Trello loves data! Here at Trello we try to incorporate data into our decision making across the company, from marketing campaigns to product roadmaps. Our goal is to delight users, and we do this using data.
What does that look like?
Our Trello data science team is lean and mean, sitting outside any one team so that we can get the best view of top data needs and priorities across the company. Trello data comes from a lot of different places…MongoDB, Snowplow, Salesforce, our in-house billing platform, etc. Our data engineers load all our compliant data into Redshift (a lot of stuff we don’t collect to respect user privacy), and that’s where the magic happens. Redshift has been great for storing and querying data from millions and millions of users – we have a lot to work with!
Because we love data, we spend a lot of time thinking about the best tools for the job, and how to scale data accessibility as the company grows. We use Mode Analytics for the bulk of our reporting needs. With the ability to build on top of SQL queries with HTML, Python, and Liquid we can make reports for our data consumers to adapt to their purposes, and this cuts down on manual work and repetition tremendously.
We try to balance our hands-on work analyzing data for strategic initiatives with dedicated time on projects ensuring the continued reliability and accessibility of our data (maintaining our database so we can continue to query things quickly, adding tracking and building pipelines to get new data, writing up documentation and resources for data consumers).
On a typical day, we might work on a variety of projects such as:
- Analyze usage and impact of a new feature like Reactions or the Power-Ups directory
- Break down revenue by country to help direct internationalization efforts
- Work with engineers to set up tracking on a missing part of the mobile sign-up funnel
- Visualize team and board networks to better understand collaboration at Trello
- Work with the experimentation team and the user research team to combine our data powers
- And more!
For more reading on data at Trello, check out this blog post from our experimentation team on how we prioritize testing!