Run revenue-aware A/B tests with Statsig and subscription analytics
Statsig is an experimentation platform that runs A/B tests as features ship and reports their effect on business metrics such as user growth and revenue. A subscription analytics platform can forward lifecycle events — trial starts and cancellations among them — into Statsig, so an experiment's verdict is not limited to a click-through number. This guide covers three experiment shapes worth trying, the two-console setup, and how to wire a first feature gate.
Three experiment shapes worth trying
A frequent mistake is to judge an experiment on one metric and miss the collateral movement in others. The same change can lift one number and quietly depress another, and no two apps respond identically. Treat each shape below as a hypothesis to test on your own audience, not a guaranteed win.
Experiment 1: trim signup friction
The top of the funnel carries the largest user counts, so small conversion-rate gains there compound into meaningful active-user growth. Fewer required fields generally means fewer abandonments: if you currently ask for first and last name, try one name field, or collect only an email and request the rest later.
This is not universal. When personalization is central to the product, a shorter initial form can hurt engagement or retention — which is exactly why the experiment should read those metrics too, not just signup completion.
Experiment 2: reduce option overload
Early users can be overwhelmed by choice, and the more effort an action requires, the fewer people complete it. A game that opens with difficulty selection plus character setup, or a cloud console that demands a node size, an optimization family and a storage type before anything runs, asks a newcomer to make decisions they lack context for. Many games solve this with a guided single path that opens up later.
Reducing choice can backfire in three ways: the remaining options may feel too simplistic for the user's knowledge, a long tutorial can lose a bored user, and an advanced user may leave for a platform that offers more freedom. Test the reduction against engagement and retention, not just task completion.
Experiment 3: vary in-app purchase pricing and presentation
Apps that sell in-app purchases can experiment with price, discount and presentation. A price test shows different values to different users and produces fast, legible results: a higher price may reduce purchase count while still raising total revenue if the drop is small; a lower price may raise purchase count while lowering revenue if many buyers would have paid the higher amount anyway.
Short-term revenue is only part of the picture. More buyers at a lower price can improve retention and later purchases, and Statsig can track those longer-term effects to show which variation wins over time. A tiered presentation is another lever: offering Small, Medium and Large versions where Small and Large look like good value lets the Medium tier act as an upsell path toward Large.
Connect the two consoles
The setup is a short sequence of steps split across the two products. Complete the Statsig side first, because the subscription analytics side needs the key it produces.
In Statsig
- Create a free Statsig account.
- Open the Statsig integration page and enable the subscription analytics integration.
- Open the Statsig API Key page and copy the Server Secret Key.
In the subscription analytics dashboard
- Open your project in the subscription analytics dashboard and locate the Integrations card in the left menu; select + New.
- Choose Statsig from the Integrations menu.
- Paste the Server Secret Key you copied from the Statsig console.
After these steps, subscription events begin appearing on your Statsig metrics page. The subscription analytics documentation covers the integration in more depth if a step does not behave as expected.
Wire your first feature gate
A feature gate is the mechanism behind a Statsig A/B test: it is a named switch whose targeting rules determine which users get the new experience. The workflow runs as follows: install the Statsig SDK for your language or framework (the JavaScript SDK is one example); create a new feature gate in the Statsig console, a step the Statsig docs walk through; add targeting that defines who receives the variation, such as 50% of Android users; initialize the SDK in your app; and branch your app's behavior on whether the user passes the gate. The initialization and gate-check snippets appear on this page as exact code blocks, so use those as the reference for the two code steps instead of retyping them from memory. With the code live, Statsig automatically compares gate-passers against non-passers. The user IDs from the Statsig SDK are joined to the events arriving from the subscription analytics platform, and that join is what lets Statsig report how the experiment moved growth and revenue. After a few days of traffic, the gate's Pulse Results section shows the comparison data.
Pre-launch checklist
- The gate has a name you will recognize in the Pulse Results section.
- Targeting is explicit about platform and share of users, not left at a default you cannot explain.
- The SDK initializes with a stable user ID that matches the identity the subscription analytics platform reports events against.
- The metric set includes at least one revenue measure and one retention or engagement measure, so a single-metric win cannot hide a regression.
- You have decided in advance how long the test runs before reading results.
Telling two failures apart. If no events appear in Statsig at all, the integration step is the likely culprit — check that the subscription analytics integration is enabled and the Server Secret Key was pasted correctly. If events appear but the gate shows no split, the SDK initialization or the user-ID join is the likely culprit. These are different problems with different fixes, and separating them saves debugging time.
Why run experiments at all
Testing each feature as it ships turns opinion into evidence about how users actually respond. That evidence supports the changes you keep and informs what you build next. Connecting Statsig to a subscription analytics platform extends the picture from feature usage to the subscription outcomes that fund further development.
Frequently asked questions
What does the subscription analytics integration actually send to Statsig?
Subscription lifecycle events, including trial starts and cancellations. Those events are joined to the Statsig SDK's user IDs so experiment results can be read against revenue and growth metrics.
Where do I find the key the subscription analytics platform needs?
In the Statsig API Key page, copy the Server Secret Key, then paste it into the Statsig integration entry in your subscription analytics project.
How long before results are visible?
The Pulse Results section of the gate begins showing comparison data after a few days of running the test. Treat that as a starting point for reading results, not a guarantee that a few days is enough traffic for a confident decision.
Is a lower in-app purchase price always worse for revenue?
No. A lower price can increase purchase volume, and if many of those buyers would not have purchased at the higher price, total revenue can rise. It can also improve retention and later purchases. The only reliable answer comes from testing price against revenue and retention together.
Can simplifying signup hurt my app?
Yes, in apps where personalization is core. A shorter initial form can reduce engagement or retention in that case, so the experiment should measure those outcomes alongside signup completion.