Treatwell accelerates product experimentation across marketplace teams
How Treatwell’s engineers replaced assumptions with experiments to optimize checkout flows and increase conversions.
Treatwell provides salon management software connected to Europe’s leading booking platform. They work with over 75,000 salon partners across 15 countries, processing over 5 million bookings every month.
Executive summary
Treatwell, Europe’s leading beauty and wellness marketplace, has built experimentation into the core of its product development culture. Using GrowthBook, the company’s three marketplace engineering teams now run experiments as their standard operating mode, testing feature changes before rollout and making data-driven decisions that improve booking flow efficiency and reduce cancellations.
By connecting directly to Treatwell’s Redshift data warehouse and integrating through Java, React, Swift, and Kotlin SDKs, GrowthBook has helped Treatwell close a critical measurement gap, enabling faster, more confident iteration and more scientific collaboration between product, data, and engineering teams.
Challenge: from pre-post analysis to true experimentation
Before GrowthBook, Treatwell’s experimentation capabilities were limited. Teams relied mainly on “before vs. after” analysis to infer impact from new releases, a method that left gaps in understanding and confidence.
“We had a visible gap in our ability to measure change. We needed a more scientific way to test hypotheses and understand customer behavior.”
— Marek Maciusowicz, Head of Engineering for Marketplace and Payments
Treatwell sought a platform that:
- Empowered cross-functional product teams to run experiments without heavy data-science bottlenecks.
- Integrated with existing data pipelines via mParticle and Redshift.
- Supported a gradual adoption model with flexible pricing and scalability.
Solution: GrowthBook’s warehouse native platform
After evaluating multiple vendors, Treatwell chose GrowthBook for its open, flexible architecture and self-hosted data model.
“GrowthBook’s ability to connect directly to our data warehouse was a key differentiator. It fits perfectly with our infrastructure and privacy approach.”
— Marek Maciusowicz, Head of Engineering for Marketplace and Payments
Why GrowthBook
Impact: measurable wins and unexpected learnings
Streamlining the booking flow
GrowthBook-powered experiments in the checkout process improved booking completion and reduced cancellations through UI and messaging clarity.
Surprising results from a click-reduction test
One test removed a step between checking staff availability and checkout.
Some feared users would act unintentionally, which could lead to more of them dropping out of the flow, but conversion rates increased, validating a customer-first, data-driven approach.
“Experimentation showed what customers actually do rather than what we assume they’ll do.”
— Marek Maciusowicz, Head of Engineering for Marketplace and Payments
Learning from neutral outcomes
An experiment highlighting salon portfolio images didn’t improve conversions, despite a strong belief that it would. This lesson has reinforced the importance of rigorous experimentation within the marketplace teams.
“We expected higher engagement, but when we were proven wrong, we had to develop more experiments rather than stick with our pre-conceived expectations.”
— Marek Maciusowicz, Head of Engineering for Marketplace and Payments
Experimentation at scale: a cross-functional effort
Each marketplace team blends roles to move from idea to insight quickly:
- Product Managers: Define hypotheses and own outcomes
- Designers & Researchers: Shape variant designs and interpret user reactions
- Engineers: Implement and monitor experiments using GrowthBook SDKs
- Data Scientists: Validate results and contrast them against broader analytics
With shared visibility and tight iteration loops, most experiments now move from start to launch in under three weeks, even with taking on more complex experiments.
How Treatwell experiments
Tech Stack
- Data Warehouse: Amazon Redshift
- CDP: mParticle
- SDKs: Java, React, Swift, Kotlin
- Analysis: GrowthBook, Mixpanel, Looker
Workflow
- Hypothesis
Product managers define the opportunity and target metric. - Design & Build
Designers and engineers collaborate on variant creation. - Launch in GrowthBook
Engineers deploy experiments that use feature flags and randomized variation assignment. - Monitor & Analyze
Product Managers and Data Analysts validate results using GrowthBook + Redshift data. - Learn & Scale
Teams integrate learnings into future product iterations.
Typical timelines
- Implementation: < 3 weeks
- Runtime: 3–4 weeks
- Decision: Data-driven “go” or “no go” to full user base
Results
- Full adoption across 3 marketplace engineering teams
- 10x more experiments can be run
- 4-5x faster experiment set up (configuration and experimentation logic)
- 3-4 week average runtime for valid statistical results
- Checkout conversion uplift from UX and flow optimizations
- Culture shift: Experimentation embedded in every release
Looking ahead
Treatwell plans to expand its experimentation footprint to mobile and explore new GrowthBook capabilities, like holdout groups and the enhanced Visual Editor, to accelerate insight generation.
“We’ve built strong momentum. Experimentation is now part of how we think, build, and improve. GrowthBook made that possible.”
— Marek Maciusowicz, Head of Engineering for Marketplace and Payments
Ready to ship faster?
No credit card required. Start with feature flags, experimentation, and product analytics—free.
