Replace ConfigCat's targeting rules with Statsig

Tue Jul 08 2025

Feature flags started as simple on/off switches. Now companies need to know if those switches actually move metrics - and ConfigCat's basic targeting rules can't answer that question.

Statsig emerged when engineers from Facebook realized every feature flag should be an experiment. They built a platform that turns ConfigCat's static targeting into dynamic learning systems, processing trillions of events for companies like OpenAI and Notion. The difference isn't just philosophical; it's architectural.

Company backgrounds and platform overview

ConfigCat built their platform around simplicity. Non-technical users toggle features through clean dashboards. Basic percentage rollouts work reliably. Their transparent pricing attracts teams wanting straightforward feature management without complexity.

But simplicity has limits. ConfigCat tracks who sees which flags - useful for debugging, insufficient for product decisions. You know 50% of users got the new checkout flow. You don't know if it increased conversion.

Statsig approaches feature flags differently. Every flag carries built-in analytics and statistical testing. The platform bundles four products - experimentation, feature flags, analytics, and session replay - into one data pipeline. This integration transforms basic targeting rules into sophisticated experiments that measure actual impact.

The architectural difference shows in scale. ConfigCat handles billions of configuration downloads monthly. Statsig processes trillions of events daily, enabling companies like Notion to run 300+ concurrent experiments instead of managing simple rollouts.

"We evaluated Optimizely, LaunchDarkly, Split, and Eppo, but ultimately selected Statsig due to its comprehensive end-to-end integration," said Don Browning, SVP at SoundCloud.

Feature and capability deep dive

Core feature comparison

ConfigCat's targeting rules follow standard patterns: match users by attributes, set percentage rollouts, schedule releases across environments. These features work exactly as advertised. A product manager can enable a feature for "users in California" or "20% of premium accounts" without engineering help.

Statsig takes those same targeting capabilities and adds statistical rigor. That 20% rollout becomes a controlled experiment with:

  • Automated power analysis to determine sample size

  • CUPED variance reduction for faster results

  • Sequential testing to prevent peeking bias

  • Interaction detection between concurrent experiments

These aren't academic features. Notion's engineering team credits these statistical methods for scaling from single-digit to 300+ experiments quarterly. One engineer now handles what previously required a team of four.

"Statsig enabled us to ship at an impressive pace with confidence," said Software Engineer Wendy Jiao at Notion.

Analytics and developer experience

ConfigCat shows basic flag exposure data. You see which users received which variants. The analytics stop there - no conversion tracking, no funnel analysis, no impact measurement. Teams export data to other tools for deeper analysis.

Statsig embeds complete product analytics alongside feature flags. Every flag automatically tracks:

  • Conversion metrics: Did users complete the desired action?

  • Engagement patterns: How did behavior change?

  • Revenue impact: What's the actual business value?

  • User segments: Which groups responded differently?

Bluesky leveraged these analytics to detect unexpected Brazilian user surges. Instead of guessing why traffic spiked, they analyzed user journeys and adapted features in real-time based on actual usage patterns.

The developer experience reflects each platform's philosophy. ConfigCat provides clean APIs and straightforward SDKs across 30+ languages. Statsig matches that SDK coverage but adds warehouse-native deployment. Your data stays in Snowflake or BigQuery - experiments run directly on your infrastructure without data duplication.

Engineers appreciate Statsig's transparency. Click any metric to see the underlying SQL query. Debug unexpected results by examining the exact calculations. Brex engineers reported being "significantly happier" after switching, citing clearer experiment results and reduced debugging time.

Pricing models and cost analysis

Pricing structure comparison

ConfigCat charges based on configuration downloads - each time a client checks for updated flags counts against your quota. The math gets tricky fast. A mobile app checking flags every 30 seconds burns through 2,880 downloads per user daily. Their free tier's 10 million monthly downloads support roughly 3,500 active users at that rate.

Statsig flips the model: feature flags cost nothing. Zero limits on flag checks, unlimited MAUs, no configuration download caps. You pay only for analytics events and session replays. The free tier includes:

  • Unlimited feature flags

  • 1 million analytics events

  • 50,000 session replays

  • Complete experimentation platform

Real-world cost implications

Consider a B2B SaaS with 50,000 monthly active users. Each generates 15 sessions with 8 flag checks per session. ConfigCat counts that as 6 million downloads monthly - requiring their $99/month plan just for basic flagging.

The same usage on Statsig remains free for feature flags. Adding comprehensive analytics (10 events per session) generates 7.5 million events monthly. That fits comfortably in Statsig's $150/month tier - which includes full experimentation, not just flags.

Enterprise scaling amplifies the difference. ConfigCat's dedicated plans require custom negotiations with unclear pricing. Statsig publishes transparent volume discounts, often 50% or more at scale. But the real savings come from consolidation. Brex reduced platform costs by over 20% by replacing separate tools with Statsig's unified platform.

"The biggest benefit is having experimentation, feature flags, and analytics in one unified platform. It removes complexity and accelerates decision-making," said Sumeet Marwaha, Head of Data at Brex.

Decision factors and implementation considerations

Implementation complexity and support

ConfigCat wins the setup race. Drop in their SDK, create flags in the dashboard, start toggling features. Their focused approach means fewer decisions during implementation. Small teams ship basic feature flags within hours.

Statsig requires more upfront planning because it delivers more value. You're not just managing flags - you're designing experiments. Key implementation considerations include:

  • Metric definition: What outcomes matter for each feature?

  • Statistical power: How many users needed for conclusive results?

  • Rollout strategy: Progressive rollouts or immediate 50/50 splits?

  • Analysis cadence: Daily metric reviews or weekly deep dives?

The platforms differ most in support philosophy. ConfigCat provides responsive developer support via chat and email. Statsig adds dedicated customer data scientists who help design statistically valid experiments. These specialists review your test plans, suggest metric improvements, and interpret complex results.

Notion's experience illustrates the value. Their data science team initially worried about supporting hundreds of experiments. Statsig's automated analysis and expert guidance let one engineer handle the entire experimentation platform.

Enterprise readiness and scalability

Both platforms handle enterprise traffic, but their architectures serve different needs. ConfigCat optimizes for configuration delivery at scale - billions of flag checks monthly with 99.99% uptime. Their CDN-backed infrastructure ensures fast flag evaluation globally.

Statsig operates at a different scale entirely: trillions of events daily across all product surfaces. This isn't just about handling traffic. Processing this volume enables sophisticated analytics that ConfigCat can't match:

  • Real-time metric computation across millions of users

  • Complex funnel analysis with arbitrary event sequences

  • Automated anomaly detection in experiment results

  • Cross-experiment interaction effects

Data governance often determines platform choice. Statsig's warehouse-native deployment keeps sensitive data in your Snowflake, BigQuery, or Databricks instance. ConfigCat requires sending user data to their cloud infrastructure - a dealbreaker for many regulated industries.

Ancestry scaled from 70 to 600 experiments annually after adopting Statsig. This 8x growth required more than traffic handling. It needed automated statistical analysis, clear result visualization, and workflows supporting dozens of concurrent tests. Pure feature flag tools can't enable this experimentation velocity.

Bottom line: why Statsig transforms basic targeting into growth engines

ConfigCat's targeting rules work exactly as designed - they turn features on and off for specific users. But modern product teams need more than reliable switches. They need to know if those switches drive growth.

Statsig replaces static targeting with dynamic experimentation. Every ConfigCat rule - "enable for 20% of users" or "target premium accounts" - becomes a measurable experiment in Statsig. You don't just roll out features; you prove their impact with statistical confidence.

The cost structure reinforces this philosophy. ConfigCat charges for basic flag delivery, limiting growth. Statsig gives unlimited flags free, charging only for the analytics that make those flags valuable. Teams get comprehensive experimentation tools for less than ConfigCat's feature flag pricing.

Real companies see real results from this approach:

These teams didn't just manage feature rollouts. They built learning engines that improve with every release.

"We were able to increase our experimentation velocity by over 8x while maintaining statistical rigor," reported the data team at Ancestry.

The architectural difference matters. ConfigCat processes configuration downloads; Statsig processes user behavior. ConfigCat tells you who saw what; Statsig tells you what actually worked. One manages features. The other optimizes outcomes.

Closing thoughts

Feature flags evolved from simple switches to sophisticated experimentation platforms. ConfigCat serves teams needing basic targeting rules, but Statsig transforms those rules into growth opportunities through integrated analytics and rigorous testing.

The transition from ConfigCat to Statsig isn't just a tool swap - it's a mindset shift from shipping features to shipping improvements. Every release becomes a learning opportunity. Every targeting rule becomes a hypothesis to validate.

Want to explore further? Check out Statsig's migration guide for moving from traditional feature flag tools, or dive into their experimentation best practices based on lessons from companies running thousands of tests monthly.

Hope you find this useful!



Please select at least one blog to continue.

Recent Posts

We use cookies to ensure you get the best experience on our website.
Privacy Policy