June 2025 Summaries

The company is launching its Managed Warehouse product, which aims to solve common problems associated with data warehouse and event tracking setups, such as complicated setup and maintenance, delayed data refresh, and costly querying. This new product offers a one-click setup, zero maintenance, and real-time data arrival, making it easier for companies to detect issues quickly. The Managed Warehouse is powered by ClickHouse, an open-source database optimized for fast analytics at scale, and provides raw SQL access and other benefits of a true Warehouse Native platform without the cost. Additionally, the company is also launching Feature Usage Analytics, which takes advantage of the Managed Warehouse's benefits to provide insights into feature flag usage, evaluation, and rules, making it easier to manage and debug issues, as well as identify stale flags that are no longer being used. The GrowthBook SDK sends events to the Managed Warehouse every time a feature flag is evaluated, allowing for aggregated and displayed usage data within seconds.

The author reflects on their experience testing an experimentation platform's new MCP (Model Context Protocol) feature, which allows for seamless integration with AI tools like Claude Desktop and ChatGPT, enabling users to ask questions about experiments without navigating through the platform. This marks a fundamental shift in how experimentation platforms will evolve to become more fluid and accessible, allowing users to access their platforms in various modes such as conversation, visualizations, or precise controls, removing micro-barriers between thought and action. The author notes that while AI is still a poor experimentation partner due to its agreeable nature, the MCP feature enables experimentation without the need for users to come to the platform, making it more efficient and effective. As experimentation platforms become more fluid, they must also become more sophisticated in terms of governance and guardrails to ensure seamless interactions. The future of experimentation will be about meeting users where they are, through various modes, rather than conversational AI replacing experimentation platforms.

GrowthBook has introduced the SQL Explorer as part of its Launch Month, a tool that aims to simplify the process of answering simple data questions by providing a lightweight, built-in SQL editor. This new feature allows users to query their data, create visualizations, and save results directly within GrowthBook, eliminating the need to switch between separate tools like Looker, Mode, or Tableau. The SQL Explorer is designed for quick analyses without requiring the full capabilities of a business intelligence platform, and GrowthBook plans to add more features and integrate it deeper into the product in the coming weeks. The company is encouraging users to try out the SQL Explorer and provide feedback, which will help shape its development and future enhancements.

Feature flags, or feature toggles, are a powerful tool that allows developers to enable or disable specific functionalities in an application without deploying new code, offering significant advantages in managing software releases. These conditional statements, often represented as if/else blocks, enable teams to decouple code deployment from feature release, allowing for safer and more controlled rollouts. By using systems like GrowthBook, teams can define flag configurations, evaluate them through SDKs, and manage flag delivery to ensure features can be toggled on or off instantly. This approach minimizes risk by facilitating progressive rollouts, enabling real-world testing in production environments, and reducing mean time to recovery during issues. Moreover, feature flags lay the groundwork for A/B testing and experimentation, helping companies make data-driven decisions to optimize product features. GrowthBook, an open-source platform, provides robust tools for managing feature flags and experiments, offering flexibility, transparency, and cost control, while also supporting integration with existing infrastructure and data warehouses. Effective use of feature flags not only improves deployment reliability but also offers a competitive edge by enhancing the agility and responsiveness of development teams.

This week introduces Insights, a new section in the sidebar featuring an updated Executive Dashboard, Learnings and Timeline pages, and metric analyses to help optimize experimentation programs. The revamped dashboard provides a high-level view of an organization's experimentation program, with features such as velocity, win rate, and impact metrics that can be filtered by project and date range. The Learnings page serves as a searchable knowledge base for completed experiments, allowing team members to learn from past successes and failures. The Timeline page offers a visual representation of experiment phases and statuses, while the Metric Effects and Correlations pages provide in-depth analyses of effect sizes and correlations between metrics. These new features are designed to help teams refine their experimentation workflows, identify bottlenecks, and make data-driven decisions.

GrowthBook has declared June as its official Launch Month, during which major new features and changes will be announced every week, allowing users to try them out early before the final release at the end of the month. The first week's launches include the introduction of the GrowthBook MCP Server, which enables AI tools to communicate directly with services like GrowthBook, and enhancements to Safe Rollouts, making them even safer by minimizing user impact when things go wrong. Additionally, custom decision criteria can now be defined for the Experiment Decision Framework on a per-experiment basis, offering more flexibility and control. The search experience within GrowthBook has also been revamped, making it easier to find feature flags and metrics with improved filtering capabilities that can be easily shared or revisited. These updates aim to provide users with more powerful tools and a better overall experience, with more information available in the company's announcement blog post and documentation.