Userpilot

I tested Userpilot by installing the JavaScript snippet on a SaaS web app, building a 4-step onboarding flow with progress checklist, modal walkthrough, and a tooltip series, and rolling it out to a 200-user activation cohort over two weeks. The cohort I had in mind was product-led growth teams running in-product onboarding and feature adoption flows, distinct from Sprig (in-product UX feedback microsurveys), distinct from Maze (pre-launch prototype testing), and distinct from Pendo (peer competitor with deeper analytics weight).

The no-code flow builder is the strongest part. Building the 4-step checklist took about 15 minutes including conditional branching for users who skipped a step. Targeting by user attribute (plan tier, days-since-signup, feature flags from Segment) worked first try without needing engineering to redeploy. The A/B testing layer let me run two onboarding variants against each other and watch activation rate diverge in the dashboard within 48 hours, which converts directly into a shipped decision rather than another meeting. The analytics tab tied flow completion rate to actual product events I had already instrumented in Mixpanel via the Segment integration, which closed the loop on whether onboarding moved the needle on activation.

What surprised me was how brittle complex flows became. A 6-step tooltip series with conditional branching broke on Safari mobile when the underlying app rearranged the DOM mid-session, leaving 3 testers staring at floating tooltips pointing at nothing. Selector-based element targeting (the standard approach) is fragile when frontend engineers change CSS classes between releases without notice, and there is no good "alarm me when my flow target disappeared" surface in the dashboard. The pricing also escalates by MAU bucket: crossing 10k MAU pushed pricing into a tier where Pendo becomes price-competitive with deeper analytics included.

Userpilot is the right pick for PLG teams up to roughly 10k MAU running mid-complexity in-product onboarding and feature adoption flows where speed-to-launch matters more than analytics depth. Skip it if your scale is enterprise (Pendo or WalkMe are more battle-tested), if your need is in-product feedback rather than guided flows (Sprig microsurveys are purpose-built), or if your frontend changes constantly without coordinating with the growth team (the selector-based targeting model breaks under that pressure).