Glossary: A/B Testing

A/B testing (or split testing) compares two versions of a page, flow, or campaign, shown to different segments of users, to see which one performs better against a defined goal. It’s a core method for making data-driven product and growth decisions.

Examples

• A team tests two versions of a landing page headline to see which drives more wallet connects.

• A mint flow is A/B tested with and without a countdown timer to measure impact on conversion rate.

• An incentive campaign compares two reward structures to see which drives higher retention.

A/B testing in crypto apps

In a crypto app, the goal of a test is usually an onchain action: a wallet connect, a first swap, a deposit or a mint. These outcomes are rarer than clicks, so tests need more traffic or more time to reach a clear result. Teams also need a stable way to keep each visitor in the same variant, and a way to join that variant to the wallet that transacts later.

Example

A perp DEX tests two onboarding screens. Variant A shows a deposit button first; variant B shows a short tutorial. Each variant gets 5,000 visitors. A converts 3.2% of visitors to a first deposit and B converts 4.1%. Before the team ships B, it checks that the gap holds over two full weeks and that B’s depositors trade again the next week.

How to measure A/B tests with Formo

Formo does not assign variants. Your feature flag or testing tool assigns them, and you record the variant on a custom event, for example an experiment_viewed event with a variant property. In a funnel, filter the first step by that property to compare how each variant converts to a wallet connect, a transaction or any later event.

Related reading

Explore Conversion Rate, Funnel Analytics, Product Metrics, and Event Tracking.

FAQs

What’s needed to run a valid A/B test?

A clear goal metric, random and comparable audience segments, and enough traffic to reach statistical significance.

How long should an A/B test run?

Long enough to reach a meaningful sample size and account for normal day-to-day or weekly variation in traffic.

Can A/B testing be used for onchain flows?

Yes. Teams can test different contract interactions, fee structures, or UI flows and measure the impact on onchain conversion rate.

Can Formo measure an A/B test?

Yes, if you record the variant. Your testing tool assigns variants, you send the variant as a property on a custom event, and a funnel filtered by that property shows how each variant converts to wallet connects, transactions or later events.