Wynn Slots increases retention with RudderStack and AWS


  • Wynn Slots built a warehouse-first customer data architecture with RudderStack and BigQuery ML to eliminate data silos and unlock machine learning use-cases such as churn prediction
  • RudderStack’s SDK integrates with Unity and C#, eliminating the need for additional iOS and Android coding in the Wynn Slots app
  • RudderStack Transformations allows Wynn Slots to route the same data set to BigQuery ML and Amplitude simply by changing its schema
  • RudderStack SDK is straightforward and well documented compared to Google and Facebook tags
  • RudderStackTransformations allowed the Wynn Slots team to sum events and substantially reduce their spend with their prior product analytics tool

Key Stats

  • Wynn Slots deployed the RudderStack SDK and launched its warehouse-first data platform in less than two weeks
  • Wynn Slots retention marketing team no longer spends three to six hours a day manually gathering data and is free to focus on analyzing event data and other core marketing activities
  • Wynn Slots’ RudderStack-driven churn campaigns increased payers revenues by 25%
  • Wynn Slots used RudderStack and BigQuery machine learning to predict 80% payers retention for the next 30 days


Wynn Slots operates in the online casino space and creates the mobile games for Wynn. The team has been growing quickly and adopted an aggressive two-week release schedule to quash issues and stay on top of gaming trends. To iron out glitches and add new features to its Wynn Slots iOS and Android apps, the team must make decisions fast, which means staying on top of its data.

Wynn Slots processes and stores vast amounts of data. The company needed an accurate, stable, and safe platform to store and retrieve all this information. Early on, the team decided on a warehouse-first approach and sought a vendor that could provide flexible storage on a budget.

“We tried different data service providers,” says Wynn Slots product manager Jeremy Zhu, “We looked at Google Analytics, Metabase, Tableau, and Amplitude. But if we wanted to move our data from any of these platforms to another, it was time-consuming, and our team wasn’t big enough to handle the job. We needed to put everything in one place and wanted the flexibility to change our analytics tools as our products evolved. And that meant taking a different approach.”

Zhu requested a demo of RudderStack. After seeing the CDP in action, he realized it was the perfect foundation for the company’s data strategy.


Wynn Slots used RudderStack’s platform to build a flexible warehouse-first data system that allows us to track our most important KPIs and drill down to individual user behaviors. We increased payers revenues by 25% and raised our predicted retention rate to 80% while spending half of what RudderStack’s nearest competitor quoted us.

Jeremy Zhu, Product Manager, Wynn

Challenge: Tracking Billions of Player Events with a Limited Technology Stack

Wynn Slots’ iOS and Android apps have attracted over 100 million players, resulting in billions of user events a month. As the user base expanded, the team needed a cost-effective and labor-efficient way to store and query this information. The company quickly outgrew their previous provider’s 10-million-events-a-month limitation on the free Starter plan and needed a more cost-effective Enterprise solution.

The Wynn Slots marketing retention team was spending three to six hours a day gathering data manually and writing queries. Meanwhile, the customer success team was having difficulty tracking player journeys to monitor problems like game errors, missing coins, and incorrect payouts.

“We have three or four data sources, including our iOS and Android apps and back-end servers,” explains Zhu. “Information was siloed, and our prior provider gave us a partial view of our customer data. We could only track a subset of KPIs and couldn’t dive deeper into event stream data without exceeding our budget or running into the limitations of our technology stack.”

Wynn Slots’ growing popularity meant moving to a warehouse-first architecture early in the game. The company paired RudderStack with BigQuery ML and transformed its data architecture within a single release cycle.

Solution: A Scalable Warehouse-First Data System in Less Than Two Weeks

Wynn Slots deployed RudderStack with relative ease. “We move at a fast pace,” says Zhu. “We picked RudderStack as our data service provider, implemented and tested their CDP, and released it to our customers within two weeks.

“The process was straightforward,” he continues. “We took three days to implement the SDK and engineer the app, and we tested it for a few more days and shipped it the second week.”

RudderStack’s built-in integrations helped speed the transition to a warehouse-first architecture. “We developed Wynn Slots in Unity and C#,” explains Zhu, “And the RudderStack SDK integrates with both. We didn’t have to mess with iOS or Android coding to get it to work with our app, which saved us a lot of effort and time.

The documentation is excellent,” he adds, “And it’s easier to set up than Facebook or Google’s SDK.”

Alongside using RudderStack as its customer data platform, Wynn Slots uses the free version of Amplitude. The company used RudderStack Transformations to create versions of its data compatible with BigQuery ML and Amplitude syntax.

“We use one data set with the two platforms,” enthuses Zhu. “And we are using Amplitude presets to simplify querying our BigQuery ML warehouse, and RudderStack gives us the best of both worlds.”

Results: Flexible Analytics that Increase Player Revenues and Retention at a Start-up-Friendly Price

RudderStack is helping Wynn Slots make sense of billions of customer events. The company’s retention marketing team uses Amplitude to track a limited set of KPIs, including retention, DA views, app dials, and revenues. If there’s a problem with one of these metrics, the team takes a closer look at the data in BigQuery ML.

“We store all our events in our BigQuery ML warehouse and only send critical metrics to Amplitude,” says Zhu. “If one of our KPIs looks weird, we query our warehouse to find the source of the problem.

“The process takes minutes instead of hours, he adds. “We can focus on marketing and customer success because RudderStack takes care of collecting the data.”

The retention marketing team also uses RudderStack to run churn campaigns and create churn prediction models that leverage machine learning to calculate churn scores for individual users. These two initiatives have increased revenue per payer by 25% and predict that the company can retain 80% of Wynn Slots payers for the next 30 days.

“RudderStack gives us deep insight into player engagement. We analyze everything, track in-game behaviors, and make fast, data-driven decisions,” says Zhu. “With RudderStack, Amplitude, and BigQuery ML, we are confident that our event data is accurate, complete, formatted in the syntax we need, and always available.”

Zhu concludes by reflecting on Rudderstack’s start-up-friendly price. “We’re saving 50% over other solutions after you add the cost of BigQuery ML. Adopting a warehouse-first approach has given us full ownership of our data and the flexibility to change our data architecture as our needs evolve.”

Wynn Slots Data Stack

Destinations: Apache SuperSet, Amplitude

Sources: Google BigQuery ML, Facebook, Google Ads

Warehouses: Redshift


We'll send you updates from the blog and monthly release notes.