One Shared Memory for Four AI Agents
736 sessions and 28 shared memories in one versioned vault
Read case studyI audited the tracking for a yoga school, from Google Tag Manager through GA4 to Google Ads. In two days I fixed more than eight issues across 15+ tags, repaired the conversions for the English and German booking pages, and filtered bot traffic out of the analytics data.
A yoga school that runs teacher trainings and retreats abroad, for students mostly from German-speaking countries. It also offers private coaching and online classes.
A teacher training is a big decision made from far away, so almost every student finds the school online, compares programs and dates on the website, then registers. The site runs in English and German, and an external marketing team runs the school’s Google Ads. Both the owner and that team rely on GA4 and Ads data to decide which campaigns and programs get the budget.
The tracking was in place, but nobody could trust the data coming out of it.
These problems feed each other. A tag that misfires in GTM leaves a gap in GA4, and a gap in GA4 or Ads means campaigns are optimized on the wrong numbers.
I followed the data from the tag that fires in the browser to the conversion that arrives in Google Ads, and fixed each break where it started.
Audit of every tag
I went through the 15+ tags in the GTM container: which tags fire, on which triggers, what they send, and whether it arrives at the next layer. I found more than eight issues and fixed the tags, triggers, and connections involved.
Both booking goals repaired
I traced the English and German thank-you pages separately through GA4 events and Google Ads conversion tracking, so each language counts its own registrations and neither is lost or counted twice.
Every status to green
I worked through the conversion actions until none showed “Unverified” or “Needs attention”. A green status alone doesn’t prove the data is right, so I also checked what arrived at each step.
Bot and internal traffic out of GA4
I filtered out traffic from the countries the school doesn’t target, which removed the bot traffic from China, Russia, and the US, and excluded the school’s own visits, so the reports count real prospective students only.
A report and the current setup
I delivered a written report of every issue found and every tag I fixed or set up, together with documentation of the complete current setup, so the owner and the marketing team know exactly what runs and why.
| Layer | Tool | Role |
|---|---|---|
| Website | WordPress, caching and multilingual plugins | Program pages and thank-you pages in two languages |
| Tag management | Google Tag Manager | 15+ tags, with their triggers and variables |
| Analytics | Google Analytics 4, Google Site Kit | Traffic, key events, WordPress dashboard |
| Advertising | Google Ads | One conversion action per language for bidding |
| Testing | GTM Preview, Tag Assistant, GA4 DebugView | Verifying each tag and the data arriving downstream |
The tags fire correctly, both booking goals are counted, GA4 reports on real visitors, and Google Ads receives the conversions it bids on.
| Before | After |
|---|---|
| Broken tags on the English and German thank-you pages | Both goals tracked separately in GA4 and Google Ads |
| Conversion actions marked “Unverified” or “Needs attention” | Every conversion action green |
| Bot traffic from China, Russia, and the US in GA4 | Untargeted countries filtered out |
| The school’s own visits counted as regular traffic | Internal traffic excluded |
| 8+ issues across 15+ tags | Every issue fixed and documented |
The owner was very happy with the result and thankful for the work, and the marketing team said the audit helped them and that the deliverables came remarkably fast.
If you aren’t sure your tracking data can be trusted, hire me.
“Sultan is a genuinely kind person, and just as responsible in his work. He went through every tag behind our English and German bookings, fixed what was broken, and explained it all in a report we could actually follow. He keeps our tracking in great shape and looks after it with real care.”
Tag management, data layers, and Consent Mode V2 that fire once, with consent, and Google Ads can bid on.
736 sessions and 28 shared memories in one versioned vault
Read case studySales documents and the books in one system
Read case studyAbout 30 hacked sites cleaned, locked, and watched daily
Read case studySending stopped on its own, every site on a new key
Read case studyProducts and articles published without a deploy
Read case studyNo health data sent to Google, bookings measured
Read case studyThe company's main hiring platform, built in-house
Read case studyFive chat channels in one inbox on every phone
Read case studyConsent to conversion, end to end, in 2 days
Read case studyOne event standard across every data flow
Read case studyNext step
See how I work and what it costs, or email me. I usually reply within one working day.