Migrating from Statsig
Step-by-step path to move off Statsig to MaxLift in an afternoon.
Statsig was a great choice. Since the OpenAI acquisition (Sept 2025) and the handover of the platform and customer base to Amplitude (May 2026), many teams are re-evaluating. This guide is the practical path from Statsig to MaxLift: what maps 1:1, what differs, and a migration you can complete in an afternoon for a typical setup.
Concept mapping
Here is how key concepts and SDK calls map between Statsig and MaxLift:
| Statsig | MaxLift | Notes |
|---|---|---|
| Feature Gate | Boolean flag | direct equivalent |
| Dynamic Config | JSON flag | direct equivalent |
| Experiment | Experiment | A/B/n, same mental model |
| statsig user object | {unitId, attributes} | your stable ID + attribute map |
| checkGate(user, 'gate') | variation('flag', ctx, false) | OpenFeature-compatible |
| getExperiment(user, 'exp').get('param') | experiment('key', ctx).payload.param | direct equivalent in evaluation context |
| Warehouse Native (WHN) mode | The whole product | we ONLY do warehouse-native + counters preview — no 'sync your events to us' mode exists |
What you gain / what changes (honest)
Gain: independence (no bundle agenda); flat $299/mo pricing (no event-volume tiers); metrics computed in YOUR warehouse with reproducible SQL; vertical experiment templates; and OpenFeature-standard SDK portability.
Changes to plan for: MaxLift readouts refresh from your warehouse on a cadence (hourly on Growth) rather than streaming live dashboards — right for decisions measured in days, different feel from Pulse. Fewer SDK languages today (JS/TS first-class; Java/Go in progress).
Migration steps (typical web/app setup, ~an afternoon)
- Export your config: Pull gates, configs, and experiments via the Statsig Console API. We provide a mapping script that converts the JSON export into MaxLift flags/experiments as drafts.
- Map identity: Whatever userID you sent Statsig is your MaxLift unitId. That same ID must exist in your warehouse event tables (already met for WHN Statsig users).
- Swap SDK calls: Init once with your MaxLift key, and swap call sites. Both SDKs evaluate locally, so latency remains sub-millisecond.
- Recreate metrics: Point MaxLift metric definitions at your existing warehouse tables. Your conversion tables don't move.
- Cut over at natural boundaries: Finish in-flight tests on Statsig to avoid re-randomization; launch new tests on MaxLift.
- Verify: Launch a built-in A/A experiment; SRM check green within a day confirms assignment and event flow.
FAQ
Can we keep Statsig running during transition? Yes — parallel running behind an env variable is the recommended cutover for cautious teams.
Contracts? We don't do annual lock-in below Enterprise. Monthly, cancel anytime; our pilot is $500 credited, with a 5x-or-refund guarantee.