Find bad features before your customers do

Releasing new features shouldn’t be a risky, nail-biting experience.

Split monitors every feature you release, detects features that degrade performance, and alerts the team responsible so they can quickly turn it off.

Instantly identify the feature that broke

Split continuously compares your feature flags and performance data to find anomalies caused by new features.

Push a change and within 30 minutes we will identify if any features caused a better or worse user experience. We calculate causation; no need for manual data correlation.

Complex releases with many features? Looking for subtle changes to a baseline? No sweat, Split handles data attribution and analysis.

Instantly notify the right team and kill the feature

Once a bad feature is identified, Split notifies the team responsible for the feature flag.

A kill switch on each feature flag means that the team can simply disable the feature in production. No rollbacks or hotfixes required.

Powered by your product and engineering data

Monitor features for movement in any key metric that matters for your application.

Select key performance, service level objective (SLO), errors, and business metrics and Split will notify you when a newly released feature is affecting them.

Split plugs into popular data sources such as Google Analytics, Sentry, Segment, mParticle, and New Relic. Or collect performance data within your app via track calls or RUM agent.

Ensure every (code) change is for the better

Vistaprint validates the roll out of every new feature against performance metrics, killing bad features within minutes.

Make releases a non-event with alerting

Getting new features in the hands of your customers shouldn’t be risky. Split monitors, alerts, and kills features that aren’t ready for prime time.