GrowthBook Files

Release Highlights

Product Analytics (preview) Turn your warehouse data and metrics into actionable product insights. Explore user behavior, share dashboards, and make smarter decisions about what to build next. This release is just the beginning and we have much more planned for Product Analytics in the coming months.

Metric Slices Automatically break down metrics into key “slices”. This can drastically reduce the number of metrics you need to create and can help organize experiment results. For example, if you have an Orders fact table, you can enable “Auto Slices” on the product_category and device_type columns. Now, when you add a Revenue metric to an experiment, it will also include breakdowns for each of these, so you can easily compare mobile vs desktop revenue or revenue from different product categories. You no longer need to maintain 10+ separate metrics that are all slight variants of each other!

Incremental Refresh We revamped our Data Pipeline Mode to drastically reduce data warehouse costs and increase query speeds for long-running experiments and high-traffic apps. We do this by storing intermediate results back into the data warehouse and incrementally refreshing it throughout the experiment. We’re starting with BigQuery, where we’ve seen some users save up to 85% in query costs.

Statsig Importer and Code Migration Tool OpenAI recently acquired Statsig and the future of the platform is uncertain. Everyone is asking “will Statsig shut down?” We built an importer and migration tool so teams can be ready to switch to GrowthBook on a moment’s notice. Our Importer lets you instantly copy over all feature gates, dynamic configs, segments, and more. Our Code Migration tool (built on Claude Code) can seamlessly migrate your app from using the Statsig SDK to the GrowthBook SDK.

Official Metrics Many companies have a core set of metrics that have been validated and battle tested over time. GrowthBook lets you mark these metrics as “Official”, which gives them a special badge in the UI, elevates them in the app, and locks down editing. We’ve had Official metrics for a while, but they had to be managed entirely through the API. In this release, we now let admins manage these metrics directly from the UI, making them more accessible and easy to use.

Custom Validation Hooks GrowthBook has always been extremely customizable, but this release brings it to a new level. Self-hosted users can now write fully custom javascript code snippets that hook into our validation logic. Require feature flags to have tags, block targeting rules that contain PII, enforce custom naming conventions - the possibilities are endless. The validation code runs in a sandbox based on V8 isolates for ultimate security and performance. We only have a couple hooks to start, but we’d love to hear your ideas and use cases for additional ones!

New SQL Template Variables You can now access custom field values and phase data within your metric and experiment SQL! This unlocks 3 important use cases:

1. Add highly custom query optimization by taking advantage of non-date partition keys.
2. Re-use the same SQL definitions when only minor tweaks are needed per-experiment.
3. More accurately join experiment exposure events to experiment phases.

Edge Remote Eval Remote Eval lets client-side SDKs offload feature flag evaluation to a back-end server, which prevents leaking targeting logic to your users. Previously, this required running your own GrowthBook Proxy servers (and dealing with all of the associated infrastructure and maintenance). We now offer a simpler option based on Cloudflare Workers. Deploy a blazing fast and cheap Remote Eval server on Cloudflare’s global infrastructure with zero maintenance required.

Quality-of-Life Improvements

Big thanks to all of our users who reported bugs, provided feedback, and worked with us to make GrowthBook better. This release includes many quality-of-life improvements that, while small on their own, add up to make a big impact. Here are a few highlights:

• Brand new search algorithm for features, metrics, and experiments. Results are now much more relevant and search is faster overall, even when there are thousands of potential matches.
• Ability to create feature rules in multiple environments at once
• Better column-type detection for Fact Tables when using BigQuery
• You can now add metric row filters based on boolean columns in your Fact Tables
• We stopped sending feature.updated webhooks for unpublished draft changes (reducing noise).
• Slack and Discord notifications now include more details about the change in GrowthBook that triggered the notification.
• Custom Pre-launch experiment checklist items can now be scoped to specific projects.
• Much faster database schema browser when there are hundreds or thousands of tables
• New settings option to disallow creating legacy metrics. This can help ease the transition to using fact tables. There is more planned here so stay tuned!
• The experiment results table is now sortable, so you can quickly see the best (or worst) performing metrics based on either significance or lift.
• Complete revamp of the included sample data and experiments to better showcase the GrowthBook platform and all that it can do.
• Plus dozens of bug fixes and smaller improvements.

New Contributors

• @kosmoz made their first contribution in https://github.com/growthbook/growthbook/pull/4011
• @aalfalah made their first contribution in https://github.com/growthbook/growthbook/pull/4518

Plus a ton of returning contributors: @ahdriel @gazzdingo @mknowlton89 @bryce-fitzsimons @Kevin-Chant @madhuchavva @royalfig @kosmoz @tzjames @lukesonnet @aalfalah @msamper @Auz @jdorn @nodirnasirov @Bohdan-Kim @lukebrawleysmith @natasha-growthbook

Full Changelog: https://github.com/growthbook/growthbook/compare/v4.1.0...v4.2.0