Mixpanel vs Amplitude
With the increasing emphasis on data driven product optimization choosing an appropriate analytics platform has never been more important. At Snowball we made the transition from Mixpanel (arguably the industry leader) to the upstart Amplitude about six months ago. We haven't looked back.
A surge in Snowball downloads in February (good) leading to a proportional surge in our Mixpanel bill in March (bad) prompted the move from Mixpanel to Amplitude. Though Mixpanel defends their pricing structure with the high costs of infrastructure, we realized that as a management team we couldn’t be in a position to significantly decrease our runway in the case of unexpected growth. The Amplitude model of unlimited events was tantalizing and we moved to their systems in March.
In addition to pricing stability for our analytics needs we realized two other cost benefits. Previously we were doing a ROI calculation on every significant new event we wanted to instrument via Mixpanel. By removing this barrier we not only saved time (=money) but also enabled engineers to run experiments without a cost/benefit discussion. The second benefit was a reduction in programmer time spent optimizing how we were sending events to reduce costs; instead we are now spending that time in analysis mode, which has led to a virtuous circle for our product.
The next major benefit we've realized is direct access to our underlying data in Redshift. This not only gives us total data portability and ownership but also allows us to start building real time service dependencies on our analytics data. Specifically, we now have the capability to do things like dynamically change the size of a control / treatment group from our service based on how the respective groups are performing in real time. This kind of a tight feedback loop has long existed in ad-tech systems and is just now being enabled in other domains, with Amplitude leading the charge.
The final, and perhaps most important, benefit that we realized was a significant improvement in our analysis capabilities. This is made possible through Amplitude’s behavioral cohorts, which allows us to build groups of users based on commonly performed events (“saw 5 notifications upon startup”) and use that group to segment retention, engagement etc.
In conclusion, we’ve found Amplitude to be the best choice for an analytics platform; it’s better priced, more sophisticated, and provides a level of openness that competitors do not provide. Most importantly, it shifts the mentality of a team with a fast growing product from minimizing (analytics) cost to maximizing insights.
