G2

Nirav M.

Senior Software Engineer

Enterprise (> 1000 emp.)

9/11/2026

"Dead-Simple A/B Test Wiring with Powerful Bucketing, Holdouts, and Overrides"

4/5

What do you like best about Statsig?

Experiment assignment is dead simple to wire into backend code.

We run dozens of concurrent A/B tests on our search and service pages, and the SDK handles bucketing, holdouts and layers without us building any of that plumbing. The override tool is what I use most day to day. I can force myself into a variant for QA on staging in seconds instead of hunting for the right user id. Gate rollouts by platform, app version or percentage have made ramping new features a lot less scary. Review collected by and hosted on G2.com.

What do you dislike about Statsig?

Debugging why a user landed in a specific variant is harder than it should be. When a gate rule, a market exclusion and an experiment layer all interact, the console doesn't give a clear trace of which rule won, so we end up reading SDK logs instead. Overrides are also easy to forget about and there's no nudge when a stale one is still sitting on your account. Would love a per-user "why this assignment" view and an expiry on overrides. Review collected by and hosted on G2.com.

What problems is Statsig solving and how is that benefiting you?

Before Statsig, every experiment meant hand rolling bucketing logic and a one-off flag, and cleanup was always someone's forgotten TODO. Now assignment, ramping and kill switches all live in one place, so shipping a new variant of a page is a config change plus the feature code, not a week of plumbing. We can run a multi-week marketplace test across web, iOS and Android with consistent bucketing per user, which we couldn't do reliably before. The biggest win is confidence: rolling back a bad variant takes seconds instead of a deploy. Review collected by and hosted on G2.com.