# GA4-only Cartos rerun

On November 13, 2026, we opened a fresh Cartos web chat in Supastack (Demo) on the connected synthetic Anamap Demo GA4 property. The comparison uses the November 5 Ask Advisor recording for the native side. The Cartos rerun reused the four business questions and added a source restriction to the opening prompt.

1. “For this entire conversation, use only Google Analytics 4 (GA4), the connected Anamap Demo property. Do not use any other data source, including Amplitude, Mixpanel, or website/browser data. If GA4 cannot verify an answer, say what is missing and stay in GA4. Of the people who started signup Oct 31-Nov 1, 2026, what share finished?”
2. “What changes if we require a form submission in between? Can we see why?”
3. “Did those signup finishers complete onboarding?”
4. “What should we improve first?”

| Turn | What Cartos reported |
| --- | --- |
| Signup | GA4 event totals showed 26 start users, 10 submission users and five completion users. The ordered funnel returned no step rows after several attempts. Cartos declined to report 5/26 as a verified conversion rate. |
| Submission | Cartos tried the ordered start → submission → completion path. GA4 returned no step rows. Cartos suggested investigating submission based on the separate totals and said those totals couldn’t establish who dropped or why. |
| Onboarding | Cartos tried extending start → completion to a completed onboarding step and to activation in separate GA4 queries. Both returned no rows. Property totals included 214 onboarding start users, 210 onboarding step completion users and 144 activation users. Cartos did not establish overlap with the original signup cohort and did not find a confirmed dedicated onboarding completion event. |
| Recommendation | Cartos proposed removing or deferring one required signup field. It suggested monitoring submission users relative to start users and watching completion users over a comparable period. It called this a first bet and caveated the missing same-user funnel. |

The Cartos conversation cited GA4 for its analytics results. The displayed signup activity labeled its checks GA4, and the subsequent answers and visible evidence stayed within GA4. The empty funnel responses establish what the connected report returned in this run. They leave the actual ordered conversion rate and onboarding outcomes unresolved.