How Strava is creating the ultimate active community experience using Snowplow and Snowflake

Learn how Strava is improving its product experience after moving to Snowplow BDP and the Snowflake Data Cloud.

Strava, the subscription platform at the center of connected fitness, has transformed the way it analyzes and improves its product features. By moving to Snowplow’s Behavioral Data Platform (BDP) and the Snowflake Data Cloud, the platform has reduced the complexity of its event tracking and significantly improved its feature instrumentation coverage, resulting in greater insights into how users interact with the product.

The company is now using its new data architecture to perform advanced digital analytics for its product experiments with precise accuracy. These include experiments with its Routes feature that have helped significantly increase the number of Route views per user, as well as four primary actions Strava measures as proxies for deciding to take a particular route: Route saves, Route downloads, Route records, and Route shares per user.


Strava is one of the world’s most popular exercise tracking apps, with over 100 million active people in more than 190 countries.

In 2018, the company’s data team set out to transform how they collect and process behavioral data to foster a culture of continuous improvement. The team chose Snowplow’s Behavioral Data Platform (BDP) and the Snowflake Data Cloud to drive this shift, and have since used the technologies to make it easier for other Strava teams to access and use behavioral data.

More recently, the company’s product team wanted to figure out how they could leverage Snowplow and Snowflake to increase the analytics coverage of several product features, including Strava’s Routes product, which allows users to discover new routes and places around them.


Prior to Snowplow and Snowflake, Strava used a mix of in-house and third-party analytics tools, which made it difficult for product teams to accurately measure how users interacted with different features. This was because feature instrumentation, which refers to the process of adding data collection mechanisms to a specific attribute on a website or application, was a tedious process that was error-prone with Strava’s previous data setup.

Instrumenting a feature was something that was difficult and time consuming for our analysts, product engineers, and data engineers. And this limited the appetite for vertical teams to add tracking to their features. We thought that if we could reduce the complexity of event tracking, instrumentation coverage would improve dramatically.


In addition, Strava’s previous event tracking did not provide the rich behavioral data needed to measure and improve the performance of key product features. It also prevented the company from achieving its desired event tracking volume because it would have been too expensive.


Why Snowplow and Snowflake?

The product team needed an easier way to implement feature instrumentation across the Strava platform. They also needed a solution that would capture the detailed insights needed to measure user behavior and improve the overall product experience. This led the team to Snowplow BDP and Snowflake, which were already being used in other areas of Strava.

We were previously using a mix of in-house and third-party tools that were expensive and difficult to use. We like Snowplow’s open and transparent approach that’s built on existing technologies we’re familiar with. And it helps us keep our data within our own walls instead of sending it to a vendor server.


The product team saw Snowplow as the answer to their challenges, thanks to its ability to reduce friction in implementing instrumentation and improving event tracking across all product features. With robust event tracking, the team could gain a complete view of user interactions to inform product decisions and take Strava to the next level.

Strava’s decision to use Snowflake was based on the need to move to a data cloud platform that could meet the company’s growing data needs.

Limitations encountered with the previous data platform made it clear that a robust solution was needed to not only meet current needs, but also scale effortlessly in the face of future growth.

Additionally, Strava wanted to extract deeper insights from its data, which required a platform capable of handling complex analytical tasks. This included building attribution framework data models and using Jupyter notebooks to automate power analyses and experiment results – capabilities that are now enhanced by Snowflake.

How Strava powers product analytics with Snowplow and Snowflake

One of Strava’s features that benefited from Snowplow and Snowflake was the ‘Routes’ product.

Routes is designed to help users find the right route for their activity. After implementing a series of product enhancements to make it easier for Strava users to find the right route for them, the team wanted to measure the success of the project.

To measure the performance of a particular Route page, the product team relies on Snowplow’s event logging. These events feed into Snowflake, where the Strava team uses ETL (Extract, Transform, Load) processes to prepare the data for visualization in Tableau. This architecture also powers the company’s dashboards, experiments, machine learning pipelines, and many other high-level business metrics.

Snowplow makes it easy for the Strava team to add instrumentation to new features and implement many different event types, such as screen enter events and click events. By unifying this data with the full breadth of its enterprise data in the Snowflake Data Cloud, Strava can effectively improve its products and enhance the customer experience.


Using this modern data architecture, Strava was able to measure the performance of its product experiments with pinpoint accuracy.

With Snowplow data, we were able to measure project success through an A/B test. In our experiment, we hypothesized that our new Route Detail Page will help Strava users feel like they have enough information to take the next step with a route, resulting in increased engagement with the product.


The team measured user engagement with control and variant Route detail pages by using page views as the main metric. They also measured how easy it was for users to make route decisions through secondary metrics which included saves, downloads, shares, and Route recordings.

These metrics were tracked through Snowplow, analyzed in Snowflake, and visualized in Tableau, which together have helped Strava perform advanced digital analytics on the Routes feature. As a result, the product team can make more informed decisions, optimize the user experience, and achieve better results by leveraging data-driven insights.

Using these tools, we were able to measure the impact of the changes we made and learned that we delivered increased value to members of the Strava community who enjoy our Routes feature.


The experiment results were outstanding. The team saw a significant increase in page views of the variant Route page versus the control version. The number of Route downloads, Route records, Route saves, and Route shares per user also increased.

The combined power of Snowplow and Snowflake made this dramatic improvement possible. Snowplow’s comprehensive first-party event tracking and instrumentation provided Strava’s product team with the detailed insights they needed to accurately measure and improve their offering. At the same time, Snowflake provided the scalability to handle large volumes of behavioral data. Together, the technologies provide Strava with the insights it needs to create the ultimate active community experience.

Get Started with Digital Analytics today!

With Snowplow Digital Analytics, Powered by Snowflake, it is now easier than ever for businesses to generate impactful analytics to optimize their products.  Snowplow Digital Analytics is a Data Cloud-first alternative to Google Analytics which provides out-of-the-box visualizations without limited customer views and a lack of data ownership. Click here to find out more.

Ready to start creating rich, first-party data?

Image of the Snowplow app UI