At Doist, we love making smart bets. Sometimes, the smartest decision isn’t to pick the biggest or shiniest tool out there but to build a small tool that does exactly what’s needed. That’s how Bitmapist came to life—a powerful, open-source cohort analytics library that’s been quietly driving smarter decisions and saving us millions of dollars.
Why We Built Bitmapist
Several years ago, we faced a common startup challenge: we needed robust cohort analytics to gain a deeper understanding of how people use Todoist and Twist . We needed answers to specific, critical questions about user behavior, such as:
How many users adopted our newest feature this week?
What percentage of users who were active last month are active today?
Which features drive long-term user retention?
At our scale, commercial analytics platforms often charge astronomical prices or provide limited flexibility for custom queries. True to our philosophy that tools should adapt to our needs (and budget), we opted to build a tailored solution instead of settling or overspending.
Why Bitmaps?
The secret behind Bitmapist’s performance lies in its name—bitmaps. Using Redis bitmaps , Bitmapist stores massive amounts of event data with minimal memory usage, allowing us to track millions of user interactions and billions of events efficiently. This approach enables real-time analytics queries, returning cohort data in milliseconds.
Here are a few powerful questions this approach effortlessly answers for us:
... continue reading