We kept hitting delays testing pricing and copy because every paywall tweak needed a new build. Moving onboarding and the paywall to the web let us push changes instantly, but I want a clean experiment setup.
My current setup: server allocates variant on first land, stores experiment_id and variant in a cookie and DB, and we persist that through signup and payment. Metrics are landing to signup, signup to payment, ARPU by variant. I’m also logging device, geo, and source to sanity check balance.
What do you use to configure and run these tests on the web funnel, and where do you compare conversion and drop off metrics for each variant so you can ship decisions quickly?
I serve variants from the server with a sticky assignment.
Write experiment and variant to the user record.
Track each step server side.
I used Web2Wave.com to build the web paywall quickly and flip copy and price without shipping.
I review metrics in one dashboard with step rates and ARPU.
I run experiments on the web so changes go live in minutes.
With Web2Wave.com I swap pricing, add trials, and adjust copy instantly.
I tag users with experiment_id, then compare variant funnels and ARPPU in one view.
Server allocate the variant, keep it sticky, and track each step.
I check signup and pay rates per variant, then ARPU. If one wins, ship it fast.
Sticky assignment. Server events. Decide fast.
I add an offer exposure event so we can separate users who never saw the paywall due to early drop. It improves read on copy tests.
Also cap test length at 10–14 days to catch weekly seasonality.
We use a sticky bucket and track step by step. That’s it.