Battle for Product Data Model
Statsig
The key battle is over who owns the product data model, because the team that already defines events, users, funnels, and success metrics can bolt on experimentation with much less setup. Amplitude already sits on that behavioral data and now packages testing into the same workflow. Statsig is pushing the opposite direction, using one event stream for flags, experiments, and analytics so a release can become a measured test without moving data between tools.
-
In practice, analytics first adjacency is simple. A product team already tracking signup, retention, and checkout events in Amplitude can reuse those same metrics for an A/B test, often through guided and no code workflows, instead of instrumenting a separate experimentation stack.
-
Statsig is trying to remove that same friction from the other side. Developers put a feature behind a flag, exposures and outcomes flow through the same data layer, and the same system powers dashboards, funnels, user views, and experiment readouts. That makes release control and measurement feel like one product.
-
Split shows why the category is converging. Its pitch is that once teams use feature management and release monitoring, the groundwork for experimentation is already in place, and under Harness it is now being folded into a broader software delivery suite. The land grab is for adjacent budget, not just better tests.
Going forward, the winners will be the vendors that turn existing telemetry into the next product someone buys. Analytics platforms will keep moving down into testing, and experimentation platforms will keep moving up into analytics. That favors products like Statsig that make event collection, metric definition, rollout, and analysis feel like one continuous workflow.