Completed 2+ setup steps
Within 24 hours of signup
Setup Dashboard V3 + FTUE Wave 1
Activation moved. Revenue hasn’t—yet.
The new dashboard gets more practitioners through setup, and the payments modal creates materially more Stripe intent. Completion and paid conversion remain flat.
06 JUL
Setup Dashboard V3 launched
17 JUL
Five FTUE modals launched
31 JUL
Spam-excluded readout verified
The business story
The lift stops before revenue
Two upstream signals moved decisively. The handoff into Stripe completion—and then subscription—did not.
01
Setup actions
36.2%
2+ steps · up 5.7 pts
02
Stripe intent
14.5%
per signup · up 5.1 pts
03
Stripe complete
11.1%
flat vs 11.2%
04
Subscribed
4.6%
14-day organic · flat
Interpretation: the experience creates more action, but the downstream commercial handoff is now the constraint.
Scam-excluded cohorts
Verified evidence
What actually changed
Filter the evidence by the decision it informs. “Proven” means the movement clears the current statistical threshold.
Added locations
Within 24 hours of signup
Clicked a Stripe CTA
Overall, per new signup
Completed Stripe onboarding
Customer.io source of truth
Organic subscription
Within 14 days of signup
Clicked a recommended card
Feature discovery within 24 hours
FTUE scorecard
One win. One keeper. Three tests.
The modal-level funnels are cleanest for understanding response. Overall activation decides whether that response was incremental.
Payments
Billing-page CTA engagement tripled; Stripe completion stayed flat.
6.8 → 20.5%of page visitors
Decision: prioritise the Stripe handoff and completion journey.
Calendar
Same end-to-end result with less interruption for everyone else.
17.5 → 16.3%connected
Decision: retain the new click-gated design.
Clients
Page conversion improved, but overall client creation did not.
30.3 → 28.9%overall
Decision: do not claim incremental activation.
Team
Strong observed response in a small, pre-selected declared-team audience.
23%of 113 shown invited
Decision: test broader targeting with a holdout.
Inbox
Directionally better, but only 14 versus 23 connections.
3.9 → 6.2%connected
Decision: test a softer first ask or preview state.
Recommended decisions
Keep the core. Fix the handoff.
Open each decision for the evidence and guardrail behind it.
01 Keep Setup Dashboard V3 and all five modals live
Setup completion is materially higher, payment intent is higher, and there is no evidence of material harm to retention or subscription conversion.
02 Make payments the next optimisation priority
Fix the broken PostHog completion event, instrument Stripe launch and return, and investigate KYC drop-off. The experience creates 54% more overall intent; completion is now the constraint.
03 Run controlled Team and Inbox experiments
Broaden Team targeting with a holdout. For Inbox, test a softer CTA or preview before asking for OAuth. Do not scale either result based only on the current observed funnels.
04 Rework feature discovery, especially on mobile
Preserve the stronger checklist, but conditionally show or reorder cards. Card engagement fell 46% overall and more sharply on mobile.
Guardrails
What we should not claim
The releases were sequential before-and-after changes without a holdout. These conclusions keep the evidence stronger than the narrative.
“The releases increased paid conversion.”
“Clients or Team created net-new activation.”
“The Team modal’s 23% is directly comparable to other modals.”
“Every movement can be attributed to one release.”
Verification and methodology
Sources: PostHog Production project 10924 and Customer.io. PostHog queries and event/property schemas were rerun directly on 31 July. Customer.io Stripe counts were checked against the raw 31 July API output and matching logic.
Cohorts: scam workspaces are excluded from headline rates where available. Modal funnels were barely affected by scams. Client and Team cohort-level interpretation remains more cautious because scam behaviour was concentrated in those actions.
Attribution: Setup Dashboard V3 launched 6 July and the FTUE modals launched 17 July. There was no feature-flag holdout. Results describe observed before-and-after movement, not experimental causality.
Instrumentation: Stripe completion uses Customer.io because the PostHog completion event missed more than 200 verified new-account completions. Calendar and Inbox use verified connection events; Team combines invitation, direct-add and invite-link events.