This is a 5-part series
Click below to navigate to the next chapter
Throughout this series, we’ve explored how the arrival of the smartphone exploded the demand for ‘mobile-first experiences’. We’ve seen new expectations from mobile users, who expect their user journeys to be fast, smooth and personalized to their unique requirements.
We’ve identified that despite major obstacles to effective analytics on mobile (including challenges around user identification and the limitations of packaged tools), the need for robust mobile analytics is only growing in our competitive world. A reliable system for capturing and processing behavioral data on mobile is no longer a nice-to-have.
Taken together, the challenges around effective mobile analytics might make the task seem daunting, yet there’s good reason for us to take heart. Budding communities, innovative solutions and helpful tooling are landing into the mobile analytics market all the time. There are also admirable examples of forward-thinking organizations making strides in their mobile analytics – delivering rich behavioral data to equip their product teams to build better user experiences. Strava is one such example.
Strava at a glance
| Strava is the world’s most popular platform for athletes to record and share their sporting activity. Once dubbed ‘the social network for athletes’, today Strava is home to 64 million active users in over 195 countries. |
If you’ve digitally recorded a cycle, run, swim, or walk recently, there’s a strong chance you used Strava. Strava is the world’s preferred platform for athletes to record and upload their physical activities, with 40 million workouts uploaded each week.
Monitoring and maintaining a platform for so many users would be a challenge in itself. But Strava’s teams of product specialists are not satisfied merely maintaining the status quo, setting their sights instead on continuous improvement. Like the athletes they serve, Strava is constantly striving to make incremental gains in order to stay ahead of the competition.
At heart of this endeavor is behavioral data – lots of it. In fact, Strava captures and processes up to 4bn events from their app platforms in a single day – the sort of scale that would normally cause infrastructure to crash and cloud computing costs to spiral. Strava not only needed a reliable way to manage their vast behavioral data asset, but a system in place to help them surface the data to the analysts and product teams who needed it most.
A question of scale
The volume of behavioral data uploaded to Strava each day made implementing reliable analytics a huge challenge. A typical day might see anywhere from 3 to 4 billion events entering the data warehouse, while a relatively small data team was tasked with managing, cleaning and organizing the data to make it available for the wider business.
“Whether that’s a product decision or a business decision, we want employees to be able to get insights from the data, without having to be an analyst or an expert.” – David Worth, Engineering Manager at Strava
Massive, unwieldy tables of data cannot be easily handled, let alone queried and modeled down into actionable chunks. On the other hand, a constant stream of requests for engineering resources from analysts can easily turn into bottlenecks or ‘data breadlines’. Strava’s team needed to find a way to ‘democratize’ their data, putting it in the hands of analysts who could serve product teams with granular insights that would empower them to continuously iterate on Strava’s rich suite of features.
Building a self-serve data culture
Strava set about building a data stack that would enable them to make data accessible to everyone within the organization who needed it. Strava’s Engineering Manager David Worth, and Data Engineer Daniel Huang, sought a solution that would give them control of their behavioral data set, without the rigid limitations of a packaged analytics solution.
“We would not have achieved our current level of self-serve data without Snowplow. It has enabled us to democratize our data culture, significantly improving our analytics coverage and deepening our insights.” – Daniel Huang, Data Engineer at Strava
Among their list of requirements, two aspects were especially important: reliable infrastructure that could cope with the eye-watering volumes of data Strava deals with, and a scalable approach to cloud costs that wouldn’t spiral out of control.
With these requirements in mind, Strava’s data team looked to Snowplow. Snowplow’s open source framework gave Strava’s data team the flexibility to be in total control of their data. Snowplow is architected from the ground up to cope with behavioral data at scale – and with fixed, tiered pricing, costs would not skyrocket as volumes increased.
Crucially, Snowplow enabled Strava to capture their data across their platforms in a way that made sense to them. With flexible tooling, Strava’s team could choose how their data should be structured and open up the end-to-end tracking process to the analysts, without relying on support from engineers.
Mobile tracking on their terms
Strava’s product teams are equipped with analysts who support product managers with insights from behavioral data. With Snowplow, those analysts can instrument tracking where they need it, making use of Snowplow’s custom events and entities to structure their data as required.
Not only does this relieve David, Daniel and the rest of the engineering team from a tide of analytics requests, but with greater autonomy over their data capture, Strava’s product analysts can self-serve insights to support their continuous cycles of iterate, learn, improve.
“When we talk about having Snowplow managed for us, where we have a full TechOps team, not just a single person, and it’s less expensive than a full time employee, is a super easy win. I don’t want to have to bring one or two people on to manage our infrastructure when we have more impactful things we could be doing”
– David Worth, Engineering Manager at Strava
While the product managers are empowered to enhance their mobile app, Strava’s data engineers are free to focus on more meaningful projects. Snowplow BDP frees Strava from the hassle of managing their data infrastructure, so they can channel their efforts into refining their data stack and laying the groundwork for an even stronger analytics cycle.
This leaves Strava in prime position to build on their lead as the world’s athletic platform of choice, driven by a steady flow of behavioral data that informs key decisions in the organization on a daily basis.
