links":211},"6ab3da06c46d8dcdda81f614","Company Resources",[212,215,219,223,227],{"id":213,"link":214},"6acf951e9e446caf0f92ef44",{"label":30,"href":31},{"id":216,"link":217},"6ab3da06c46d8dcdda81f610",{"label":218,"href":55},"About Trackian",{"id":220,"link":221},"6ab3da06c46d8dcdda81f611",{"label":222,"href":47},"Trackian Blog & News",{"id":224,"link":225},"6ab3da06c46d8dcdda81f612",{"label":226,"href":59},"Your Career in Trackian",{"id":228,"link":229},"6ab3da06c46d8dcdda81f613",{"label":230,"href":231},"Affiliate & Reseller Program","\u002Faffiliate-reseller-program\u002F",{"id":233,"title":234,"links":235},"6ab3fa0f456178c9ec754921","Follow us",[236,241,246],{"id":237,"link":238},"6ab3fa0f456178c9ec75491e",{"label":239,"href":240},"LinkedIn","https:\u002F\u002Fwww.linkedin.com\u002Fcompany\u002Ftrackiancom\u002F",{"id":242,"link":243},"6ab3fa0f456178c9ec75491f",{"label":244,"href":245},"Facebook","https:\u002F\u002Fwww.facebook.com\u002Ftrackiancom",{"id":247,"link":248},"6ab3fa0f456178c9ec754920",{"label":249,"href":250},"X","https:\u002F\u002Ftwitter.com\u002Ftrackiancom",{"formId":252,"title":254,"placeholder":305,"submitLabel":306,"submittingLabel":307,"successMessage":308,"invalidEmailMessage":309,"failureMessage":310,"privacyNote":311},{"id":253,"title":254,"fields":255,"submitButtonLabel":268,"confirmationType":269,"confirmationMessage":270,"redirect":283,"emails":284,"kind":302,"updatedAt":303,"createdAt":304},2,"Newsletter",[256,262],{"id":257,"name":258,"label":259,"width":260,"required":261,"blockName":260,"blockType":258},"6ab3da06c46d8dcdda81f5e1","email","Email",null,true,{"id":263,"name":264,"label":265,"width":260,"defaultValue":260,"required":266,"blockName":260,"blockType":267},"6ab3da06c46d8dcdda81f5e2","website","Website",false,"text","Subscribe","message",{"root":271},{"type":272,"format":273,"indent":274,"version":5,"children":275,"direction":282},"root","",0,[276],{"type":277,"format":273,"indent":274,"version":5,"children":278,"direction":282},"paragraph",[279],{"type":267,"format":274,"style":273,"mode":280,"detail":274,"version":5,"text":281},"normal","Subscribed — welcome aboard.","ltr",{"url":260},[285],{"id":286,"emailTo":287,"cc":260,"bcc":260,"replyTo":288,"emailFrom":289,"subject":290,"message":291},"6ab3da06c46d8dcdda81f5e3","marketing@trackian.com","{{email}}","newsletter@trackian.com","New newsletter signup",{"root":292},{"type":272,"format":273,"indent":274,"version":5,"children":293,"direction":282},[294,298],{"type":277,"format":273,"indent":274,"version":5,"children":295,"direction":282},[296],{"type":267,"format":274,"style":273,"mode":280,"detail":274,"version":5,"text":297},"New newsletter signup: {{email}}",{"type":277,"format":273,"indent":274,"version":5,"children":299,"direction":282},[300],{"type":267,"format":274,"style":273,"mode":280,"detail":274,"version":5,"text":301},"{{*:table}}","newsletter","2026-09-23T18:27:04.130Z","2026-09-23T13:54:14.300Z","Enter your e-mail address...","Subscribe Now","Subscribing…","You're subscribed. Check your inbox to confirm.","Enter a valid email address, e.g. name@company.com.","Couldn't subscribe right now. Please try again in a minute.","Your email is secure with us and we won't send you any spam.","Trackian® Copyright 2026 © All Rights Reserved",[314,319,324,329,334],{"id":315,"link":316},"6ab3da06c46d8dcdda81f616",{"label":317,"href":318},"Sitemap","\u002Fwebsite-sitemap\u002F",{"id":320,"link":321},"6ab3da06c46d8dcdda81f617",{"label":322,"href":323},"Terms & Conditions","\u002Fprivacy-center\u002Fterms-and-conditions\u002F",{"id":325,"link":326},"6ab3da06c46d8dcdda81f618",{"label":327,"href":328},"Privacy Policy","\u002Fprivacy-center\u002Fprivacy-policy\u002F",{"id":330,"link":331},"6ab3da06c46d8dcdda81f619",{"label":332,"href":333},"Cookie Policy","\u002Fprivacy-center\u002Fcookie-policy\u002F",{"id":335,"link":336},"6ab3da06c46d8dcdda81f61a",{"label":337,"href":338},"Accessibility Policy","\u002Faccessibility-statement\u002F","2026-10-14T14:43:42.900Z","2026-09-23T13:54:14.470Z","footer",{"id":5,"gtmContainerId":343,"recaptchaSiteKey":344,"consent":345,"defaultMeta":350,"formMessages":353,"ui":360,"integrations":407,"siteLocales":482,"updatedAt":515,"createdAt":516,"globalType":517},"GTM-NQT76P65","6Lc_ks0rAAAAAFrQGfMbWZRrXIUBhP9GxLqcK7Kx",{"message":346,"policyLink":347,"acceptLabel":348,"essentialLabel":349},"We use cookies to measure how the site is used. You can accept analytics cookies or continue with essential cookies only.",{"label":332,"href":333},"Accept all","Essential only",{"title":351,"description":352,"image":260},"Trackian: Marketing Intelligence","Trackian turns your marketing and business data into ranked daily actions.",{"requiredField":354,"invalidEmail":309,"requestFailed":355,"submitting":356,"success":357,"recaptchaNotice":358,"recaptchaPrivacy":327,"recaptchaTerms":359},"This field is required.","Couldn't send right now. Please try again in a minute.","Sending…","Thanks, we'll be in touch shortly.","This site is protected by reCAPTCHA and the Google {privacy} and {terms} apply.","Terms of Service",{"backToInsights":361,"backToCareers":362,"onThisPage":363,"tableOfContentsLabel":364,"searchTriggerLabel":365,"searchPlaceholder":366,"searchLabel":367,"searchTitle":368,"searchNoResults":369,"searchNoResultsHint":370,"searchEmptyHint":371,"searchUnavailable":372,"searchGroupPages":373,"searchGroupInsights":374,"searchGroupCaseStudies":375,"searchGroupIntegrations":376,"searchGroupHelp":377,"searchRecentLabel":378,"searchSuggestedLabel":379,"searchClearRecent":380,"searchClearLabel":381,"searchKeyboardHint":382,"readMore":383,"readCaseStudy":384,"viewRole":385,"noOpenRoles":386,"readingTime":387,"relatedInsights":388,"share":389,"copyLink":390,"copied":391,"caseStudyChallenge":392,"caseStudySolution":393,"caseStudyResults":394,"applyNow":395,"applicationsClosed":396,"closedBadge":397,"jobDepartmentLabel":398,"jobLocationLabel":399,"jobTypeLabel":400,"jobSalaryLabel":401,"jobHiringManagerLabel":402,"jobDeadlineLabel":403,"honeypotLabel":404,"formSubmitting":356,"cookieConsentLabel":405,"localeNoticeMessage":406},"← Back to Insights","← Back to Careers","On this page","Table of contents","Search Trackian.com…","Search pages, insights, case studies…","Search","Search Trackian","No results for \"{query}\"","Try a different search term.","Start typing to search pages, insights and case studies.","Search is unavailable right now. Please try again later.","Pages","Insights","Case Studies","Integrations","Help center","Recent","Suggested","Clear recent","Clear search","to search","Read More","Read case study","View role","No open roles right now. Check back soon or follow us for updates.","{minutes} min read","Related Insights","Share","Copy link","Copied!","Challenge","Solution","Results","Apply now","Applications closed","Closed","Department","Location","Type","Salary","Hiring Manager","Deadline","Leave this field empty","Cookie consent","This document is available in English only.",{"railAria":408,"collectionsHeading":409,"allApps":410,"collectionEssentials":411,"collectionNew":412,"collectionSoon":413,"categoriesHeading":414,"categoryAll":415,"filtersHeading":416,"facetAvailability":417,"facetAccess":418,"availabilityAvailable":419,"availabilityAlpha":420,"availabilityBeta":421,"availabilityComingSoon":413,"accessRead":422,"accessDual":423,"resetFilters":424,"searchPlaceholder":425,"searchLabel":426,"sortLabel":427,"sortRecommended":428,"sortName":429,"sortNewest":430,"viewing":431,"showMore":432,"filtersButton":416,"filtersHint":433,"done":434,"chipsFilters":416,"chipsRemove":435,"chipsResetSort":436,"chipsClearAll":437,"dismiss":438,"emptyTitle":439,"emptyBody":440,"clearFilters":441,"connect":442,"comingSoon":413,"newTag":412,"agentOnly":443,"backToStore":444,"overview":445,"whatIs":446,"whatItDoes":447,"relatedApps":448,"factType":400,"factSetup":449,"factAgent":450,"factCategory":451,"factAdded":452,"typeInternal":453,"typeExternal":454,"setupOauth":455,"setupKey":456,"setupManual":457,"builtByTrackian":458,"builtByVendor":459,"powerBlended":460,"powerDashboards":461,"powerSentinel":462,"powerMcp":463,"powerMessaging":464,"powerExternal1":465,"powerExternal2":466,"powerExternal3":467,"screenshotOpen":468,"screenshotPrev":469,"screenshotNext":470,"screenshotGoTo":471,"screenshotAlt":472,"screenshotCounter":473,"ctaTitle":474,"ctaBody":475,"ctaBodyExternal":476,"ctaConnect":477,"ctaNewcomer":478,"ctaStartFree":479,"metaTitle":480,"metaDescriptionFallback":481},"Browse apps","Collections","All apps","Essentials","New","Coming soon","Categories","All","Filters","Availability","AI agent access","Available","Alpha","Beta","Read only","Dual layer","Reset filters","Search apps…","Search apps","Sort","Recommended","A to Z","Newest","Viewing {shown} of {total} apps","Show {n} more","Changes apply as you tick.","Done","Remove filter","Reset sort","Clear all","Dismiss","No platforms match your filter","Try clearing the search or picking a different filter.","Clear filters","Connect","Agent only","All Platforms","Overview","What is {name}?","What does it do?","Related apps","Setup","AI agent","Category","Added","Full analysis","Agent only (Beta)","Log in with the vendor","API key","Plugin, module or store keys","Built by Trackian","Built by {name} · connected through Trackian","Daily metrics blended with GA4 and your store on one timeline","Dashboards and KPI targets for every metric","Sentinel patterns and nightly findings","Read access for your AI agent through Trackian's MCP","Sentinel alerts and reports delivered to your channel","Every endpoint the {name} API exposes, read only","Fresh data on every call, nothing cached or modelled","Ask your AI agent in plain language, it picks the endpoint","Open screenshot {n} full size","Previous screenshot","Next screenshot","Go to screenshot {n}","{name} screenshot {n}","{n} of {total}","Connect {name} to Trackian","Trackian lines its data up with the rest of your stack and watches it every night.","Your AI assistant reads it through Trackian's MCP, next to everything else you connect.","Connect {name}","New to Trackian?","Start for free","{name} integration | Trackian","Connect {name} to Trackian and see its data next to the rest of your marketing stack.",[483,487,491,495,499,503,507,511],{"id":484,"code":485,"label":486},"6ab3da06c46d8dcdda81f61b","en","English",{"id":488,"code":489,"label":490},"6ab3da06c46d8dcdda81f61c","bg","Български",{"id":492,"code":493,"label":494},"6ab3da06c46d8dcdda81f61d","de","Deutsch",{"id":496,"code":497,"label":498},"6ab3da06c46d8dcdda81f61e","el","Ελληνικά",{"id":500,"code":501,"label":502},"6ab3da06c46d8dcdda81f61f","es","Español",{"id":504,"code":505,"label":506},"6ab3da06c46d8dcdda81f620","it","Italiano",{"id":508,"code":509,"label":510},"6ab3da06c46d8dcdda81f621","pl","Polski",{"id":512,"code":513,"label":514},"6ab3da06c46d8dcdda81f622","ro","Română","2026-10-14T10:59:48.664Z","2026-09-23T13:54:14.554Z","siteSettings",{"collections":519,"globals":729},{"pages":520,"posts":657,"caseStudies":665,"jobs":677,"testimonials":684,"faqs":690,"plans":697,"redirects":715,"search":720,"payload-locked-documents":725},{"text":521,"richText":651,"relations":653},[522,523,524,525,526,527,528,529,530,531,532,533,534,535,536,537,538,539,540,541,542,543,544,545,546,547,548,549,550,551,552,553,554,555,556,557,558,559,560,561,562,563,564,565,566,567,568,569,570,571,572,573,574,575,576,577,578,579,580,581,582,583,584,585,586,587,588,589,590,591,592,593,594,595,596,597,598,599,600,601,602,603,604,605,606,607,608,609,610,611,612,613,614,615,616,617,618,619,620,621,622,623,624,625,626,627,628,629,630,631,632,633,634,635,636,637,638,639,640,641,642,643,644,645,646,647,648,649,650],"title","layout.hero.badge","layout.hero.title","layout.hero.lead","layout.hero.primaryCta.label","layout.hero.secondaryCta.label","layout.hero.alerts.*.title","layout.hero.alerts.*.ago","layout.hero.alerts.*.body","layout.logoMarquee.label","layout.audienceSplit.title","layout.audienceSplit.lead","layout.audienceSplit.cards.*.eyebrow","layout.audienceSplit.cards.*.body","layout.audienceSplit.cards.*.link.label","layout.featureBento.eyebrow","layout.featureBento.title","layout.featureBento.lead","layout.featureBento.tiles.*.title","layout.featureBento.tiles.*.body","layout.integrationHub.title","layout.integrationHub.cta.label","layout.tabbedFeature.title","layout.tabbedFeature.lead","layout.tabbedFeature.tabs.*.title","layout.tabbedFeature.tabs.*.body","layout.tabbedFeature.tabs.*.caption","layout.tabbedFeature.cta.label","layout.statFeature.title","layout.statFeature.body","layout.statFeature.chips.*.text","layout.serviceGrid.eyebrow","layout.serviceGrid.title","layout.serviceGrid.lead","layout.serviceGrid.note","layout.serviceGrid.cta.label","layout.serviceGrid.services.*.title","layout.serviceGrid.services.*.body","layout.testimonialGrid.title","layout.ctaBand.title","layout.ctaBand.lead","layout.ctaBand.primaryCta.label","layout.ctaBand.secondaryCta.label","layout.ctaBand.cards.*.title","layout.ctaBand.cards.*.body","layout.ctaBand.steps.*.title","layout.ctaBand.steps.*.body","layout.ctaBand.steps.*.links.*.label","layout.ctaBand.steps.*.note","layout.pricingHero.title","layout.pricingHero.lead","layout.pricingHero.billingLabels.monthly","layout.pricingHero.billingLabels.annual","layout.pricingHero.billingLabels.save","layout.pricingHero.footnotes.vat","layout.pricingHero.footnotes.payments","layout.pricingHero.segments.*.tabLabel","layout.pricingHero.segments.*.title","layout.pricingHero.segments.*.body","layout.pricingHero.segments.*.comparison.eyebrow","layout.pricingHero.segments.*.comparison.title","layout.pricingHero.segments.*.comparison.expandLabel","layout.pricingHero.segments.*.comparison.collapseLabel","layout.pricingHero.segments.*.comparison.srLabels.included","layout.pricingHero.segments.*.comparison.srLabels.notIncluded","layout.pricingHero.segments.*.comparison.columns.*.note","layout.pricingHero.segments.*.comparison.rows.*.label","layout.pricingHero.segments.*.comparison.rows.*.hint","layout.pricingHero.segments.*.comparison.rows.*.cells.*.value","layout.faq.title","layout.faq.help","layout.form.intro","layout.postList.title","layout.postList.lead","layout.caseStudyList.title","layout.caseStudyList.lead","layout.jobList.title","layout.jobList.lead","layout.notFound.title","layout.notFound.body","layout.notFound.primaryCta.label","layout.notFound.links.*.link.label","layout.notFound.searchPlaceholder","layout.htmlSitemap.title","layout.htmlSitemap.lead","layout.htmlSitemap.groupLabels.pages","layout.htmlSitemap.groupLabels.insights","layout.htmlSitemap.groupLabels.caseStudies","layout.htmlSitemap.groupLabels.careers","layout.htmlSitemap.groupLabels.legal","layout.splitShowcase.title","layout.splitShowcase.lead","layout.splitShowcase.rows.*.title","layout.splitShowcase.rows.*.body","layout.splitShowcase.rows.*.chips.*.text","layout.screenshotStage.title","layout.screenshotStage.lead","layout.screenshotStage.hotspots.*.label","layout.screenshotStage.hotspots.*.body","layout.steps.title","layout.steps.lead","layout.steps.steps.*.title","layout.steps.steps.*.body","layout.bigQuote.quote","layout.bigQuote.role","layout.metricsBand.title","layout.metricsBand.lead","layout.metricsBand.metrics.*.label","layout.comparisonSplit.title","layout.comparisonSplit.lead","layout.comparisonSplit.withoutTrackian.label","layout.comparisonSplit.withoutTrackian.items.*.text","layout.comparisonSplit.withTrackian.label","layout.comparisonSplit.withTrackian.items.*.text","layout.logoWall.title","layout.heroSplit.badge","layout.heroSplit.title","layout.heroSplit.lead","layout.heroSplit.primaryCta.label","layout.heroSplit.secondaryCta.label","layout.heroSplit.chips.*.text","layout.heroSplit.note","layout.connectorStore.title","layout.connectorStore.lead","layout.resourceLibrary.heading","layout.resourceLibrary.featuredHeading","layout.resourceLibrary.intro","meta.title","meta.description",[652],"layout.richText.content",{"layout.testimonialGrid.testimonials":654,"layout.pricingHero.segments.*.plans":655,"layout.pricingHero.segments.*.comparison.columns.*.plan":655,"layout.faq.items":656},"testimonials","plans","faqs",{"text":658,"richText":662,"relations":664},[522,659,660,661,649,650],"excerpt","category","author.title",[663],"content",{},{"text":666,"richText":672,"relations":676},[667,668,669,670,671,649,650],"clientTitle","resultHeadline","stats.*.label","quote.text","quote.title",[673,674,675],"challenge","solution","results",{},{"text":678,"richText":682,"relations":683},[522,679,680,681,649,650],"location","type","department",[663],{},{"text":685,"richText":688,"relations":689},[686,687],"quote","ratingLabel",[],{},{"text":691,"richText":695,"relations":696},[692,693,694],"question","answer","tags.*.tag",[],{},{"text":698,"richText":713,"relations":714},[699,700,701,702,703,704,705,706,707,708,709,710,711,712],"name","featuredLabel","customLabel","customNote.monthly","customNote.annual","billedMonthlyNote","billedAnnuallyNote","perMonthLabel","idealForLabel","idealFor","highlights.*.text","guarantee","cta.monthly.label","cta.annual.label",[],{},{"text":716,"richText":717,"relations":718},[],[],{"to.reference":719},"posts",{"text":721,"richText":722,"relations":723},[],[],{"doc":724},"caseStudies",{"text":726,"richText":727,"relations":728},[],[],{"document":655},{"header":730,"footer":740,"siteSettings":757},{"text":731,"richText":738,"relations":739},[732,733,734,735,736,737],"nav.*.label","nav.*.children.*.label","signIn.label","secondaryCta.label","primaryCta.label","menuLabel",[],{},{"text":741,"richText":755,"relations":756},[742,743,744,745,746,747,748,749,750,751,752,753,754],"about","columns.*.groups.*.title","columns.*.groups.*.links.*.link.label","newsletter.title","newsletter.placeholder","newsletter.submitLabel","newsletter.submittingLabel","newsletter.successMessage","newsletter.invalidEmailMessage","newsletter.failureMessage","newsletter.privacyNote","copyright","legalLinks.*.link.label",[],{},{"text":758,"richText":900,"relations":901},[759,760,761,762,763,764,765,766,767,768,769,770,771,772,773,774,775,776,777,778,779,780,781,782,783,784,785,786,787,788,789,790,791,792,793,794,795,796,797,798,799,800,801,802,803,804,805,806,807,808,809,810,811,812,813,814,815,816,817,818,819,820,821,822,823,824,825,826,827,828,829,830,831,832,833,834,835,836,837,838,839,840,841,842,843,844,845,846,847,848,849,850,851,852,853,854,855,856,857,858,859,860,861,862,863,864,865,866,867,868,869,870,871,872,873,874,875,876,877,878,879,880,881,882,883,884,885,886,887,888,889,890,891,892,893,894,895,896,897,898,899],"consent.message","consent.policyLink.label","consent.acceptLabel","consent.essentialLabel","defaultMeta.title","defaultMeta.description","formMessages.requiredField","formMessages.invalidEmail","formMessages.requestFailed","formMessages.submitting","formMessages.success","formMessages.recaptchaNotice","formMessages.recaptchaPrivacy","formMessages.recaptchaTerms","ui.backToInsights","ui.backToCareers","ui.onThisPage","ui.tableOfContentsLabel","ui.searchTriggerLabel","ui.searchPlaceholder","ui.searchLabel","ui.searchTitle","ui.searchNoResults","ui.searchNoResultsHint","ui.searchEmptyHint","ui.searchUnavailable","ui.searchGroupPages","ui.searchGroupInsights","ui.searchGroupCaseStudies","ui.searchGroupIntegrations","ui.searchGroupHelp","ui.searchRecentLabel","ui.searchSuggestedLabel","ui.searchClearRecent","ui.searchClearLabel","ui.searchKeyboardHint","ui.readMore","ui.readCaseStudy","ui.viewRole","ui.noOpenRoles","ui.readingTime","ui.relatedInsights","ui.share","ui.copyLink","ui.copied","ui.caseStudyChallenge","ui.caseStudySolution","ui.caseStudyResults","ui.applyNow","ui.applicationsClosed","ui.closedBadge","ui.jobDepartmentLabel","ui.jobLocationLabel","ui.jobTypeLabel","ui.jobSalaryLabel","ui.jobHiringManagerLabel","ui.jobDeadlineLabel","ui.honeypotLabel","ui.formSubmitting","ui.cookieConsentLabel","ui.localeNoticeMessage","integrations.railAria","integrations.collectionsHeading","integrations.allApps","integrations.collectionEssentials","integrations.collectionNew","integrations.collectionSoon","integrations.categoriesHeading","integrations.categoryAll","integrations.filtersHeading","integrations.facetAvailability","integrations.facetAccess","integrations.availabilityAvailable","integrations.availabilityAlpha","integrations.availabilityBeta","integrations.availabilityComingSoon","integrations.accessRead","integrations.accessDual","integrations.resetFilters","integrations.searchPlaceholder","integrations.searchLabel","integrations.sortLabel","integrations.sortRecommended","integrations.sortName","integrations.sortNewest","integrations.viewing","integrations.showMore","integrations.filtersButton","integrations.filtersHint","integrations.done","integrations.chipsFilters","integrations.chipsRemove","integrations.chipsResetSort","integrations.chipsClearAll","integrations.dismiss","integrations.emptyTitle","integrations.emptyBody","integrations.clearFilters","integrations.connect","integrations.comingSoon","integrations.newTag","integrations.agentOnly","integrations.backToStore","integrations.overview","integrations.whatIs","integrations.whatItDoes","integrations.relatedApps","integrations.factType","integrations.factSetup","integrations.factAgent","integrations.factCategory","integrations.factAdded","integrations.typeInternal","integrations.typeExternal","integrations.setupOauth","integrations.setupKey","integrations.setupManual","integrations.builtByTrackian","integrations.builtByVendor","integrations.powerBlended","integrations.powerDashboards","integrations.powerSentinel","integrations.powerMcp","integrations.powerMessaging","integrations.powerExternal1","integrations.powerExternal2","integrations.powerExternal3","integrations.screenshotOpen","integrations.screenshotPrev","integrations.screenshotNext","integrations.screenshotGoTo","integrations.screenshotAlt","integrations.screenshotCounter","integrations.ctaTitle","integrations.ctaBody","integrations.ctaBodyExternal","integrations.ctaConnect","integrations.ctaNewcomer","integrations.ctaStartFree","integrations.metaTitle","integrations.metaDescriptionFallback",[],{},{"resource":903,"doors":2591,"libraryTitle":25},{"path":904,"href":905,"itemNumber":906,"track":907,"typeLabel":908,"format":-1,"length":909,"effort":910,"hasSkill":266,"platforms":911,"cards":918,"slug":904,"edition":-1,"level":995,"parentPath":-1,"gate":996,"cta":997,"title":920,"shortTitle":920,"button":998,"hero":999,"heroArt":1003,"flatCover":-1,"whatYouGet":1008,"who":1010,"inside":1013,"faq":1015,"layout":1016,"formId":1017,"picker":-1,"sheetsCopyUrl":-1,"thankYou":1018,"seo":1021,"check":-1,"toolKind":1022,"tool":1023},"trust-your-ai-scorecard","\u002Fresources\u002Ftrust-your-ai-scorecard\u002F","20","both","Scorecard","About 3 minutes","quick",[912,913,914,915,916,917],"claude","chatgpt","gemini","ga4","gads","meta",[919],{"cardId":906,"track":907,"title":920,"sub":921,"description":922,"punchline":923,"sortOrder":924,"featured":266,"featuredOrder":-1,"coverArt":925,"cover":930,"stage":992,"platforms":994},"Can You Trust Your AI's Marketing Answers?","A 3-minute scorecard","3-minute scorecard on how far you can trust your AI's answers about ads, sales and leads.","no way to know",40,{"src":926,"alt":927,"width":928,"height":928,"srcset":929},"/media/20-cover-art.8de7650b.webp","Cover art for Can You Trust Your AI's Marketing Answers?",1200,"/media/20-cover-art.8de7650b-400x400.webp 400w, /media/20-cover-art.8de7650b-900x900.webp 900w, /media/20-cover-art.8de7650b.webp 1200w",{"x0":931,"y0":932,"x1":933,"y1":934,"chip_right":935,"chip_top":936,"note":937,"loop":942,"y2":947,"outline":948,"person":990,"tone":991},0.05859375,0.15380859375,0.92,0.32509765625,0.04296875,0.0400390625,[938,939,940,941],0.28515625,0.4765625,0.92578125,0.57421875,[943,944,945,946],0.509765625,0.486328125,0.736328125,0.607421875,0.3538085937500002,[949,952,954,956,958,960,962,964,966,968,970,972,974,976,978,980,982,984,986,988],[950,951],0.1538,0.93,[953,951],0.1638,[955,951],0.1738,[957,951],0.1838,[959,951],0.1938,[961,951],0.2038,[963,951],0.2138,[965,951],0.2238,[967,951],0.2338,[969,951],0.2438,[971,951],0.2538,[973,951],0.2638,[975,951],0.2738,[977,951],0.2838,[979,951],0.2938,[981,951],0.3038,[983,951],0.3138,[985,951],0.3238,[987,951],0.3338,[989,951],0.3438,"L","dark",{"kind":993},"scene",[912,913,914,915,916,917],"item","result-first","check","Check my AI setup",[1000],{"variant":1001,"sub":921,"hook":1002,"lead":922},"default","Confident Isn't Correct",{"src":1004,"alt":1005,"width":1006,"height":1006,"srcset":1007},"/media/20-hero-art.0ec1c17e.webp","Illustration for Can You Trust Your AI's Marketing Answers?",1600,"/media/20-hero-art.0ec1c17e-400x400.webp 400w, /media/20-hero-art.0ec1c17e-900x900.webp 900w, /media/20-hero-art.0ec1c17e-1600x1600.webp 1600w",{"shop":1009,"leads":1009},[],{"forList":1011,"notForList":1012,"photoShop":-1,"photoLeads":-1},[],[],{"intro":-1,"note":-1,"pages":1014},[],[],[],5,{"startHeading":-1,"whereToStartShop":-1,"whereToStartLeads":-1,"step2":1019,"related":1020},[],[],{"title":-1,"description":-1,"image":-1},"scored",{"meta":1024,"intro":1033,"track_selector":1038,"shared_qualification":1053,"scoring":1112,"routing":1138,"tracks":1230,"placeholders":2260,"result_labels":2264,"email_capture":2278,"pdf":2299,"sources":2373,"volume_band_labels":2483,"spend_band_labels":2488,"result_screen_order":2490,"visuals":2500,"forward_email":2559,"thank_you_routes":2571,"lead_scoring":2589},{"id":1025,"title":1026,"engine_version":1027,"content_version":1028,"language":485,"currency":1029,"reuse_note":1030,"url":1031,"edition_line":1032},"ai-trust-scorecard","Can you trust your AI's marketing answers?","1.0","2026-10-09","EUR","Built on the Tracking Readiness Scorecard engine (07-tracking-readiness-scorecard\u002Fscorecard.json), with the additions introduced in 14-numbers-lying-quiz\u002Fquiz.json (helper on scored questions, result_labels.why_label, text_link CTAs with placement, null countdown fields, top-level pdf object). New additive fields in this scorecard: shared qualification questions asked before the track's own (top-level shared_qualification, with per-track option lists), a top-level routing object that resolves guide links from the track, ai_tool and main_source answers, url_ref on links, and (launch update, 9 Oct 2026) ONE primary CTA per track (ctas.check_24h with own_numbers_line), tracks[].next_step_by_band (step 2 text links), tracks[].content_links, a link on the trust-test fixes, result_labels.step_one_line, forward_email, thank_you_routes, spend_band_labels \u002F volume_band_labels, track_selector.preselect_from_url and the timing question; and (owner answers, 9 Oct 2026 evening) show_when on a qualification question (ai_tool_most, the follow-up for people who use more than one assistant) and routing.main_ai. Same structure and field names otherwise.","https:\u002F\u002Ftrackian.com\u002Fresources\u002Ftrust-your-ai-scorecard","Made by Trackian, a marketing brain for your business. Edition: October 2026.",{"headline":1026,"subheadline":1034,"meta_line":1035,"start_button":1036,"small_print":1037},"For anyone who already asks ChatGPT, Claude, Gemini or Copilot about their ads, sales or leads. Answer a few plain-language questions and get a trust score out of 100, the 3 fixes that matter most, and the right next step for you.","About 3 minutes · Score shown instantly · No email needed to see it","Start the scorecard","AI assistants are genuinely useful for marketing questions. The risk is confident answers built on data that hasn't been joined or checked. Your score is based on your answers, not a scan of your accounts. Every scored question has an \"I don't know\" option, and that is a useful answer too.",{"id":1039,"text":1040,"helper":1041,"scored":266,"options":1042,"preselect_from_url":1049},"track","Do you sell online, or do you collect leads?","If you do both, pick the one most of your ad budget goes to. You can take the other track afterwards.",[1043,1046],{"id":1044,"text":1045,"sets_track":1044},"shop","I sell online (I run an online shop)",{"id":1047,"text":1048,"sets_track":1047},"leadgen","I collect leads (forms, calls or bookings)",{"param":1050,"values":1051,"note":1052},"v",{"shop":1044,"leads":1047},"An ad URL with ?v=shop or ?v=leads pre-selects the track and shows that track's intro. Visitors can still change it.",[1054,1074,1086],{"id":1055,"text":1056,"helper":1057,"scored":266,"ask_after":1039,"options":1058},"ai_tool","Which AI assistant do you use most for questions about your marketing or sales numbers?","This one isn't scored. It helps us point you to the right next step.",[1059,1061,1063,1065,1068,1071],{"id":913,"text":1060},"ChatGPT",{"id":912,"text":1062},"Claude",{"id":914,"text":1064},"Gemini",{"id":1066,"text":1067},"copilot","Microsoft Copilot",{"id":1069,"text":1070},"several","More than one of these",{"id":1072,"text":1073},"other","Another assistant",{"id":1075,"text":1076,"helper":1057,"scored":266,"ask_after":1055,"show_when":1077,"note":1078,"options":1079},"ai_tool_most","Which one do you use most?","ai_tool == 'several'","One-tap follow-up, shown only to people who answered 'More than one of these' (owner, 9 Oct 2026 evening, #56). When it is hidden, main_source follows ai_tool directly. Routing uses this answer in place of ai_tool (routing.main_ai).",[1080,1081,1082,1083,1084],{"id":912,"text":1062},{"id":913,"text":1060},{"id":914,"text":1064},{"id":1066,"text":1067},{"id":1072,"text":1085},"Other",{"id":1087,"text":1088,"helper":1057,"scored":266,"ask_after":1075,"options_by_track":1089},"main_source","What do you ask it about most?",{"shop":1090,"leadgen":1104},[1091,1093,1096,1099,1102],{"id":915,"text":1092},"Google Analytics",{"id":1094,"text":1095},"google_ads","Google Ads",{"id":1097,"text":1098},"meta_ads","Meta ads (Facebook and Instagram)",{"id":1100,"text":1101},"store","My shop's orders and sales",{"id":1069,"text":1103},"Several of these together",[1105,1106,1107,1108,1111],{"id":915,"text":1092},{"id":1094,"text":1095},{"id":1097,"text":1098},{"id":1109,"text":1110},"lead_list","My own list of leads (from forms, calls or bookings)",{"id":1069,"text":1103},{"total_max":1113,"method":1114,"fix_ranking":1115,"rank_by":1121,"top3_rules":1125,"keep_it_label":1137},100,"Sum the points of the chosen option for each of the 12 scored questions. Each track's maximums add up to exactly 100. Qualification questions (including ai_tool, ai_tool_most and main_source) are never scored. Higher score = the AI's answers rest on more complete, better-checked data.",[1116,1117,1118,1119,1120],"1. Collect every fix id in triggers_fixes of the chosen options. Remove duplicates.","2. Apply conditional_fixes (3 or more \"I don't know\" answers puts the \"write down how your AI gets its numbers\" fix in first place).","3. Sort the rest by fix.impact (high to low), then by points lost on the triggering question (max_points minus points scored, high to low), then by question order.","4. Show the top 3. If fewer than 3 were triggered, fill the remaining places from fallback_fixes in order, skipping any already shown, and label them \"Keep it this way\".","5. Band b1 only (owner, 11 Oct 2026): the trust-test fix (s-trust-test \u002F l-trust-test) always takes a top-3 slot. If it isn't already in the top 3, it takes third place and the fix that was third drops out. If s7 \u002F l7 was \"a\" (the fix wasn't triggered), it shows labelled \"Keep it this way\". See top3_rules.",[1122,1123,1124],"impact","points_lost","question_order",[1126],{"id":1127,"when":1128,"fix":1129,"rule":1132,"place_if_missing":1133,"detail":1134,"keep_it_label_when":1135,"source":1136},"b1-trust-test","band == 'b1'",{"shop":1130,"leadgen":1131},"s-trust-test","l-trust-test","always_in_top_3",3,"Applied after steps 1-4: if the fix isn't already in the top 3, it replaces the fix in third place.","s7 == 'a' (shop) \u002F l7 == 'a' (lead-gen): the fix wasn't triggered, so it shows with keep_it_label","Owner (Evgeni), 11 Oct 2026: _tools\u002FPLAN.md Q12, _tools\u002FCOPY-NOTES.md T3","Keep it this way",{"note":1139,"main_ai":1140,"trust_test":1144,"stack_guide":1215},"Resolve url_ref values from this table. Rules are checked in order; the first match wins. Rules match on main_ai, the assistant the person uses most. Gemini, Copilot and other assistants get the tool-neutral 30-second AI answer check (Defaults #19); ChatGPT users get the ChatGPT editions (connection guides 02-04, stack guides 18); Claude users get the Claude editions. People who use more than one assistant are routed by their answer to 'Which one do you use most?' (ai_tool_most), exactly as single-tool users of that assistant (owner, 9 Oct 2026 evening, #56).",{"definition":1141,"values":1142,"fallback":1143},"ai_tool_most if ai_tool == 'several', else ai_tool","claude, chatgpt, gemini, copilot, other","If ai_tool == 'several' and ai_tool_most is missing, main_ai = 'claude' (the earlier default for 'several').",{"label_by_rule":261,"rules":1145},[1146,1153,1160,1166,1172,1178,1185,1191,1197,1203,1209],{"when":1147,"guide":1148,"section":1149,"url":1150,"folder":1151,"id":1152},"main_ai in ('gemini', 'copilot', 'other')","9 Ways AI Gets Your Marketing Numbers Wrong","The 30-second AI answer check (works in any assistant)","https:\u002F\u002Ftrackian.com\u002Fresources\u002Fnine-ways-ai-gets-numbers-wrong\u002Fchecklist","30-nine-ways-ai-gets-numbers-wrong\u002Fchecklist-one-pager.md","30-checklist",{"when":1154,"guide":1155,"section":1156,"url":1157,"folder":1158,"id":1159},"main_source in ('store', 'lead_list', 'several') AND track == 'shop' AND main_ai == 'chatgpt'","Build Your Shop's AI Analyst in ChatGPT","Chapter 5: The 20-minute cross-source trust test","https:\u002F\u002Ftrackian.com\u002Fresources\u002Fstack-guides-chatgpt\u002Fshop#chapter-5","18-stack-guides-chatgpt\u002Fshop-ai-analyst-chatgpt.md","18-shop",{"when":1161,"guide":1162,"section":1156,"url":1163,"folder":1164,"id":1165},"main_source in ('store', 'lead_list', 'several') AND track == 'leadgen' AND main_ai == 'chatgpt'","Build Your Lead-Gen AI Analyst in ChatGPT","https:\u002F\u002Ftrackian.com\u002Fresources\u002Fstack-guides-chatgpt\u002Fleadgen#chapter-5","18-stack-guides-chatgpt\u002Fleadgen-ai-analyst-chatgpt.md","18-leadgen",{"when":1167,"guide":1168,"section":1156,"url":1169,"folder":1170,"id":1171},"main_source in ('store', 'lead_list', 'several') AND track == 'shop'","Build Your Shop's AI Analyst in Claude","https:\u002F\u002Ftrackian.com\u002Fresources\u002Fstack-guides-claude\u002Fshop#chapter-5","05-stack-guides-claude\u002Fshop-ai-analyst-claude.md","05-shop",{"when":1173,"guide":1174,"section":1156,"url":1175,"folder":1176,"id":1177},"main_source in ('store', 'lead_list', 'several') AND track == 'leadgen'","Build Your Lead-Gen AI Analyst in Claude","https:\u002F\u002Ftrackian.com\u002Fresources\u002Fstack-guides-claude\u002Fleadgen#chapter-5","05-stack-guides-claude\u002Fleadgen-ai-analyst-claude.md","05-leadgen",{"when":1179,"guide":1180,"section":1181,"url":1182,"folder":1183,"id":1184},"main_source == 'ga4' AND main_ai == 'chatgpt'","Connect GA4 to ChatGPT","Chapter 5: The 10-minute trust test","https:\u002F\u002Ftrackian.com\u002Fresources\u002Fga4-to-chatgpt#chapter-5","04-ga4-to-chatgpt\u002Fguide.md","04-ga4-chatgpt",{"when":1186,"guide":1187,"section":1181,"url":1188,"folder":1189,"id":1190},"main_source == 'ga4'","Connect GA4 to Claude","https:\u002F\u002Ftrackian.com\u002Fresources\u002Fga4-to-claude#chapter-5","01-ga4-to-claude\u002Fguide.md","01-ga4-claude",{"when":1192,"guide":1193,"section":1181,"url":1194,"folder":1195,"id":1196},"main_source == 'google_ads' AND main_ai == 'chatgpt'","Connect Google Ads to ChatGPT","https:\u002F\u002Ftrackian.com\u002Fresources\u002Fgoogle-ads-to-claude-chatgpt\u002Fchatgpt#chapter-5","03-google-ads-to-claude-chatgpt\u002Fguide-chatgpt.md","03-gads-chatgpt",{"when":1198,"guide":1199,"section":1181,"url":1200,"folder":1201,"id":1202},"main_source == 'google_ads'","Connect Google Ads to Claude","https:\u002F\u002Ftrackian.com\u002Fresources\u002Fgoogle-ads-to-claude-chatgpt\u002Fclaude#chapter-5","03-google-ads-to-claude-chatgpt\u002Fguide-claude.md","03-gads-claude",{"when":1204,"guide":1205,"section":1181,"url":1206,"folder":1207,"id":1208},"main_source == 'meta_ads' AND main_ai == 'chatgpt'","Connect Meta Ads to ChatGPT","https:\u002F\u002Ftrackian.com\u002Fresources\u002Fmeta-ads-to-claude-chatgpt\u002Fchatgpt#chapter-5","02-meta-ads-to-claude-chatgpt\u002Fguide-chatgpt.md","02-meta-chatgpt",{"when":1210,"guide":1211,"section":1181,"url":1212,"folder":1213,"id":1214},"main_source == 'meta_ads'","Connect Meta Ads to Claude","https:\u002F\u002Ftrackian.com\u002Fresources\u002Fmeta-ads-to-claude-chatgpt\u002Fclaude#chapter-5","02-meta-ads-to-claude-chatgpt\u002Fguide-claude.md","02-meta-claude",{"note":1216,"rules":1217},"Resolved by track and main_ai (Defaults #19; #56 for people who use more than one assistant): ChatGPT users get the ChatGPT edition (item 18); everyone else gets the Claude edition (item 05).",[1218,1221,1224,1227],{"when":1219,"guide":1155,"url":1220,"folder":1158,"id":1159},"track == 'shop' AND main_ai == 'chatgpt'","https:\u002F\u002Ftrackian.com\u002Fresources\u002Fstack-guides-chatgpt\u002Fshop",{"when":1222,"guide":1162,"url":1223,"folder":1164,"id":1165},"track == 'leadgen' AND main_ai == 'chatgpt'","https:\u002F\u002Ftrackian.com\u002Fresources\u002Fstack-guides-chatgpt\u002Fleadgen",{"when":1225,"guide":1168,"url":1226,"folder":1170,"id":1171},"track == 'shop'","https:\u002F\u002Ftrackian.com\u002Fresources\u002Fstack-guides-claude\u002Fshop",{"when":1228,"guide":1174,"url":1229,"folder":1176,"id":1177},"track == 'leadgen'","https:\u002F\u002Ftrackian.com\u002Fresources\u002Fstack-guides-claude\u002Fleadgen",[1231,1852],{"id":1044,"name":1232,"short_name":1233,"intro":1234,"qualification":1239,"questions":1302,"fixes":1570,"fallback_fixes":1781,"conditional_fixes":1782,"bands":1786,"diy_reality":1812,"pitch":1815,"ctas":1822,"content_links":1833,"next_step_by_band":1840},"Can you trust your AI's answers about your shop?","shop scorecard",{"headline":1235,"subheadline":1236,"meta_line":1035,"start_button":998,"seasonal_variants":1237,"preselect_param":1238},"Your AI just told you which channel drives your sales. Would you bet your budget on it?","A few plain-language questions about how your AI assistant gets, combines and checks your shop's numbers. Get a trust score out of 100, the 3 fixes that matter most and the right next step.",[],"?v=shop",[1240,1257,1274,1289],{"id":1241,"text":1242,"helper":1057,"scored":266,"ask_after":1243,"options":1244},"spend_band","Roughly how much do you spend on ads per month, across all platforms?","s4",[1245,1248,1251,1254],{"id":1246,"text":1247},"under_2k","Under €2k",{"id":1249,"text":1250},"2k_5k","€2k-5k",{"id":1252,"text":1253},"5k_15k","€5k-15k",{"id":1255,"text":1256},"15k_plus","€15k+",{"id":1258,"text":1259,"helper":1057,"scored":266,"ask_after":1260,"options":1261},"volume_band","Roughly how many orders do you get a month?","s8",[1262,1265,1268,1271],{"id":1263,"text":1264},"under_100","Under 100",{"id":1266,"text":1267},"100_499","100-499",{"id":1269,"text":1270},"500_1999","500-1,999",{"id":1272,"text":1273},"2000_plus","2,000+",{"id":1275,"text":1276,"helper":1057,"scored":266,"ask_after":1277,"options":1278},"role","Which best describes you?","s12",[1279,1282,1285,1288],{"id":1280,"text":1281},"owner","Owner or founder",{"id":1283,"text":1284},"marketing_lead","Marketing lead",{"id":1286,"text":1287},"agency","Agency (answering for a client)",{"id":1072,"text":1085},{"id":1290,"text":1291,"helper":1057,"scored":266,"ask_after":1275,"options":1292},"timing","When do you want this sorted?",[1293,1296,1299],{"id":1294,"text":1295},"this_month","This month",{"id":1297,"text":1298},"this_quarter","This quarter",{"id":1300,"text":1301},"exploring","Just exploring",[1303,1342,1367,1387,1408,1429,1450,1468,1487,1506,1525,1547],{"id":1304,"text":1305,"helper":1306,"why":1307,"max_points":1308,"options":1309,"sources":1338},"s1","How do your numbers usually get to the AI?","Pick the one you use most. A connector (ChatGPT calls them apps) is a direct link between the AI and a platform such as Google Analytics, Google Ads, Meta or Shopify.","The AI can only be as accurate as what it receives. Screenshots and hand-copied numbers can be misread or cut off, and the filters you applied are invisible to it. Exports are frozen the moment you made them and often hold only the rows that were on screen. Connectors fetch fresh data, but each one brings one platform with its own rules, and every extra one in the chat is another chance for an incomplete table. A source where your platforms are already joined and checked removes most of that work before the AI starts.",10,[1310,1315,1319,1324,1329,1333],{"id":1311,"text":1312,"points":253,"triggers_fixes":1313},"a","Screenshots, or numbers I type or copy in by hand",[1314],"s-better-input",{"id":1316,"text":1317,"points":1017,"triggers_fixes":1318},"b","Exports (CSV or spreadsheet files) that I upload or paste in",[1314],{"id":1320,"text":1321,"points":1322,"triggers_fixes":1323},"c","A connector to one platform at a time",8,[],{"id":1325,"text":1326,"points":1327,"triggers_fixes":1328},"d","Several connectors, used together in the same chat",7,[],{"id":1330,"text":1331,"points":1308,"triggers_fixes":1332},"e","One source where data from all our platforms is already joined and checked",[],{"id":1334,"text":1335,"points":5,"triggers_fixes":1336,"dont_know":261},"f","I don't know",[1337],"s-ai-setup-map",[1339,1340,1341],"shopify-access","meta-ppc","openai-apps",{"id":1343,"text":1344,"helper":1345,"why":1346,"max_points":1322,"options":1347,"sources":1363},"s2","Do you ask the AI to combine numbers from different platforms in one answer?","For example: total return on ad spend across Google Ads and Meta, or cost per order using ad spend and your shop's orders.","Each platform counts a sale by its own rules. Google Ads only credits Google ads and Meta only credits Meta ads, so a customer who clicked both is a sale in each, and adding them together over-counts every time. Google Ads also reports a sale against the date of the ad interaction (usually the click), not the day the sale happened, and each account can sit in a different time zone and currency. An AI that hasn't been told these rules lines up numbers that were never designed to match.",[1348,1351,1354,1358,1361],{"id":1311,"text":1349,"points":1322,"triggers_fixes":1350},"No, one platform per question",[],{"id":1316,"text":1352,"points":1322,"triggers_fixes":1353},"Yes, and it uses our shop's real orders as the total, with each platform's rules written down",[],{"id":1320,"text":1355,"points":5,"triggers_fixes":1356},"Yes, it adds up the sales each platform reports",[1357],"s-stop-adding",{"id":1325,"text":1359,"points":1133,"triggers_fixes":1360},"Yes, but I'm not sure how it combines them",[1357],{"id":1330,"text":1335,"points":5,"triggers_fixes":1362,"dont_know":261},[1357],[1364,1365,1366],"ads-dates","ads-tz","meta-measurement",{"id":1368,"text":1369,"helper":1370,"why":1371,"max_points":1372,"options":1373,"sources":1386},"s3","Can the AI see your shop's real orders, to compare with what Google Analytics and the ad platforms say?","Through a shop connector (Shopify has official ones for Claude, ChatGPT and other AI tools) or an order export from your shop admin.","Your shop admin is the only place that knows how many orders really exist. Google Analytics misses some (cookie choices, ad blockers, broken tags), and each ad platform credits itself. Without the real orders in the room, the AI can't tell a tracking problem from a real drop in sales, and it will explain either one with the same confidence.",9,[1374,1377,1381,1384],{"id":1311,"text":1375,"points":1372,"triggers_fixes":1376},"Yes, and it's told to treat the shop's orders as the true count",[],{"id":1316,"text":1378,"points":1017,"triggers_fixes":1379},"Yes, but as just another source, with no rule about which one wins",[1380],"s-real-orders",{"id":1320,"text":1382,"points":274,"triggers_fixes":1383},"No, it only sees Google Analytics or the ad platforms",[1380],{"id":1325,"text":1335,"points":5,"triggers_fixes":1385,"dont_know":261},[1380],[1339],{"id":1243,"text":1388,"helper":1389,"why":1390,"max_points":1372,"options":1391,"sources":1405},"Have you told the AI what your key terms mean?","What counts as a sale (paid orders only? before or after refunds?), how long after an ad click a sale still counts, and which time zone and currency to use.","\"Sales\" means something different in your shop admin, Google Analytics, Google Ads and Meta. Google Ads counts a sale up to 30 days after a click unless someone changed it; Meta and Google Analytics use their own settings. A Google Ads account's time zone and currency are fixed when the account is created, and Google Analytics converts other currencies using the previous day's exchange rate. The AI only allows for the differences it has been told about.",[1392,1395,1400,1403],{"id":1311,"text":1393,"points":1372,"triggers_fixes":1394},"Yes, written down in instructions it reads every time (for example a project or custom instructions)",[],{"id":1316,"text":1396,"points":1397,"triggers_fixes":1398},"Sometimes, in the chat, when I remember",4,[1399],"s-definitions",{"id":1320,"text":1401,"points":274,"triggers_fixes":1402},"No",[1399],{"id":1325,"text":1335,"points":5,"triggers_fixes":1404,"dont_know":261},[1399],[1406,1365,1407],"ads-window","ga4-currency",{"id":1409,"text":1410,"helper":1411,"why":1412,"max_points":1308,"options":1413,"sources":1426},"s5","When you ask for totals, percentages or comparisons, how is the maths done?","If the answer shows a block of code, or a note that the AI ran an analysis, it used code. If the numbers simply appear in the text, it probably worked them out itself.","AI assistants are good at explaining numbers and unreliable at long arithmetic done in their heads. Claude and ChatGPT can write and run code for calculations, which is far more reliable, but only when that feature is on and actually used. Best of all is when the numbers arrive already calculated, so the AI only has to explain them.",[1414,1417,1420,1424],{"id":1311,"text":1415,"points":1308,"triggers_fixes":1416},"The numbers arrive already calculated; the AI explains them",[],{"id":1316,"text":1418,"points":1322,"triggers_fixes":1419},"The AI writes and runs code to calculate them",[],{"id":1320,"text":1421,"points":253,"triggers_fixes":1422},"The AI works them out in its written answer",[1423],"s-code-maths",{"id":1325,"text":1335,"points":5,"triggers_fixes":1425,"dont_know":261},[1423],[1427,1428],"claude-code-exec","chatgpt-data-analysis",{"id":1430,"text":1431,"helper":1432,"why":1433,"max_points":1327,"options":1434,"sources":1446},"s6","Does the AI tell you which data each answer is based on?","The date range, which platform and account, and whether all the data came back or only part of it.","Connectors return data in pages and have size limits, so a long table can arrive cut short, and the AI can answer from a partial table without saying so. An answer that states its date range, source and number of rows is one you can check in a minute. One that doesn't, you can't.",[1435,1438,1442,1444],{"id":1311,"text":1436,"points":1327,"triggers_fixes":1437},"Yes, our instructions ask for it in every answer",[],{"id":1316,"text":1439,"points":1133,"triggers_fixes":1440},"Sometimes, if I ask",[1441],"s-show-working",{"id":1320,"text":1401,"points":274,"triggers_fixes":1443},[1441],{"id":1325,"text":1335,"points":5,"triggers_fixes":1445,"dont_know":261},[1441],[1447,1448,1449],"ga4-api","claude-code-mcp","gads-truncation",{"id":1451,"text":1452,"helper":1453,"why":1454,"max_points":1308,"options":1455,"sources":1467},"s7","Have you checked the AI's answers against the platforms themselves?","For example, asking for last month's sales and comparing the answer with the same report in Google Analytics, Google Ads, Meta or your shop admin.","This is the quickest way to know whether the AI reads your data correctly. Wrong answers usually look plausible, so checking only when something looks odd misses exactly the mistakes that matter. A one-off side-by-side test, then spot checks on any number you act on, catches most of them.",[1456,1459,1462,1465],{"id":1311,"text":1457,"points":1308,"triggers_fixes":1458},"Yes, we tested it side by side once, and we still spot-check numbers we act on",[],{"id":1316,"text":1460,"points":1397,"triggers_fixes":1461},"Only when a number looks odd",[1130],{"id":1320,"text":1463,"points":274,"triggers_fixes":1464},"No, never",[1130],{"id":1325,"text":1335,"points":5,"triggers_fixes":1466,"dont_know":261},[1130],[],{"id":1260,"text":1469,"helper":1470,"why":1471,"max_points":1308,"options":1472,"sources":1486},"When did someone last check that the tracking itself is right?","For example, a test order followed all the way through, or 30 days of Google Analytics orders compared with your shop admin. This is about the data, not the AI.","An AI that reads your data perfectly still repeats whatever the tracking recorded. If a theme update dropped the purchase tag on Tuesday, the AI will report a drop in sales from Tuesday, with reasons. Checking the AI and checking the tracking are two different jobs.",[1473,1476,1481,1484],{"id":1311,"text":1474,"points":1308,"triggers_fixes":1475},"In the last month",[],{"id":1316,"text":1477,"points":1478,"triggers_fixes":1479},"Between 1 and 6 months ago",6,[1480],"s-tracking-check",{"id":1320,"text":1482,"points":274,"triggers_fixes":1483},"More than 6 months ago, or never",[1480],{"id":1325,"text":1335,"points":5,"triggers_fixes":1485,"dont_know":261},[1480],[],{"id":1488,"text":1489,"helper":1490,"why":1491,"max_points":1322,"options":1492,"sources":1505},"s9","Does anyone check your numbers on a schedule, even when nobody has asked a question?","A person with a weekly routine counts. So does a scheduled check, if someone actually reads the result.","An AI assistant answers when you ask. A tracking break on a Friday can sit unnoticed until someone happens to ask the right question, and meanwhile the ad platforms keep learning from the wrong numbers. Some assistants can run scheduled tasks, but those repeat the same unchecked steps unless someone reviews the result.",[1493,1496,1500,1503],{"id":1311,"text":1494,"points":1322,"triggers_fixes":1495},"Yes, weekly or more often",[],{"id":1316,"text":1497,"points":1397,"triggers_fixes":1498},"Monthly, when we do the reports",[1499],"s-weekly-check",{"id":1320,"text":1501,"points":5,"triggers_fixes":1502},"Only when something looks wrong",[1499],{"id":1325,"text":1335,"points":5,"triggers_fixes":1504,"dont_know":261},[1499],[],{"id":1507,"text":1508,"helper":1509,"why":1510,"max_points":1478,"options":1511,"sources":1524},"s10","When an AI answer leads to a decision, what happens before anyone acts on it?","Decisions like moving budget, pausing a campaign, or putting a number in a report for the owner, the team or a client.","The risk isn't one wrong answer. It's a wrong answer that travels. Once an AI number has moved budget or landed in a report, nobody goes back to check it. One rule fixes most of this: someone who knows the platforms checks the number before it changes money or goes into a report.",[1512,1515,1519,1522],{"id":1311,"text":1513,"points":1478,"triggers_fixes":1514},"Someone who knows the platforms checks the key number first",[],{"id":1316,"text":1516,"points":1133,"triggers_fixes":1517},"The person who asked decides on their own",[1518],"s-check-before-acting",{"id":1320,"text":1520,"points":274,"triggers_fixes":1521},"The answer goes into a report or a message as it is",[1518],{"id":1325,"text":1335,"points":5,"triggers_fixes":1523,"dont_know":261},[1518],[],{"id":1526,"text":1527,"helper":1528,"why":1529,"max_points":1327,"options":1530,"sources":1543},"s11","Can the AI change anything in your ad accounts or your shop?","For example: create or edit campaigns, change budgets, or update products, prices, inventory or discount codes. Meta's own connector and Shopify's connector can both make changes if you allow it.","Some connectors can write as well as read. Meta's connector can create and edit campaigns, ad sets and ads. New ones start paused and switching them on needs your confirmation, but Meta doesn't describe a separate confirmation step for edits to campaigns that are already running. Shopify's connector, with write access, can update product listings and prices, inventory and discount codes, and Shopify says it doesn't review those actions. \"Don't change anything\" in your instructions helps, but it's an instruction to the AI, not a permission.",[1531,1534,1538,1541],{"id":1311,"text":1532,"points":1327,"triggers_fixes":1533},"No, it only has read-only access",[],{"id":1316,"text":1535,"points":1017,"triggers_fixes":1536},"Yes, but we approve every change, and someone checks it afterwards",[1537],"s-read-only",{"id":1320,"text":1539,"points":274,"triggers_fixes":1540},"Yes, and changes can go through without anyone checking",[1537],{"id":1325,"text":1335,"points":5,"triggers_fixes":1542,"dont_know":261},[1537],[1544,1339,1545,1546],"meta-tools","shopify-considerations","shopify-claude-help",{"id":1277,"text":1548,"helper":1549,"why":1550,"max_points":1478,"options":1551,"sources":1564},"Do you know what business data you share with the AI, and on which plan?","For example, whether people use a personal or a business plan, whether your chats can be used to train the AI's models, and whether customer names and emails ever go in.","Personal and business plans handle your data differently. On the personal plans of Claude, ChatGPT, Gemini and Copilot there is a setting that decides whether your chats help train the models; on business plans, these providers say they don't train on your business data by default. Customer names, emails and addresses are personal data under GDPR, and an analyst rarely needs them. This is not legal advice: your data protection lead or adviser has the final word.",[1552,1555,1559,1562],{"id":1311,"text":1553,"points":1478,"triggers_fixes":1554},"Yes: we've checked our plan and its training setting, and we share totals, not customer details",[],{"id":1316,"text":1556,"points":1133,"triggers_fixes":1557},"Partly: we've checked one of those, not both",[1558],"s-data-basics",{"id":1320,"text":1560,"points":274,"triggers_fixes":1561},"No, we haven't looked",[1558],{"id":1325,"text":1335,"points":5,"triggers_fixes":1563,"dont_know":261},[1558],[1565,1566,1567,1568,1569],"anthropic-consumer","openai-training","gemini-apps","copilot-consumer","gdpr",[1571,1587,1603,1620,1635,1656,1670,1684,1703,1718,1733,1748,1763],{"id":1337,"title":1572,"impact":1322,"effort":1573,"explanation":1574,"diy_steps":1575,"for_implementer":1580,"sources":1581,"triggered_by":1584},"Write down how your AI gets its numbers","About 30 minutes, with whoever set it up","Not knowing how your AI gets its numbers usually means the setup grew one chat at a time, so nobody has the full picture. Before fixing anything, write one page: which assistants and plans people use, which platforms and accounts they're connected to, who connected them, what the AI is allowed to change, and where your definitions live. Most of the other fixes take minutes once this page exists.",[1576,1577,1578,1579],"List every AI assistant people in your team use for numbers, and whether each is a personal or a business plan.","For each one, list what it's connected to: the platform, the account, who signed in, and whether it can make changes. Claude and ChatGPT list connected tools in their settings (Claude calls them connectors, ChatGPT calls them apps).","Note where instructions and definitions live, if anywhere: a project, custom instructions or a shared prompt.","Remove connections nobody uses, and any signed in with a former colleague's or agency's login.","ChatGPT: Settings > Apps. Claude: the connectors list in settings (Anthropic's help pages refer to both Customize > Connectors and Settings > Connectors; check which your plan shows). On Claude Team and Enterprise, owners can set each connector's tools to Always allow, Needs approval or Blocked, grouped into read-only and write\u002Fdelete tools. Meta's connector inherits everything the signed-in Facebook profile can reach, which for agencies can include client accounts.",[1341,1582,1583],"claude-connectors","meta-access",[1585,1586],"s1.f","rule: dont_know_count >= 3",{"id":1314,"title":1588,"impact":1322,"effort":1589,"explanation":1590,"diy_steps":1591,"for_implementer":1596,"sources":1597,"triggered_by":1600},"Swap screenshots and copy-paste for a direct connection or a complete export","15-60 minutes per platform","Screenshots and pasted tables give the AI a picture of part of your data. Numbers can be misread, rows cut off, and any filter you applied is invisible to it. A connector fetches the data itself, for the date range you ask about. If you stay with files, export the whole report, not just what was on screen.",[1592,1593,1594,1595],"Pick the platform you ask about most and connect it directly. Our connection guides walk through Google Analytics, Google Ads and Meta for Claude and ChatGPT, step by step.","If you stay with files, export the full report (not only the rows on screen), and put the date range and currency in the file name.","Before any analysis, ask the AI to repeat back the date range, the number of rows and the totals it sees. Check them against the platform.","Whichever route you use, run the trust test before relying on it.","Official routes as of October 2026: Shopify's connector (one store at a time; read-only or read and write); Meta's Ads AI Connectors (open beta since 29 April 2026; read and write); Google's GA4 and Google Ads MCP servers are self-hosted developer setups (Python, a Google Cloud project, OAuth) and read-only. Third-party vendor connectors cover most platforms without code. Large exports can be too big for the assistant to read in full.",[1339,1340,1598,1599],"ga4-mcp","gads-mcp",[1601,1602],"s1.a","s1.b",{"id":1357,"title":1604,"impact":1372,"effort":1605,"explanation":1606,"diy_steps":1607,"for_implementer":1612,"sources":1613,"triggered_by":1616},"Stop adding up the sales each platform reports","10 minutes to change your instructions, then a habit","Google Ads only credits Google ads; Meta only credits Meta ads. A customer who clicked both is a sale in each, so adding them together over-counts every time. For cross-channel questions, use your shop's real orders as the total, and have the AI treat each platform's figure as that platform's claim, not a share of the total.",[1608,1609,1610,1611],"Add this to your instructions: \"Never add sales or conversions from different ad platforms together. The shop's orders are the true total.\"","For cost per order across channels, divide total ad spend by real shop orders, not by the purchases the platforms claim.","When the AI compares platforms, ask it to list each platform's own rules (attribution window, date logic, time zone) next to its number.","Read our explainer on why Meta, Google Ads, Google Analytics and your shop never agree, then decide which number answers which question.","Google Ads' standard conversion columns put each conversion on the date of the ad interaction that led to it (usually the click), while other tools use the date the conversion happened (Google Ads Help); add \"Conversions (by conv. time)\" columns to line Google Ads up with Google Analytics, and allow 24-48 hours before comparing. Meta said in March 2026 that differences in what counts as a click \"can lead to inconsistency\" between Ads Manager and third-party tools. Within one platform, duplicates are a genuine fault: for example, a purchase sent by both the Meta pixel and the Conversions API without a shared event ID.",[1364,1614,1366,1615],"ads-conv-time","meta-dedup",[1617,1618,1619],"s2.c","s2.d","s2.e",{"id":1380,"title":1621,"impact":1372,"effort":1622,"explanation":1623,"diy_steps":1624,"for_implementer":1629,"sources":1630,"triggered_by":1631},"Give the AI your real orders, and tell it they win","15-30 minutes","Your shop's order count is the reference everything else should be compared with. Connect your shop (Shopify has an official connector; other platforms through a vendor connector or an export), and tell the AI that when the shop and a platform disagree, the shop's orders are the real count and the gap is something to explain.",[1625,1626,1627,1628],"Connect your shop with read-only access if offered, or export 30 days of daily orders and revenue without customer columns.","Add to your instructions: \"The shop's orders are the true count. Report Google Analytics and ad platform figures as gaps against it.\"","Decide which orders count (paid only? before or after refunds? without test and cancelled orders?) and write it down.","Ask: \"For each day last month, show shop orders, Google Analytics purchases and the gap.\" A sudden jump in the gap usually means tracking changed, not demand.","Shopify: access levels can't be lowered after approval without uninstalling and reconnecting; one store at a time; start a new conversation after switching stores. Shop revenue and GA4 purchase revenue differ by design (refunds, taxes, shipping, cancelled orders), so state which revenue definition is the reference. Leave customer names, emails and addresses out of exports.",[1339],[1632,1633,1634],"s3.b","s3.c","s3.d",{"id":1399,"title":1636,"impact":1372,"effort":1637,"explanation":1638,"diy_steps":1639,"link":1645,"for_implementer":1649,"sources":1650,"triggered_by":1652},"Write your definitions down once, where the AI reads them every time","About an hour, once","A one-page definitions file stops the AI guessing. It says what counts as a sale, which revenue you mean (before or after refunds and VAT), each platform's attribution window and date logic, and each account's time zone and currency. Put it where the AI reads it on every question, not in message three of an old chat.",[1640,1641,1642,1643,1644],"Note each account's time zone and currency: Google Analytics property settings, Google Ads account settings, your Meta ad account settings and your shop settings.","Note each platform's attribution setting. In Google Ads each conversion action has its own window (30 days after a click unless someone changed it); Meta and Google Analytics have their own settings.","Write down what counts as an order and as revenue for your business.","Save it as project instructions or a project file (or custom instructions, depending on your assistant), and start a new chat for each question.","Our shop stack guide includes a ready-made definitions template you can fill in.",{"type":1646,"label":1647,"url":1648},"text_link","A ready template for those definitions, and the house rules to go with them: Your AI Marketing Workspace Starter Kit","https:\u002F\u002Ftrackian.com\u002Fresources\u002Fai-marketing-workspace-starter-kit","Google Ads time zone and currency are permanently set when the account is created. GA4 converts other currencies into the property currency using the prior day's exchange rate. The Google Ads default click-through conversion window is 30 days (1-90 days, depending on the conversion source). Confirm Meta's attribution setting per ad set in Ads Manager rather than assuming a default.",[1365,1407,1406,1651],"chatgpt-projects",[1653,1654,1655],"s4.b","s4.c","s4.d",{"id":1423,"title":1657,"impact":1372,"effort":1658,"explanation":1659,"diy_steps":1660,"for_implementer":1665,"sources":1666,"triggered_by":1667},"Make the AI calculate in code, not in its head","5 minutes","Claude and ChatGPT can write and run code for calculations, which is far more reliable than arithmetic written out in a chat answer. Make sure the feature is on, and ask for it. Better still, start from numbers that arrive already calculated, so the AI only has to explain them.",[1661,1662,1663,1664],"Check the setting. In Claude, \"Code execution and file creation\" is under Settings > Capabilities (on by default on personal plans; on Team and Enterprise an owner controls it). ChatGPT writes and runs Python for some data-analysis tasks.","Add to your instructions: \"Calculate every total, ratio and comparison with code, and show the code.\"","Redo one calculation yourself for two or three rows. If they match, the method works; the inputs are a separate question.","On Gemini, Copilot or another assistant, check its help pages for an equivalent feature before relying on calculated answers.","Code fixes arithmetic, not inputs. If a connector returned a partial table, code calculates a precise wrong number. Pair this with \"Ask every answer to say what it's based on\".",[1427,1428],[1668,1669],"s5.c","s5.d",{"id":1441,"title":1671,"impact":1327,"effort":1658,"explanation":1672,"diy_steps":1673,"for_implementer":1677,"sources":1678,"triggered_by":1680},"Ask every answer to say what it's based on","Add one line to your instructions so every answer states the date range, the platform and account, the number of rows used, and whether anything came back incomplete. It takes seconds to read and turns a confident answer into a checkable one.",[1674,1675,1676],"Add to your instructions: \"Start every answer with the date range, source, account and number of rows used. Say if any data was missing, cut off or estimated.\"","Ask about date ranges that ended a few days ago. Recent data still changes (Google Analytics can take 24-48 hours to process).","If an answer arrives without that line, ask for it before you use the number.","The GA4 Data API returns up to 10,000 rows per request by default and pages beyond that. In Claude Code, MCP tool output above 10,000 tokens triggers a warning, with a default cap of 25,000. Large Google Ads MCP results can be truncated by the assistant's context window. Ask for aggregated results (totals by day or campaign) rather than raw rows.",[1447,1448,1449,1679],"ga4-fresh",[1681,1682,1683],"s6.b","s6.c","s6.d",{"id":1130,"title":1685,"impact":1372,"effort":1686,"explanation":1687,"diy_steps":1688,"for_implementer":1694,"sources":1695,"triggered_by":1696,"link":1700},"Run a 10-minute trust test before you rely on it","10 minutes per platform; 20 minutes for several together","Ask the AI five questions about a closed date range and check each answer in the platform itself: a total, a breakdown, your purchases and revenue, a top-10 list and one calculation. If all five match, the AI reads that platform correctly. If two or more don't, don't use its answers for decisions until you've fixed the connection or your instructions.",[1689,1690,1691,1692,1693],"Pick a date range that ended at least a few days ago, so the numbers have stopped changing.","Ask for a headline total, a breakdown by channel or campaign, your purchases and revenue, a top-10 list, and one ratio with \"show your working\".","Check each answer in the platform, side by side, and write both numbers down.","Use the step-by-step version for your setup, linked below: chapter 5 of the connection guide for your platform, the 20-minute cross-source test in the shop stack guide if you combine several, or the 30-second AI answer check if you use another assistant.","Run it again after any change to connectors, plans or tracking.","The test checks the AI's reading and maths, not the tracking. Passing proves the AI matches the platform's own interface; it doesn't prove the platform recorded every sale. For that, see \"Check the tracking itself, not just the AI\".",[],[1697,1698,1699],"s7.b","s7.c","s7.d",{"type":1646,"label":1701,"url_ref":1702},"Open the trust test for your setup: \"{route.trust_test.guide}\", {route.trust_test.section}","routing.trust_test",{"id":1480,"title":1704,"impact":1308,"effort":1705,"explanation":1706,"diy_steps":1707,"for_implementer":1712,"sources":1713,"triggered_by":1714},"Check the tracking itself, not just the AI","About 30 minutes, with the free Missing Orders sheet","Follow one test order all the way through, and compare 30 days of daily orders in Google Analytics with your shop admin. This tells you whether the data the AI reads is close to what really happened. A small, steady gap is normal; a big or sudden one usually means something broke.",[1708,1709,1710,1711],"Place a test order (a discount code or a cheap product is fine) and check it appears once in your shop, in Google Analytics and in your ad platforms' conversion reports.","Paste 30 days of daily orders from your shop and from Google Analytics into the free Missing Orders sheet. It shows the gap in orders and in euros.","Look for days where the gap jumped, and check what changed on the site that day.","Repeat after every theme, checkout, payment or cookie banner change.","Use a range that ended at least 3 days ago (GA4 processing can take 24-48 hours). Compare the GA4 purchase event count with the number of unique transaction IDs to spot duplicates. If you'd rather a specialist did it, the 24-hour Missing Orders Check compares your tracking with your real orders.",[1679],[1715,1716,1717],"s8.b","s8.c","s8.d",{"id":1499,"title":1719,"impact":1327,"effort":1720,"explanation":1721,"diy_steps":1722,"for_implementer":1727,"sources":1728,"triggered_by":1729},"Put a 15-minute numbers check in the diary every week","15 minutes a week","Pick a fixed time each week. Compare last week's orders in your shop with Google Analytics and each ad platform, note anything unusual, and write down any site or campaign changes. It's boring, which is exactly why it catches things.",[1723,1724,1725,1726],"Same day, same time, same person, with a named backup.","Compare last week's shop orders with Google Analytics purchases and each ad platform's reported purchases. Note the gaps.","If a gap moved much more than usual, check the tracking before you change any budget.","If an AI scheduled task does the comparing, make sure a person reads and questions the result.","Scheduled AI tasks (some assistants offer them on paid plans) repeat the same chat-based steps. Someone still has to review the output and know what normal looks like for this shop. Keep a simple log of site, tracking and campaign changes next to the weekly numbers.",[],[1730,1731,1732],"s9.b","s9.c","s9.d",{"id":1518,"title":1734,"impact":1327,"effort":1735,"explanation":1736,"diy_steps":1737,"for_implementer":1742,"sources":1743,"triggered_by":1744},"One rule: no money moves on an unchecked AI number","A 10-minute team conversation","Agree that any AI number that moves budget, pauses a campaign or goes into a report is checked in the platform first, by someone who knows it. It costs a minute per decision and stops a wrong answer from travelling.",[1738,1739,1740,1741],"Name who checks each platform, and a backup.","For each decision, check the one or two numbers it rests on in the platform itself.","When an AI number goes into a report, add the source and the date range next to it.","If people outside the team (the owner, an agency, a client) receive AI-generated numbers, tell them which ones were checked.","Keep the rule in the same place as the definitions file, so the AI can remind people of it: for example, \"End every answer that recommends a budget change with: check these numbers in the platform before acting.\"",[],[1745,1746,1747],"s10.b","s10.c","s10.d",{"id":1537,"title":1749,"impact":1322,"effort":1622,"explanation":1750,"diy_steps":1751,"for_implementer":1756,"sources":1757,"triggered_by":1759},"Give the AI read-only access, or approve every change","Analysis doesn't need write access. Where a connector offers read-only access, use it. Where it can write, read every proposed change before you approve it, and check the platform's own history afterwards. An instruction like \"never change anything\" helps, but it isn't a permission.",[1752,1753,1754,1755],"List which connectors can make changes. Meta's connector can; Shopify's can if write access was approved.","Shopify: if you don't need write access, uninstall the connection and reconnect with read-only access (it can't be lowered any other way).","Meta: new campaigns, ad sets and ads are created paused, and switching them on asks for confirmation. Edits to running campaigns are the risk. Read every proposed change before approving it, then check the account's activity history in Ads Manager.","Keep editing in the platforms. Use the AI to analyse.","On Claude Team and Enterprise, owners can set a connector's write\u002Fdelete tools to Needs approval or Blocked, organisation-wide. Meta's connector inherits everything the signed-in profile can reach; use a profile with the least access that works. Significant edits (targeting, creative, optimisation event, bid strategy, large budget changes) can send an ad set back into Meta's learning phase. There's no undo for a change the AI made for you.",[1544,1339,1545,1546,1582,1583,1758],"meta-learning",[1760,1761,1762],"s11.b","s11.c","s11.d",{"id":1558,"title":1764,"impact":1478,"effort":1765,"explanation":1766,"diy_steps":1767,"for_implementer":1772,"sources":1773,"triggered_by":1777},"Check your plan, and share totals, not customers","About 20 minutes","Find out which plan each person uses and what it does with your chats, and keep customer details out. Analysis almost never needs names, emails or addresses: totals and counts by day, channel or campaign do the job. This isn't legal advice; your data protection lead or adviser has the final word.",[1768,1769,1770,1771],"On personal plans, find the setting that decides whether chats help train the models. As of October 2026: Claude, the model training choice in privacy settings; ChatGPT, \"Improve the model for everyone\" under Data controls; Gemini, \"Keep Activity\"; Microsoft Copilot, \"Training on conversation activity\" under Privacy.","Prefer a business plan for company data. Anthropic (commercial terms), OpenAI (ChatGPT Business and Enterprise), Google (Gemini in Workspace) and Microsoft (Microsoft 365 Copilot with a work account) say they don't train their models on that business data by default.","Export orders without customer columns. With connectors, ask for totals and counts, not lists of customers.","Check each connector vendor: where it processes data, and whether you need a data processing agreement with it (GDPR Article 28).","Claude personal plans: if you allow training, Anthropic may keep de-identified data for up to five years; with training off, deleted chats leave Anthropic's systems within 30 days. Gemini with a personal account: with Keep Activity on, a subset of chats is reviewed by people, and Google advises not to enter confidential data. Shopify says data shared with a connected AI tool is subject to that tool's terms, and that the merchant is responsible for data shared with third-party services.",[1565,1774,1566,1567,1775,1568,1776,1545,1569],"anthropic-commercial","workspace-ai","m365-edp",[1778,1779,1780],"s12.b","s12.c","s12.d",[1130,1480,1499],[1783],{"fix":1337,"when":1784,"rank":1785},"dont_know_count >= 3","first",[1787,1793,1799,1806],{"id":1788,"min":274,"max":1789,"name":1790,"headline":1791,"body":1792},"b1",39,"Confident guesswork","Your AI sounds sure. Right now, there's no way to know if it's right.","AI assistants are genuinely useful for marketing questions. Based on your answers, though, yours is working from shop data that hasn't been joined or checked, so its answers can be confidently wrong without anyone noticing. That's fixable, and the first steps are free: test whether it reads your data correctly, then check the tracking underneath.",{"id":1794,"min":924,"max":1795,"name":1796,"headline":1797,"body":1798},"b2",59,"Trusted, not verified","Some good habits, but too much rests on answers nobody has checked.","Parts of your setup are sound. The gaps are the ones that produce plausible wrong numbers: platforms added together, definitions the AI has to guess, or tracking nobody has tested lately. Fix the top three below, then build the setup properly, with every source connected, each platform's rules written down and a test you can repeat.",{"id":1800,"min":1801,"max":1802,"name":1803,"headline":1804,"body":1805},"b3",60,79,"Mostly checked","A careful setup with a few blind spots.","You check more than most. The remaining gaps tend to show up when several sources are combined, or when tracking changes quietly between checks. Close them with the fixes below and the cross-source rules in the shop stack guide.",{"id":1807,"min":1808,"max":1113,"name":1809,"headline":1810,"body":1811},"b4",80,"Checked and joined","Your AI's answers rest on solid ground. Now keep it that way.","Your answers describe a setup where the AI gets complete data, calculates in code and gets checked. The work now is doing that every week: connectors change, plans change, tracking drifts, and someone has to notice.",{"heading":1813,"body":1814},"Doing this yourself, every week","Every fix above can be done without Trackian, and an AI assistant is a real help once it's set up well. The hard part is keeping it that way. Each connector brings one platform with its own rules, the AI has to be told how they differ, someone has to check its answers against the platforms, and someone has to check the tracking underneath. Every week, because connectors, plans and tracking change without telling you.",{"heading":1816,"body":1817},"Or let your AI explain numbers that are already joined and checked",[1818,1819,1820,1821],"Trackian is a marketing brain for your business. It joins your ads, site and shop data through 40+ integrations, including Google Ads, Meta Ads, Microsoft Ads, Google Analytics (GA4) and your store, whether it runs on Shopify, WooCommerce, Magento, PrestaShop or OpenCart. Your ERP or other tools connect through API, webhooks or MCP.","Every number is calculated in code from the joined data and checked every night: what moved, why, and whether it's real or broken tracking. A specialist reviews findings before they reach you.","Prefer to keep asking Claude, ChatGPT or another AI assistant? The Trackian MCP connector works in Claude, ChatGPT, Gemini, Microsoft Copilot and other MCP-ready assistants, so the numbers your AI explains are already joined, checked and calculated in code, with over 130 built-in AI tools for breakdowns. The AI explains them; it doesn't calculate them. The connector can also make changes in your ad accounts, but only after it shows you a preview and you confirm. Add it from your Trackian account.","If you'd rather not fix it yourself, our specialists can: the Fix Sprint covers GA4, Google Tag Manager, and Google Ads and Meta conversion settings. It's delivered by people and priced separately.",[1823],{"id":1824,"order":5,"type":1825,"heading":1826,"label":1827,"subtext":1828,"own_numbers_line":1829,"note":1830,"url":1831,"show_when":1832},"check_24h","primary","Step 3 - get your numbers checked in 24 hours, free.","Get my free Missing Orders Check","A Trackian specialist compares what your tracking recorded with your real orders and tells you within 24 hours what's missing, why, and what to fix first. Free - we ask a few short questions first to make sure it's a fit.","You said you spend {spend_band_label} a month on ads and get {volume_band_label} orders a month. Your AI's answers about those orders are only as good as the tracking underneath - the check compares it with your real orders.","Delivered by people, not scripts. Not sure you can trust your numbers, or the people reporting them? This is the second opinion.","https:\u002F\u002Ftrackian.com\u002Fmissing-orders-check","always",[1834],{"id":1835,"type":1646,"label":1836,"subtext":1837,"url":1838,"placement":1839},"missing_sheet","Get the free Missing Orders sheet","Paste 30 days of Google Analytics and shop numbers into a simple spreadsheet and see the gap in orders and in euros. The quickest way to check the data your AI reads.","https:\u002F\u002Ftrackian.com\u002Fresources\u002Fmissing-orders-leads-sheet\u002Forders","directly under the 3 fixes, above \"Doing this yourself, every week\"",{"b1":1841,"b2":1844,"b3":1847,"b4":1849},{"type":1646,"label":1842,"url":1843},"Step 2 of 3: learn the nine ways AI gets marketing numbers wrong, and how to catch each one in 30 seconds","https:\u002F\u002Ftrackian.com\u002Fresources\u002Fnine-ways-ai-gets-numbers-wrong",{"type":1646,"label":1845,"url_ref":1846},"Step 2 of 3: build it properly with the free stack guide \"{route.stack_guide.guide}\": every source connected, a rules file your AI reads every time, and a 20-minute cross-source trust test","routing.stack_guide",{"type":1646,"label":1848,"url_ref":1846},"Step 2 of 3: close the gaps with the cross-source rules and the 20-minute trust test in the free stack guide \"{route.stack_guide.guide}\"",{"type":1646,"label":1850,"url":1851},"Step 2 of 3: keep it this way with the owner's weekly numbers - seven numbers every Monday in about 10 minutes, with a tracking check built in","https:\u002F\u002Ftrackian.com\u002Fresources\u002Fowners-weekly-numbers\u002Fshop",{"id":1047,"name":1853,"short_name":1854,"intro":1855,"qualification":1860,"questions":1896,"fixes":2080,"fallback_fixes":2226,"conditional_fixes":2227,"bands":2229,"diy_reality":2238,"pitch":2240,"ctas":2243,"content_links":2249,"next_step_by_band":2254},"Can you trust your AI's answers about your leads?","lead-gen scorecard",{"headline":1856,"subheadline":1857,"meta_line":1035,"start_button":998,"seasonal_variants":1858,"preselect_param":1859},"Your AI says your cheapest leads come from Meta. Does your sales team agree?","A few plain-language questions about how your AI assistant gets, combines and checks your lead numbers. Get a trust score out of 100, the 3 fixes that matter most and the right next step.",[],"?v=leads",[1861,1868,1884,1891],{"id":1241,"text":1242,"helper":1057,"scored":266,"ask_after":1862,"options":1863},"l4",[1864,1865,1866,1867],{"id":1246,"text":1247},{"id":1249,"text":1250},{"id":1252,"text":1253},{"id":1255,"text":1256},{"id":1258,"text":1869,"helper":1057,"scored":266,"ask_after":1870,"options":1871},"Roughly how many leads do you get a month?","l8",[1872,1875,1878,1881],{"id":1873,"text":1874},"under_20","Under 20",{"id":1876,"text":1877},"20_49","20-49",{"id":1879,"text":1880},"50_199","50-199",{"id":1882,"text":1883},"200_plus","200+",{"id":1275,"text":1276,"helper":1057,"scored":266,"ask_after":1885,"options":1886},"l12",[1887,1888,1889,1890],{"id":1280,"text":1281},{"id":1283,"text":1284},{"id":1286,"text":1287},{"id":1072,"text":1085},{"id":1290,"text":1291,"helper":1057,"scored":266,"ask_after":1275,"options":1892},[1893,1894,1895],{"id":1294,"text":1295},{"id":1297,"text":1298},{"id":1300,"text":1301},[1897,1916,1935,1952,1967,1980,1993,2006,2021,2035,2049,2065],{"id":1898,"text":1305,"helper":1899,"why":1307,"max_points":1308,"options":1900,"sources":1915},"l1","Pick the one you use most. A connector (ChatGPT calls them apps) is a direct link between the AI and a platform such as Google Analytics, Google Ads or Meta.",[1901,1904,1906,1908,1910,1912],{"id":1311,"text":1312,"points":253,"triggers_fixes":1902},[1903],"l-better-input",{"id":1316,"text":1317,"points":1017,"triggers_fixes":1905},[1903],{"id":1320,"text":1321,"points":1322,"triggers_fixes":1907},[],{"id":1325,"text":1326,"points":1327,"triggers_fixes":1909},[],{"id":1330,"text":1331,"points":1308,"triggers_fixes":1911},[],{"id":1334,"text":1335,"points":5,"triggers_fixes":1913,"dont_know":261},[1914],"l-ai-setup-map",[1340,1341],{"id":1917,"text":1344,"helper":1918,"why":1919,"max_points":1322,"options":1920,"sources":1934},"l2","For example: total cost per lead across Google Ads and Meta, or cost per real lead using ad spend and your own list of leads.","Each platform counts a lead by its own rules. Google Ads only credits Google ads and Meta only credits Meta ads, so one enquiry from someone who clicked both can be a lead in each, and adding them together over-counts. Google Ads also reports a lead against the date of the ad interaction (usually the click), not the day the enquiry arrived, and each account can sit in a different time zone. An AI that hasn't been told these rules lines up numbers that were never designed to match.",[1921,1923,1926,1930,1932],{"id":1311,"text":1349,"points":1322,"triggers_fixes":1922},[],{"id":1316,"text":1924,"points":1322,"triggers_fixes":1925},"Yes, and it uses our own lead list as the total, with each platform's rules written down",[],{"id":1320,"text":1927,"points":5,"triggers_fixes":1928},"Yes, it adds up the leads each platform reports",[1929],"l-stop-adding",{"id":1325,"text":1359,"points":1133,"triggers_fixes":1931},[1929],{"id":1330,"text":1335,"points":5,"triggers_fixes":1933,"dont_know":261},[1929],[1364,1365,1366],{"id":1936,"text":1937,"helper":1938,"why":1939,"max_points":1372,"options":1940,"sources":1951},"l3","Can the AI see your real list of leads, and which ones turned into customers?","For example, an export from your form tool, inbox, booking system or CRM, with junk and duplicates marked.","Your own lead list is the only record of the enquiries that really arrived, and only your team knows which were junk and which became customers. Ad platforms count form submissions, including spam and duplicates, and can miss calls and bookings. Without your list, the AI can't tell cheap leads from good ones, or a tracking problem from a real drop in enquiries.",[1941,1944,1947,1949],{"id":1311,"text":1942,"points":1372,"triggers_fixes":1943},"Yes, and it's told to treat our list as the true count",[],{"id":1316,"text":1378,"points":1017,"triggers_fixes":1945},[1946],"l-real-leads",{"id":1320,"text":1382,"points":274,"triggers_fixes":1948},[1946],{"id":1325,"text":1335,"points":5,"triggers_fixes":1950,"dont_know":261},[1946],[],{"id":1862,"text":1388,"helper":1953,"why":1954,"max_points":1372,"options":1955,"sources":1965},"What counts as a lead (any form? only real enquiries? calls and bookings too?), whether two forms from one person are one lead or two, how long after an ad click a lead still counts, and which time zone to use.","\"Leads\" means something different in your inbox, Google Analytics, Google Ads and Meta. In Google Ads, each conversion action can count every conversion after a click or only one, and Google recommends counting one for leads. Google Ads counts a lead up to 30 days after a click unless someone changed it, and an account's time zone is fixed when it's created. The AI only allows for the differences it has been told about.",[1956,1958,1961,1963],{"id":1311,"text":1393,"points":1372,"triggers_fixes":1957},[],{"id":1316,"text":1396,"points":1397,"triggers_fixes":1959},[1960],"l-definitions",{"id":1320,"text":1401,"points":274,"triggers_fixes":1962},[1960],{"id":1325,"text":1335,"points":5,"triggers_fixes":1964,"dont_know":261},[1960],[1966,1406,1365],"ads-count",{"id":1968,"text":1410,"helper":1411,"why":1412,"max_points":1308,"options":1969,"sources":1979},"l5",[1970,1972,1974,1977],{"id":1311,"text":1415,"points":1308,"triggers_fixes":1971},[],{"id":1316,"text":1418,"points":1322,"triggers_fixes":1973},[],{"id":1320,"text":1421,"points":253,"triggers_fixes":1975},[1976],"l-code-maths",{"id":1325,"text":1335,"points":5,"triggers_fixes":1978,"dont_know":261},[1976],[1427,1428],{"id":1981,"text":1431,"helper":1432,"why":1433,"max_points":1327,"options":1982,"sources":1992},"l6",[1983,1985,1988,1990],{"id":1311,"text":1436,"points":1327,"triggers_fixes":1984},[],{"id":1316,"text":1439,"points":1133,"triggers_fixes":1986},[1987],"l-show-working",{"id":1320,"text":1401,"points":274,"triggers_fixes":1989},[1987],{"id":1325,"text":1335,"points":5,"triggers_fixes":1991,"dont_know":261},[1987],[1447,1448,1449],{"id":1994,"text":1452,"helper":1995,"why":1454,"max_points":1308,"options":1996,"sources":2005},"l7","For example, asking for last month's leads and comparing the answer with the same report in Google Ads, Meta, Google Analytics or your own lead list.",[1997,1999,2001,2003],{"id":1311,"text":1457,"points":1308,"triggers_fixes":1998},[],{"id":1316,"text":1460,"points":1397,"triggers_fixes":2000},[1131],{"id":1320,"text":1463,"points":274,"triggers_fixes":2002},[1131],{"id":1325,"text":1335,"points":5,"triggers_fixes":2004,"dont_know":261},[1131],[],{"id":1870,"text":2007,"helper":2008,"why":2009,"max_points":1308,"options":2010,"sources":2020},"When did someone last check that the lead tracking itself is right?","For example, a test enquiry sent through every form, or 12 weeks of reported leads compared with the leads you really received. This is about the data, not the AI.","An AI that reads your data perfectly still repeats whatever the tracking recorded. If a form plugin update stopped the lead event on Tuesday, the AI will report a drop in leads from Tuesday, with reasons, while your inbox stays as busy as ever. Checking the AI and checking the tracking are two different jobs.",[2011,2013,2016,2018],{"id":1311,"text":1474,"points":1308,"triggers_fixes":2012},[],{"id":1316,"text":1477,"points":1478,"triggers_fixes":2014},[2015],"l-tracking-check",{"id":1320,"text":1482,"points":274,"triggers_fixes":2017},[2015],{"id":1325,"text":1335,"points":5,"triggers_fixes":2019,"dont_know":261},[2015],[],{"id":2022,"text":1489,"helper":1490,"why":2023,"max_points":1322,"options":2024,"sources":2034},"l9","An AI assistant answers when you ask. A broken form on a Friday can sit unnoticed until someone happens to ask the right question, and meanwhile the ad platforms keep learning from the wrong numbers. Some assistants can run scheduled tasks, but those repeat the same unchecked steps unless someone reviews the result.",[2025,2027,2030,2032],{"id":1311,"text":1494,"points":1322,"triggers_fixes":2026},[],{"id":1316,"text":1497,"points":1397,"triggers_fixes":2028},[2029],"l-weekly-check",{"id":1320,"text":1501,"points":5,"triggers_fixes":2031},[2029],{"id":1325,"text":1335,"points":5,"triggers_fixes":2033,"dont_know":261},[2029],[],{"id":2036,"text":1508,"helper":2037,"why":1510,"max_points":1478,"options":2038,"sources":2048},"l10","Decisions like moving budget, pausing a campaign, telling sales which campaign brings the best leads, or reporting to the owner or a client.",[2039,2041,2044,2046],{"id":1311,"text":1513,"points":1478,"triggers_fixes":2040},[],{"id":1316,"text":1516,"points":1133,"triggers_fixes":2042},[2043],"l-check-before-acting",{"id":1320,"text":1520,"points":274,"triggers_fixes":2045},[2043],{"id":1325,"text":1335,"points":5,"triggers_fixes":2047,"dont_know":261},[2043],[],{"id":2050,"text":2051,"helper":2052,"why":2053,"max_points":1327,"options":2054,"sources":2064},"l11","Can the AI change anything in your ad accounts?","For example: create or edit campaigns, change budgets or switch ads on. Meta's own connector can make changes, and so can some vendor connectors, if you allow it.","Some connectors can write as well as read. Meta's connector can create and edit campaigns, ad sets and ads. New ones start paused and switching them on needs your confirmation, but Meta doesn't describe a separate confirmation step for edits to campaigns that are already running. \"Don't change anything\" in your instructions helps, but it's an instruction to the AI, not a permission.",[2055,2057,2060,2062],{"id":1311,"text":1532,"points":1327,"triggers_fixes":2056},[],{"id":1316,"text":1535,"points":1017,"triggers_fixes":2058},[2059],"l-read-only",{"id":1320,"text":1539,"points":274,"triggers_fixes":2061},[2059],{"id":1325,"text":1335,"points":5,"triggers_fixes":2063,"dont_know":261},[2059],[1544,1583],{"id":1885,"text":1548,"helper":2066,"why":2067,"max_points":1478,"options":2068,"sources":2079},"For example, whether people use a personal or a business plan, whether your chats can be used to train the AI's models, and whether names, emails or phone numbers from your lead list ever go in.","Personal and business plans handle your data differently. On the personal plans of Claude, ChatGPT, Gemini and Copilot there is a setting that decides whether your chats help train the models; on business plans, these providers say they don't train on your business data by default. A lead list is full of personal data under GDPR (names, emails, phone numbers), and an analyst rarely needs any of it. This is not legal advice: your data protection lead or adviser has the final word.",[2069,2072,2075,2077],{"id":1311,"text":2070,"points":1478,"triggers_fixes":2071},"Yes: we've checked our plan and its training setting, and we share counts, not people's details",[],{"id":1316,"text":1556,"points":1133,"triggers_fixes":2073},[2074],"l-data-basics",{"id":1320,"text":1560,"points":274,"triggers_fixes":2076},[2074],{"id":1325,"text":1335,"points":5,"triggers_fixes":2078,"dont_know":261},[2074],[1565,1566,1567,1568,1569],[2081,2088,2097,2111,2129,2143,2149,2157,2169,2184,2196,2206,2215],{"id":1914,"title":1572,"impact":1322,"effort":1573,"explanation":2082,"diy_steps":2083,"for_implementer":1580,"sources":2085,"triggered_by":2086},"Not knowing how your AI gets its numbers usually means the setup grew one chat at a time, so nobody has the full picture. Before fixing anything, write one page: which assistants and plans people use, which platforms and accounts they're connected to, who connected them, what the AI is allowed to change, where your lead list comes from, and where your definitions live. Most of the other fixes take minutes once this page exists.",[1576,1577,2084,1579],"Note where instructions and definitions live, if anywhere, and which lead export (form tool, inbox, booking system or CRM) people paste in.",[1341,1582,1583],[2087,1586],"l1.f",{"id":1903,"title":1588,"impact":1322,"effort":1589,"explanation":1590,"diy_steps":2089,"for_implementer":2091,"sources":2092,"triggered_by":2094},[1592,2090,1594,1595],"If you stay with files, export the full report (not only the rows on screen), and put the date range in the file name.","Official routes as of October 2026: Meta's Ads AI Connectors (open beta since 29 April 2026; read and write); Google's GA4 and Google Ads MCP servers are self-hosted developer setups (Python, a Google Cloud project, OAuth) and read-only. Third-party vendor connectors cover most platforms without code. Reading Meta lead form submissions needs a separate permission (leads_retrieval) that isn't on Meta's connector permission list. Large exports can be too big for the assistant to read in full.",[1340,1598,1599,2093],"meta-mcp-start",[2095,2096],"l1.a","l1.b",{"id":1929,"title":2098,"impact":1372,"effort":1605,"explanation":2099,"diy_steps":2100,"for_implementer":2105,"sources":2106,"triggered_by":2107},"Stop adding up the leads each platform reports","Google Ads only credits Google ads; Meta only credits Meta ads. Someone who clicked both can be a lead in each, so adding them together over-counts. For cross-channel questions, use your own lead list as the total, and have the AI treat each platform's figure as that platform's claim, not a share of the total.",[2101,2102,2103,2104],"Add this to your instructions: \"Never add leads or conversions from different ad platforms together. Our lead list is the true total.\"","For cost per lead across channels, divide total ad spend by real leads from your list (junk and duplicates removed), not by the leads the platforms claim.","When the AI compares platforms, ask it to list each platform's own rules (attribution window, date logic, counting setting) next to its number.","Read our explainer on why the platforms never agree, then decide which number answers which question.","Google Ads' standard conversion columns put each conversion on the date of the ad interaction that led to it (usually the click), while other tools use the date the conversion happened (Google Ads Help); add \"Conversions (by conv. time)\" columns to line Google Ads up with Google Analytics, and allow 24-48 hours before comparing. Meta said in March 2026 that differences in what counts as a click \"can lead to inconsistency\" between Ads Manager and third-party tools. Within one platform, check that lead conversion actions count \"One\" per click, and that a lead isn't sent by both a pixel and a server-side event without a shared event ID.",[1364,1614,1366,1966,1615],[2108,2109,2110],"l2.c","l2.d","l2.e",{"id":1946,"title":2112,"impact":1372,"effort":2113,"explanation":2114,"diy_steps":2115,"link":2120,"for_implementer":2123,"sources":2124,"triggered_by":2125},"Give the AI your real lead list, without the personal details","30-60 minutes the first time, 10 minutes after that","Your own list of enquiries is the reference everything else should be compared with. Export it with only what the analysis needs: date, source or campaign, lead type (form, call, booking) and outcome (junk, real, became a customer). Tell the AI that when your list and a platform disagree, your list is the real count and the gap is something to explain.",[2116,2117,2118,2119],"Export 12 weeks of leads from your form tool, inbox, booking system or CRM, with date, source or campaign, type and outcome. Leave out names, emails and phone numbers.","Mark junk and duplicates. Only your team can do this part well.","Add to your instructions: \"Our lead list is the true count. Report Google Ads, Meta and Google Analytics figures as gaps against it.\"","Ask: \"For each week, show real leads, leads reported by Google Ads and by Meta, and the gap.\" Then: \"Which campaigns brought leads that became customers?\"",{"type":1646,"label":2121,"url":2122},"Step by step, for Claude or ChatGPT: Connect Your Lead List to Claude or ChatGPT","https:\u002F\u002Ftrackian.com\u002Fresources\u002Flead-list-to-claude-chatgpt","If leads don't carry their source, fix that first: save UTM parameters and the Google click ID (gclid) in hidden form fields so each lead keeps its campaign. Use the same time zone on both sides. Keep the export free of personal data, or anonymise it before upload.",[],[2126,2127,2128],"l3.b","l3.c","l3.d",{"id":1960,"title":1636,"impact":1372,"effort":1637,"explanation":2130,"diy_steps":2131,"link":2136,"for_implementer":2137,"sources":2138,"triggered_by":2139},"A one-page definitions file stops the AI guessing. It says what counts as a lead (and what's junk), whether calls and bookings count, how each platform counts and dates a lead, and each account's time zone. Put it where the AI reads it on every question, not in message three of an old chat.",[2132,2133,2134,1643,2135],"Note each account's time zone: Google Analytics property settings, Google Ads account settings, your Meta ad account settings and your form or booking tool.","Note how each platform counts. In Google Ads, check each lead conversion action's counting setting (\"One\" or \"Every\") and its window (30 days after a click unless someone changed it).","Write down what counts as a lead for your business, what counts as junk, and whether two enquiries from one person are one lead or two.","Our lead-gen stack guide includes a ready-made definitions template you can fill in.",{"type":1646,"label":1647,"url":1648},"Google recommends \"One\" conversion per click for leads and \"Every\" for purchases. Google Ads time zone is permanently set when the account is created. The default click-through conversion window is 30 days (1-90 days, depending on the conversion source). Confirm Meta's attribution setting per ad set in Ads Manager rather than assuming a default.",[1966,1365,1406,1651],[2140,2141,2142],"l4.b","l4.c","l4.d",{"id":1976,"title":1657,"impact":1372,"effort":1658,"explanation":1659,"diy_steps":2144,"for_implementer":1665,"sources":2145,"triggered_by":2146},[1661,1662,1663,1664],[1427,1428],[2147,2148],"l5.c","l5.d",{"id":1987,"title":1671,"impact":1327,"effort":1658,"explanation":1672,"diy_steps":2150,"for_implementer":2151,"sources":2152,"triggered_by":2153},[1674,1675,1676],"The GA4 Data API returns up to 10,000 rows per request by default and pages beyond that. In Claude Code, MCP tool output above 10,000 tokens triggers a warning, with a default cap of 25,000. Large Google Ads MCP results can be truncated by the assistant's context window. Ask for aggregated results (totals by week or campaign) rather than raw rows.",[1447,1448,1449,1679],[2154,2155,2156],"l6.b","l6.c","l6.d",{"id":1131,"title":1685,"impact":1372,"effort":1686,"explanation":2158,"diy_steps":2159,"for_implementer":2162,"sources":2163,"triggered_by":2164,"link":2168},"Ask the AI five questions about a closed date range and check each answer in the platform itself: a total, a breakdown, your leads, a top-10 list and one calculation. If all five match, the AI reads that platform correctly. If two or more don't, don't use its answers for decisions until you've fixed the connection or your instructions.",[1689,2160,1691,2161,1693],"Ask for a headline total, a breakdown by channel or campaign, your leads (form submissions or key events), a top-10 list, and one ratio with \"show your working\".","Use the step-by-step version for your setup, linked below: chapter 5 of the connection guide for your platform, the 20-minute cross-source test in the lead-gen stack guide if you combine several, or the 30-second AI answer check if you use another assistant.","The test checks the AI's reading and maths, not the tracking. Passing proves the AI matches the platform's own interface; it doesn't prove the platform recorded every enquiry. For that, see \"Check the lead tracking itself, not just the AI\".",[],[2165,2166,2167],"l7.b","l7.c","l7.d",{"type":1646,"label":1701,"url_ref":1702},{"id":2015,"title":2170,"impact":1308,"effort":2171,"explanation":2172,"diy_steps":2173,"for_implementer":2178,"sources":2179,"triggered_by":2180},"Check the lead tracking itself, not just the AI","About 30 minutes, with the free Missing Leads sheet","Send a test enquiry through every form, and compare 12 weeks of the leads Google Ads and Google Analytics report with the leads you actually received. This tells you whether the data the AI reads is close to what really happened. A few missing is normal; a big or sudden gap usually means something broke.",[2174,2175,2176,2177],"Send a test enquiry through each form (and a test booking, if you take them) and check each appears once in your inbox, in Google Analytics and in your ad platforms' conversion reports.","Paste 12 weeks of reported leads and real leads into the free Missing Leads sheet. It shows the gap each week and what the uncounted leads are worth.","Look for weeks where the gap jumped, and check what changed on the site or forms that week.","Repeat after every website, form plugin or cookie banner change.","Use a range that ended at least 3 days ago (GA4 processing can take 24-48 hours). Check that each form fires its lead event once, on a successful submission only, not on page load or a failed validation. If you'd rather a specialist did it, the 24-hour Missing Leads Check compares your tracking with the leads you actually received.",[1679],[2181,2182,2183],"l8.b","l8.c","l8.d",{"id":2029,"title":2185,"impact":1327,"effort":1720,"explanation":2186,"diy_steps":2187,"for_implementer":2190,"sources":2191,"triggered_by":2192},"Put a 15-minute lead numbers check in the diary every week","Pick a fixed time each week. Compare last week's real leads with what Google Ads, Meta and Google Analytics report, note anything unusual, and write down any site, form or campaign changes. It's boring, which is exactly why it catches things.",[1723,2188,2189,1726],"Compare last week's real leads (junk removed) with each platform's reported leads. Note the gaps.","If a gap moved much more than usual, check the forms and tracking before you change any budget.","Scheduled AI tasks (some assistants offer them on paid plans) repeat the same chat-based steps. Someone still has to review the output and know what normal looks like for this business. Keep a simple log of site, form and campaign changes next to the weekly numbers.",[],[2193,2194,2195],"l9.b","l9.c","l9.d",{"id":2043,"title":1734,"impact":1327,"effort":1735,"explanation":2197,"diy_steps":2198,"for_implementer":1742,"sources":2201,"triggered_by":2202},"Agree that any AI number that moves budget, pauses a campaign, changes what sales hears about lead quality or goes into a report is checked first, by someone who knows the platforms. It costs a minute per decision and stops a wrong answer from travelling.",[1738,2199,1740,2200],"For each decision, check the one or two numbers it rests on in the platform itself and in your lead list.","If people outside the team (the owner, an agency, a client, sales) receive AI-generated numbers, tell them which ones were checked.",[],[2203,2204,2205],"l10.b","l10.c","l10.d",{"id":2059,"title":1749,"impact":1322,"effort":1622,"explanation":1750,"diy_steps":2207,"for_implementer":1756,"sources":2210,"triggered_by":2211},[2208,1754,2209,1755],"List which connectors can make changes. Meta's connector can; some vendor connectors can too. Check each one's description.","Where a vendor offers a read-only option, use it for analysis.",[1544,1582,1583,1758],[2212,2213,2214],"l11.b","l11.c","l11.d",{"id":2074,"title":2216,"impact":1478,"effort":1765,"explanation":2217,"diy_steps":2218,"for_implementer":2220,"sources":2221,"triggered_by":2222},"Check your plan, and share counts, not people","Find out which plan each person uses and what it does with your chats, and keep people's details out. A lead list is personal data, and analysis almost never needs names, emails or phone numbers: counts by week, source, campaign and outcome do the job. This isn't legal advice; your data protection lead or adviser has the final word.",[1768,1769,2219,1771],"Export leads without names, emails and phone numbers. With connectors, ask for counts, not lists of people.","Claude personal plans: if you allow training, Anthropic may keep de-identified data for up to five years; with training off, deleted chats leave Anthropic's systems within 30 days. Gemini with a personal account: with Keep Activity on, a subset of chats is reviewed by people, and Google advises not to enter confidential data. Lead form data in ad platforms is personal data too; keep it out of chats unless the question truly needs it.",[1565,1774,1566,1567,1775,1568,1776,1569],[2223,2224,2225],"l12.b","l12.c","l12.d",[1131,2015,2029],[2228],{"fix":1914,"when":1784,"rank":1785},[2230,2232,2234,2236],{"id":1788,"min":274,"max":1789,"name":1790,"headline":1791,"body":2231},"AI assistants are genuinely useful for marketing questions. Based on your answers, though, yours is working from lead data that hasn't been joined or checked, so its answers can be confidently wrong without anyone noticing. That's fixable, and the first steps are free: test whether it reads your data correctly, then check the lead tracking underneath.",{"id":1794,"min":924,"max":1795,"name":1796,"headline":1797,"body":2233},"Parts of your setup are sound. The gaps are the ones that produce plausible wrong numbers: platforms added together, definitions the AI has to guess, junk counted as leads, or forms nobody has tested lately. Fix the top three below, then build the setup properly, with every source connected, your lead list as the reference and a test you can repeat.",{"id":1800,"min":1801,"max":1802,"name":1803,"headline":1804,"body":2235},"You check more than most. The remaining gaps tend to show up when several sources are combined, or when a form or tag changes quietly between checks. Close them with the fixes below and the cross-source rules in the lead-gen stack guide.",{"id":1807,"min":1808,"max":1113,"name":1809,"headline":1810,"body":2237},"Your answers describe a setup where the AI gets complete data, calculates in code and gets checked against your real leads. The work now is doing that every week: connectors change, plans change, forms and tracking drift, and someone has to notice.",{"heading":1813,"body":2239},"Every fix above can be done without Trackian, and an AI assistant is a real help once it's set up well. The hard part is keeping it that way. Each connector brings one platform with its own rules, the AI has to be told how they differ, someone has to check its answers against the platforms and your lead list, and someone has to check the forms and tracking underneath. Every week, because connectors, plans and tracking change without telling you.",{"heading":1816,"body":2241},[2242,1819,1820,1821],"Trackian is a marketing brain for your business. It joins your ads, site and lead data through 40+ integrations, including Google Ads, Meta Ads, Microsoft Ads, Google Analytics (GA4), Google Search Console, and form and booking tools such as Typeform and Calendly. Your CRM connects through API, webhooks or MCP, so you can see which leads turned into deals - not just which forms were filled in.",[2244],{"id":1824,"order":5,"type":1825,"heading":1826,"label":2245,"subtext":2246,"own_numbers_line":2247,"note":1830,"url":2248,"show_when":1832},"Get my free Missing Leads Check","A Trackian specialist compares what your tracking recorded with the leads you actually received and tells you within 24 hours what's missing, why, and what to fix first. Free - we ask a few short questions first to make sure it's a fit.","You said you spend {spend_band_label} a month on ads and get {volume_band_label} leads a month. Your AI's answers about those leads are only as good as the tracking underneath - the check compares it with the leads you actually received.","https:\u002F\u002Ftrackian.com\u002Fmissing-leads-check",[2250],{"id":1835,"type":1646,"label":2251,"subtext":2252,"url":2253,"placement":1839},"Get the free Missing Leads sheet","Paste 12 weeks of reported leads next to the leads you really received and see the gap each week. The quickest way to check the data your AI reads.","https:\u002F\u002Ftrackian.com\u002Fresources\u002Fmissing-orders-leads-sheet\u002Fleads",{"b1":2255,"b2":2256,"b3":2257,"b4":2258},{"type":1646,"label":1842,"url":1843},{"type":1646,"label":1845,"url_ref":1846},{"type":1646,"label":1848,"url_ref":1846},{"type":1646,"label":1850,"url":2259},"https:\u002F\u002Ftrackian.com\u002Fresources\u002Fowners-weekly-numbers\u002Fleadgen",{"track_name":2261,"other_track_name":2262,"note":2263},"The visitor's track's name (tracks[track].name), e.g. \"Can you trust your AI's answers about your shop?\". Used on the results-PDF cover date line: \"Can you trust your AI's answers about your shop? · taken on 11 October 2026\".","The other track's short_name (tracks[other].short_name), e.g. \"lead-gen scorecard\" in result_labels.other_track: \"Take the lead-gen scorecard instead\".","Defined 11 Oct 2026 (owner-approved tool fixes, _tools\u002FCOPY-NOTES.md T14), as session 5 builds them.",{"score_label":2265,"score_format":2266,"self_report_note":2267,"next_step_heading":2268,"top_fixes_heading":2269,"top_fixes_intro":2270,"fix_effort_label":2271,"fix_steps_label":2272,"fix_implementer_label":2273,"why_label":2274,"retake":2275,"other_track":2276,"step_one_line":2277},"Your AI trust score","{score}\u002F100","This score reflects your answers, not a technical check of your AI setup or your tracking. If you answered \"I don't know\" a lot, the real picture could be better or worse.","Your next step","Your 3 most important fixes","Ranked by how likely each one is to put a wrong number in front of you with a confident explanation attached. All of them can be done without Trackian.","Time needed","What to do","For whoever implements this","Why this matters","Retake the scorecard","Take the {other_track_name} instead","You've done step 1 of 3: your AI trust score and your 3 fixes.",{"position":2279,"delivery":2280,"headline":2281,"body":2282,"field_label":2283,"field_placeholder":2284,"work_email_note":2285,"first_name_label":2286,"site_url_label":2287,"site_url_placeholder":2288,"site_url_note":2289,"phone_label":2290,"site_url_error":2291,"phone_error":2292,"privacy_line":2293,"button":2294,"success":2295,"error_invalid":2296,"skip_link":2297,"consent_note":2298},"after_score","download","Get my results PDF: my AI trust score and the 3 fixes","Your score, your answers, the step-by-step fixes and your next step, as a PDF you can forward to your team, developer or agency. Fill in four quick fields and the download starts straight away.","Work email","you@yourcompany.com","A work email helps us match your results to your company. Any address gets the PDF.","First name","Your website","yourcompany.com","Filled in from your work email. Agencies: type your client's site.","Phone (optional)","That website address doesn't look right. Please check it and try again.","That phone number doesn't look right. Please check it and try again.","We use your email to give you your results and to follow up about them. Marketing emails only if you ask for them. Privacy policy: https:\u002F\u002Ftrackian.com\u002Fprivacy-center\u002Fprivacy-policy\u002F","Get my results PDF","Show the thank-you route (thank_you_routes) with the download link.","That email address doesn't look right. Please check it and try again.","No thanks, I've seen enough","The results PDF is a requested service and is given without marketing consent (Defaults #19). There's no marketing box on the form: the opt-in is offered on the thank-you view (LAUNCH-BLOCKS 9.5, owner 11 Oct 2026).",{"filename":2300,"sections":2301,"delivery":2372},"trackian-ai-trust-score-{date}.pdf",[2302,2307,2314,2318,2333,2343,2347,2355,2359,2362,2365,2370],{"id":2303,"title":2265,"score_line":2304,"subtitle":2305,"date_line":2306},"cover","{score}\u002F100 · {band_name}","{band_headline}","{track_name} · {ai_tool_name} · taken on {date}",{"id":2308,"heading":2309,"body":2310},"what_this_is","What this score is, and isn't",[2311,2312,2313],"This score adds up 12 answers about how your AI assistant gets your numbers, combines them, calculates with them and gets checked. The higher it is, the more complete and better-checked the data behind its answers.","AI assistants are genuinely useful for marketing questions: they're fast, they explain well and they never get tired of \"why?\". The risk isn't the AI itself. It's confident answers built on data that hasn't been joined or checked. In a US survey of marketers from February 2026, 47.1% said they run into AI inaccuracies several times a week.","This is based on your answers, not a technical check of your AI setup, your accounts or your tracking. If you answered \"I don't know\" several times, the first fix below is the one to start with.",{"id":2315,"heading":2316,"body":2317},"band","{band_name}","{band_body}",{"id":2319,"heading":2268,"by_band":2320},"next_step",{"b1":2321,"b2":2324,"b3":2327,"b4":2330},{"body":2322,"link":2323},"Start with the trust test in your top 3 fixes: five questions to your AI about a closed date range, each checked in the platform. Your version: \"{route.trust_test.guide}\", {route.trust_test.section}. Then read the nine ways AI gets marketing numbers wrong, so you can catch each one in 30 seconds.","next_step_by_band.b1",{"body":2325,"link":2326},"Fix the three items below, then build your setup properly with the free stack guide \"{route.stack_guide.guide}\": every source connected, a rules file your AI reads every time, 15 prompts labelled by how far to trust them, and a 20-minute cross-source trust test.","next_step_by_band.b2",{"body":2328,"link":2329},"Close the gaps below, then use the cross-source rules and the 20-minute trust test in the free stack guide \"{route.stack_guide.guide}\" to keep combined answers honest.","next_step_by_band.b3",{"body":2331,"link":2332},"You've built by hand what most setups lack: complete data, calculations in code and regular checks. Keep it that way with a 10-minute Monday routine, the owner's weekly numbers, which has a tracking check built in.","next_step_by_band.b4",{"id":2334,"heading":2335,"intro":2336,"columns":2337,"row_rule":2341,"full_points_text":2342},"answers","Your answers","Every question, your answer, and why it matters.",[2338,2339,2340],"Question","Your answer","Why it matters","List all 12 scored questions in order. Show the question's why text in the third column, except for full-points answers, which show full_points_text.","Nothing to fix here. Keep it this way.",{"id":2344,"heading":2269,"intro":2345,"per_fix":2346},"fixes","Ranked by how likely each one is to put a wrong number in front of you with a confident explanation attached. Each one can be done without Trackian. Forward this page to whoever manages your AI setup, website or ads.","title, explanation, Time needed, What to do (numbered), For whoever implements this (boxed)",{"id":2348,"heading":2349,"by_track":2350},"first_check","Check the data underneath, too",{"shop":2351,"leadgen":2353},{"body":2352,"link_label":1836,"url":1838},"Your AI can only repeat what your tracking recorded. The free Missing Orders sheet takes about 30 minutes: paste 30 days of daily orders and revenue from your shop and from Google Analytics, and it shows the gap in orders and in euros, flags the days that look wrong, and tells you whether the gap is within normal range, worth checking or serious.",{"body":2354,"link_label":2251,"url":2253},"Your AI can only repeat what your tracking recorded. The free Missing Leads sheet takes about 30 minutes: paste 12 weeks of the leads Google Ads and Google Analytics report, next to the leads you actually received and how many were junk, and it shows the gap each week and what the uncounted leads are worth.",{"id":2356,"heading":2357,"body":2358},"habit","Keep it honest: a 15-minute weekly check","Use the steps of the weekly check fix for the track (s-weekly-check or l-weekly-check), even if it wasn't in your top 3.",{"id":2360,"use":2361},"diy_reality","track.diy_reality",{"id":2363,"use":2364},"pitch","track.pitch",{"id":2366,"heading":2367,"body":2368,"cta_order":2369},"next_steps","Your next steps","You've done step 1 of 3: your AI trust score and your 3 fixes. Step 2: tracks[].next_step_by_band[band] (label and URL as plain text). Step 3: the check block (ctas.check_24h: heading, subtext, own_numbers_line and URL as plain text).",[2319,1824],{"id":341,"text":2371},"Based on your answers on {date}. This is not a technical audit of your AI setup, website or ad accounts, and not legal advice. Made by Trackian, a marketing brain for your business. Edition: October 2026. https:\u002F\u002Ftrackian.com · Privacy policy: https:\u002F\u002Ftrackian.com\u002Fprivacy-center\u002Fprivacy-policy\u002F","Instant download on the thank-you screen after the email is entered. No email is sent.",[2374,2377,2380,2383,2386,2389,2392,2395,2398,2401,2404,2407,2410,2413,2416,2419,2422,2425,2428,2431,2434,2437,2440,2443,2446,2449,2452,2455,2458,2461,2464,2467,2470,2473,2476,2479],{"id":1364,"title":2375,"url":2376},"Google Ads Help - Conversion discrepancies with other tools (by conv. time columns; allow 24-48 hours; this page words the reporting date as \"ad impression date\", see ads-conv-time for the click-date rule)","https:\u002F\u002Fsupport.google.com\u002Fgoogle-ads\u002Fanswer\u002F7457111",{"id":1614,"title":2378,"url":2379},"Google Ads Help - Understand conversions by time (standard columns report conversions by the date the ad was clicked; \"by conv. time\" columns use the conversion date)","https:\u002F\u002Fsupport.google.com\u002Fgoogle-ads\u002Fanswer\u002F9549009",{"id":1406,"title":2381,"url":2382},"Google Ads Help - About conversion windows (30-day default, 1-90 days)","https:\u002F\u002Fsupport.google.com\u002Fgoogle-ads\u002Fanswer\u002F3123169",{"id":1966,"title":2384,"url":2385},"Google Ads Help - About conversion counting options (One for leads, Every for purchases)","https:\u002F\u002Fsupport.google.com\u002Fgoogle-ads\u002Fanswer\u002F3438531",{"id":1365,"title":2387,"url":2388},"Google Ads Help - Time zone and currency are permanently set at account setup","https:\u002F\u002Fsupport.google.com\u002Fgoogle-ads\u002Fanswer\u002F9842104",{"id":1407,"title":2390,"url":2391},"Google Analytics Help - Currency conversion using the prior day's exchange rate","https:\u002F\u002Fsupport.google.com\u002Fanalytics\u002Fanswer\u002F9796179",{"id":1679,"title":2393,"url":2394},"Google Analytics Help - Data freshness (processing can take 24-48 hours)","https:\u002F\u002Fsupport.google.com\u002Fanalytics\u002Fanswer\u002F11198161",{"id":1447,"title":2396,"url":2397},"Google Analytics Data API - runReport (row limits and paging)","https:\u002F\u002Fdevelopers.google.com\u002Fanalytics\u002Fdevguides\u002Freporting\u002Fdata\u002Fv1\u002Frest\u002Fv1beta\u002Fproperties\u002FrunReport",{"id":1598,"title":2399,"url":2400},"Google Analytics MCP server (GitHub, official, read-only developer setup)","https:\u002F\u002Fgithub.com\u002Fgoogleanalytics\u002Fgoogle-analytics-mcp",{"id":1599,"title":2402,"url":2403},"Google Ads MCP server (GitHub, official, read-only developer setup)","https:\u002F\u002Fgithub.com\u002Fgoogleads\u002Fgoogle-ads-mcp",{"id":1449,"title":2405,"url":2406},"The Ad Spend - ChatGPT and Google Ads: what changed (large MCP results truncated by the context window; connectors renamed apps) - reported","https:\u002F\u002Ftheadspend.com\u002Fblog\u002Fchatgpt-google-ads-what-changed",{"id":1448,"title":2408,"url":2409},"Claude Code docs - MCP (tool output warning above 10,000 tokens, default cap 25,000)","https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fmcp",{"id":1427,"title":2411,"url":2412},"Anthropic Help - Create and edit files with Claude (code execution for data analysis; Settings > Capabilities; all plans)","https:\u002F\u002Fsupport.claude.com\u002Fen\u002Farticles\u002F12111783-create-and-edit-files-with-claude",{"id":1582,"title":2414,"url":2415},"Anthropic Help - Use connectors to extend Claude's capabilities (Always allow, Needs approval, Blocked on Team and Enterprise)","https:\u002F\u002Fsupport.claude.com\u002Fen\u002Farticles\u002F11176164-use-connectors-to-extend-claude-s-capabilities",{"id":1428,"title":2417,"url":2418},"OpenAI Help - Data analysis with ChatGPT (writes and runs Python for some tasks)","https:\u002F\u002Fhelp.openai.com\u002Fen\u002Farticles\u002F8437071-data-analysis-with-chatgpt",{"id":1651,"title":2420,"url":2421},"OpenAI Help - Projects in ChatGPT (project instructions and files)","https:\u002F\u002Fhelp.openai.com\u002Fen\u002Farticles\u002F10169521-projects-in-chatgpt",{"id":1341,"title":2423,"url":2424},"OpenAI Help - Troubleshooting apps in ChatGPT (Settings > Apps)","https:\u002F\u002Fhelp.openai.com\u002Fen\u002Farticles\u002F20001497-troubleshooting-apps-in-chatgpt",{"id":1566,"title":2426,"url":2427},"OpenAI Help - Opting out of model training (\"Improve the model for everyone\"; business data not trained on by default)","https:\u002F\u002Fhelp.openai.com\u002Fen\u002Farticles\u002F5722486-how-do-i-opt-out-of-my-content-being-used-to-improve-model-performance",{"id":1565,"title":2429,"url":2430},"Anthropic - Updates to Consumer Terms and Privacy Policy (model training choice on Free, Pro, Max; five-year retention if allowed; commercial terms excluded)","https:\u002F\u002Fwww.anthropic.com\u002Fnews\u002Fupdates-to-our-consumer-terms",{"id":1774,"title":2432,"url":2433},"Anthropic Privacy Center - Data relationship for commercial products","https:\u002F\u002Fprivacy.claude.com\u002Fen\u002Farticles\u002F9267385-what-is-the-data-relationship-between-anthropic-the-customer-and-the-user",{"id":1567,"title":2435,"url":2436},"Google - Gemini Apps Privacy Hub (Keep Activity; human review; don't enter confidential data)","https:\u002F\u002Fsupport.google.com\u002Fgemini\u002Fanswer\u002F13594961",{"id":1775,"title":2438,"url":2439},"Google Workspace - Generative AI in Google Workspace Privacy Hub (no training on customer data without permission)","https:\u002F\u002Fknowledge.workspace.google.com\u002Fadmin\u002Fgenerative-ai\u002Fgenerative-ai-in-google-workspace-privacy-hub?hl=en",{"id":1568,"title":2441,"url":2442},"Microsoft Support - Microsoft Copilot privacy controls (Training on conversation activity; not for work or school accounts)","https:\u002F\u002Fsupport.microsoft.com\u002Fen-us\u002Fmicrosoft-copilot\u002Fmicrosoft-copilot-privacy-controls",{"id":1776,"title":2444,"url":2445},"Microsoft Learn - Enterprise data protection in Microsoft 365 Copilot (prompts and responses not used to train foundation models)","https:\u002F\u002Flearn.microsoft.com\u002Fen-us\u002Fcopilot\u002Fmicrosoft-365\u002Fenterprise-data-protection",{"id":1544,"title":2447,"url":2448},"Meta for Developers - Ads MCP server: ad creation and management tools (created paused; confirmation before activation)","https:\u002F\u002Fdevelopers.facebook.com\u002Fdocumentation\u002Fads-commerce\u002Fads-ai-connectors\u002Fads-mcp-server\u002Fads-mcp-server-tools-ad-creation-and-management",{"id":2093,"title":2450,"url":2451},"Meta for Developers - Ads MCP server: Get started (token permissions list)","https:\u002F\u002Fdevelopers.facebook.com\u002Fdocumentation\u002Fads-commerce\u002Fads-ai-connectors\u002Fads-mcp-server\u002Fads-mcp-server-get-started",{"id":1340,"title":2453,"url":2454},"PPC Land - Meta opens its ad system to Claude and ChatGPT with new AI connectors (open beta, 29 April 2026) - reported","https:\u002F\u002Fppc.land\u002Fmeta-opens-its-ad-system-to-claude-and-chatgpt-with-new-ai-connectors\u002F?year=2026",{"id":1583,"title":2456,"url":2457},"Jon Loomer - Meta Ads AI Connectors and Claude (connector inherits the profile's access) - reported","https:\u002F\u002Fwww.jonloomer.com\u002Fmeta-ads-ai-connectors-claude\u002F",{"id":1758,"title":2459,"url":2460},"Meta Business Help Center - About the learning phase","https:\u002F\u002Fwww.facebook.com\u002Fbusiness\u002Fhelp\u002F112167992830700",{"id":1366,"title":2462,"url":2463},"Meta, \"Simplifying Ad Measurement for a Social-First World\" (3 March 2026), as quoted by Campaign Middle East","https:\u002F\u002Fcampaignme.com\u002Fmeta-updates-ad-measurement-for-a-social-first-world\u002F",{"id":1615,"title":2465,"url":2466},"Meta for Developers - Deduplicate pixel and server events","https:\u002F\u002Fdevelopers.facebook.com\u002Fdocs\u002Fmarketing-api\u002Fconversions-api\u002Fdeduplicate-pixel-and-server-events\u002F",{"id":1339,"title":2468,"url":2469},"Shopify Help Center - Authorizing access (read-only or write; lowering access needs reinstall; one store at a time)","https:\u002F\u002Fhelp.shopify.com\u002Fen\u002Fmanual\u002Fai-powered-tools\u002Fconnecting-ai-tools\u002Fauthorizing-access",{"id":1545,"title":2471,"url":2472},"Shopify Help Center - Considerations (tools can update product listings and change prices; actions not reviewed by Shopify; merchant responsibility)","https:\u002F\u002Fhelp.shopify.com\u002Fen\u002Fmanual\u002Fai-powered-tools\u002Fconnecting-ai-tools\u002Fconsiderations",{"id":1546,"title":2474,"url":2475},"Shopify Help Center - Shopify connector for Claude (with approval it can change products, inventory and discount codes; orders and customers are read-only; one store at a time)","https:\u002F\u002Fhelp.shopify.com\u002Fen\u002Fmanual\u002Fai-powered-tools\u002Fconnecting-ai-tools\u002Fshopify-connector-for-claude",{"id":1569,"title":2477,"url":2478},"GDPR (Regulation (EU) 2016\u002F679), including Article 28","https:\u002F\u002Feur-lex.europa.eu\u002Feli\u002Freg\u002F2016\u002F679\u002Foj",{"id":2480,"title":2481,"url":2482},"survey-us","Neil Patel - AI hallucination data study (US marketers, February 2026: 47.1% encounter AI inaccuracies several times a week)","https:\u002F\u002Fneilpatel.com\u002Fblog\u002Fai-hallucination-data-study\u002F",{"shop":2484,"leadgen":2486},{"under_100":2485,"100_499":1267,"500_1999":1270,"2000_plus":1273},"under 100",{"under_20":2487,"20_49":1877,"50_199":1880,"200_plus":1883},"under 20",{"under_2k":2489,"2k_5k":1250,"5k_15k":1253,"15k_plus":1256},"under €2k",[2491,2492,2493,2494,2495,2496,2497,2498,2499],"Your AI trust score: {score}\u002F100 and the band name, with the self-report note.","Band headline and body.","Your 3 most important fixes (the trust-test fix carries the routed trust-test link).","Free sheet link (content_links.missing_sheet).","\"Doing this yourself, every week\" and the Trackian pitch.","Your next step: step 1 of 3 (result_labels.step_one_line), step 2 (tracks[].next_step_by_band[band], a text link), step 3 (the one CTA, ctas.check_24h, with its own_numbers_line).","Forward this to your boss (forward_email, a text link).","Email capture for the results PDF (optional).","Links: \"Retake the scorecard\" and \"Take the {other_track_name} instead\".",[2501,2508,2516,2523,2529,2535,2544,2551],{"id":2502,"type":2503,"placement":2504,"title":2505,"shows":2506,"alt":2507},"V1","PHOTO","intro","Asking the AI, then checking","welcome-screen photo, shown above the headline on the generic intro and on both track intros - an owner at a desk in a small, light European office in the late afternoon; a laptop with an AI chat window, angled away so nothing on it is readable, and beside it a printed report with a pen resting on one line, as if comparing the answer with the source; calm and uncluttered; no assistant branding; nothing that marks it as a shop or a service business, so it suits both tracks; no readable text or numbers, no logos or brand marks; no real or named people; crops to a wide banner at phone width","An owner at a desk compares an AI assistant's answer on a laptop with a printed report, pen in hand.",{"id":2509,"type":2510,"placement":2511,"title":2512,"shows":2513,"caption":2514,"alt":2515},"V2","DIAGRAM","track:shop","How your shop's numbers reach the AI","shown at the start of the shop track, above its first question - left: four sources, Google Analytics, Google Ads, Meta and Your shop admin (the real orders, outlined). Middle: four lanes for the ways numbers reach the assistant: screenshots or copy-paste, exports, connectors (one platform each, with its own rules), one source already joined and checked. Right: AI assistant -> answer. Three small labels along the path: what it receives, what it has been told (your definitions), who does the maths (in its head, in code, or already calculated). The one point: an answer is only as good as what goes in and how it is worked out. Compact; reads top to bottom at phone width","What goes in, what the AI is told and who does the maths decide how far to trust the answer.","Diagram of Google Analytics, Google Ads, Meta and your shop admin reaching an AI assistant by screenshots, exports, connectors or one joined and checked source, with what it receives, what it has been told and who does the maths shaping the answer.",{"id":2517,"type":2510,"placement":2518,"title":2519,"shows":2520,"caption":2521,"alt":2522},"V3","before:s7","Checking the AI, and checking the tracking","shown above question 7 of the shop track - three boxes in a row: AI assistant's answer, The platform's own report (Google Analytics, Google Ads, Meta), Your shop admin (real orders). A short double arrow between the first two, labelled The trust test - does the AI read the platform correctly? A second double arrow between the platform and your shop admin, labelled The tracking check - did the platform record every order? Underneath, two small habits: a weekly check, even when nobody asks; a person checks the number before money moves. The one point: checking the AI and checking the tracking are two different jobs. Compact; the boxes stack top to bottom at phone width","Checking the AI and checking the tracking are two different jobs.","Diagram of two checks: the trust test compares the AI's answer with the platform's own report, and the tracking check compares the platform with your shop's real orders, with a weekly check and a person checking before money moves.",{"id":2524,"type":2510,"placement":2525,"title":2526,"shows":2527,"caption":2514,"alt":2528},"V4","track:leadgen","How your lead numbers reach the AI","shown at the start of the lead-gen track, above its first question - left: four sources, Google Analytics, Google Ads, Meta and Your own lead list (inbox, form tool, booking system or CRM; the real leads, outlined). Middle: four lanes for the ways numbers reach the assistant: screenshots or copy-paste, exports, connectors (one platform each, with its own rules), one source already joined and checked. Right: AI assistant -> answer. Three small labels along the path: what it receives, what it has been told (your definitions), who does the maths (in its head, in code, or already calculated). The one point: an answer is only as good as what goes in and how it is worked out. Compact; reads top to bottom at phone width","Diagram of Google Analytics, Google Ads, Meta and your own lead list reaching an AI assistant by screenshots, exports, connectors or one joined and checked source, with what it receives, what it has been told and who does the maths shaping the answer.",{"id":2530,"type":2510,"placement":2531,"title":2532,"shows":2533,"caption":2521,"alt":2534},"V5","before:l7","Checking the AI, and checking the lead tracking","shown above question 7 of the lead-gen track - three boxes in a row: AI assistant's answer, The platform's own report (Google Analytics, Google Ads, Meta), Your lead list (the enquiries that really arrived). A short double arrow between the first two, labelled The trust test - does the AI read the platform correctly? A second double arrow between the platform and your lead list, labelled The tracking check - did the platform record every real lead? Underneath, two small habits: a weekly check, even when nobody asks; a person checks the number before money moves. The one point: checking the AI and checking the tracking are two different jobs. Compact; the boxes stack top to bottom at phone width","Diagram of two checks: the trust test compares the AI's answer with the platform's own report, and the tracking check compares the platform with the leads you really received, with a weekly check and a person checking before money moves.",{"id":2536,"type":2537,"placement":2538,"dynamic":261,"title":2539,"shows":2540,"data":2541,"caption":2542,"alt":2543},"V6","CHART","result","Your AI trust score and band","sits directly under the score line and above the band headline, on both tracks - a horizontal bar from 0 to 100 split into four bands, each labelled with its name; the visitor's band in the highlight colour, the other three muted; a marker above the bar at the visitor's score, labelled with the score. Static mockup, example: score 64, Mostly checked highlighted. Full width at phone size, band names under the bar; not clickable","dynamic - the visitor's AI trust score 0-100 on the band scale, the same on both tracks: 0-39 Confident guesswork, 40-59 Trusted, not verified, 60-79 Mostly checked, 80-100 Checked and joined; their band highlighted and their score marked; no average, benchmark or comparison","Based on your answers, not a check of your accounts. Your band also sets step 2 below.","Bar from 0 to 100 split into four bands, from Confident guesswork to Checked and joined, with your AI trust score marked and your band highlighted.",{"id":2545,"type":2510,"placement":2538,"dynamic":261,"title":2546,"shows":2547,"data":2548,"caption":2549,"alt":2550},"V7","Your 3 fixes at a glance","sits under the heading Your 3 most important fixes, before the full fixes, on both tracks - a numbered strip of three short cards, 1 -> 2 -> 3, each with the fix title and its time needed; a card filled from the fallback list carries a small Keep it this way tag. Stacked at phone width, a row of three on wider screens. Plain text, not buttons or links; the full steps follow below. The trust-test fix keeps its routed link in the full fix below, never on the card. Static mockup, example: 1 Check the tracking itself, not just the AI - Time needed: About 30 minutes, with the free Missing Orders sheet; 2 Stop adding up the sales each platform reports - Time needed: 10 minutes to change your instructions, then a habit; 3 Give the AI your real orders, and tell it they win - Time needed: 15-30 minutes","dynamic - the same 3 fixes, in the same order, as the fixes list below (scoring.fix_ranking, including the fix Write down how your AI gets its numbers in first place after 3 or more I don't know answers): each fix's title and its effort (fix.effort, labelled Time needed); fixes filled from fallback_fixes tagged Keep it this way","Start with number 1. The full steps for each fix are below.","Your three most important fixes in ranked order, each with its title and the time it needs.",{"id":2552,"type":2537,"placement":2553,"dynamic":261,"title":2554,"shows":2555,"data":2556,"caption":2557,"alt":2558},"V8","pdf","Your points, question by question","results PDF only, beside Your answers - one horizontal bar per scored question, in question order; each bar's full length is that question's maximum and the filled part is the visitor's points; full-points questions in a neutral colour, questions that lost points in the highlight colour; short labels on the left, points such as 6\u002F10 on the right. Static mockup, example values","dynamic - the visitor's points on each scored question against its max_points, with the same labels on both tracks except question 3: How data gets to the AI 10, Combining platforms 8, Your real orders (shop) or Your lead list (lead-gen) 9, Definitions given to the AI 9, Who does the maths 10, Answers say what they're based on 7, Checked against the platforms 10, Tracking itself checked 10, Scheduled checks 8, Who acts on AI answers 6, Write access 7, Data protection basics 6","Where your points went. The answers table explains each one.","Bar chart of your points on each of the 12 questions against the most each could score, with the questions that lost points highlighted.",{"link_label":2560,"delivery":2561,"subject":2562,"body":2563,"fill":2564},"Forward this to your boss","mailto: the visitor sends it from their own inbox; Trackian sends nothing. Fill it only from the visitor's own answers.","Our AI trust score: {score}\u002F100 ({band_name})","Hi {first_name},\n\nI ran our setup through Trackian's free AI trust scorecard (about 3 minutes).\n\nResult: {score}\u002F100, {band_name}.\nThe biggest issue it found: {top_fix_title}.\nWhat it means for us: at our {spend_band_label} monthly ad spend and {volume_band_label} {orders_or_leads} a month, budget moves on what our AI tells us, so its numbers need to match our real {orders_or_leads}.\n\nTrackian offers a free check: a specialist compares what our tracking recorded with our real {orders_or_leads} and tells us within 24 hours what's missing and why. I'd like to do it this month: {check_url}\n\nFull results: https:\u002F\u002Ftrackian.com\u002Fresources\u002Ftrust-your-ai-scorecard\n\n{sender}",{"top_fix_title":2565,"orders_or_leads":2566,"check_url":2567,"spend_band_label":2568,"volume_band_label":2569,"first_name":2570,"sender":2570},"title of the first of the 3 fixes shown","orders (shop) \u002F leads (lead-gen)","ctas.check_24h.url for the track","spend_band_labels[spend_band]","volume_band_labels[track][volume_band]","left blank",{"high":2572,"medium":2582,"low":2586,"note":2588},{"when":2573,"headline":2574,"body":2575,"button":2576,"url":2577,"url_by_segment":2578,"url_rule":2581},"spend_band == '15k_plus' OR erp == 'yes' OR lead_score >= 70","Your results are downloading. Want to go through them with a specialist?","Pick a time that suits you. A Trackian specialist will look at your answers before the call.","Book a call","https:\u002F\u002Ftidycal.com\u002Fteam\u002Ftrackian\u002Fmeet",{"agency":2579,"shop":2580,"leadgen":2580},"https:\u002F\u002Ftidycal.com\u002Fteam\u002Ftrackian\u002Fmeet-agency","https:\u002F\u002Ftidycal.com\u002Fteam\u002Ftrackian\u002Fmeet-brands","role == agency -> url_by_segment.agency; otherwise url_by_segment[track] (shop or lead-gen); if the segment is unknown -> url (generic)",{"when":2583,"headline":2584,"body":2585},"lead_score >= 30","Your results are downloading.","If it didn't start, use the download link below. Next: {next_step_by_band[band].label}: {next_step_by_band[band].url}.",{"when":2587,"headline":2584,"body":2585},"else","erp is not collected in this scorecard, so the high route fires on spend or lead score. The result screen always shows the one CTA; routing only changes the thank-you screen and the follow-up. Lead scoring: lead-scoring.md. {next_step_by_band[band].url} is the step's url, or for b2 and b3 the URL their url_ref resolves to (routing.stack_guide). Medium = nurture (owner, 11 Oct 2026): the Medium route starts at lead score 30 (it was 40) and shows the same thank-you text as Low, with no specialist review; only High offers the call.",{"spec":2590,"scorecard_id":1025},"lead-scoring.md (v2, LAUNCH-BLOCKS s.9)",[],1792069939104]