Here at Mixpanel, we pride ourselves on being data-informed. What that means is that we use data to inform our decisions but don’t mindlessly defer all of our thinking over to it. In the case of assessing what content performed well, and how that affects our strategy going forward, there’s a certain amount of art that goes with the science. For example, if a post about how Mixpanel customers can make sure they’re complying with GDPR is widely read, that doesn’t mean we need to write more about GDPR necessarily, but perhaps it does mean that providing useful, timely information is something we should prioritize.
These were some of our top posts from the last year. What did we learn from them? In general, when we focus on our areas of expertise or provide original data and insights, our readership responds. These are the posts that your fellow Signal readers elevated above the rest.
In this post, we took over 4.6 billion email sends and did what we do best here at Mixpanel: we explored the data. It gave us the answers to some questions about email campaigns that every marketer is interested in and helped us craft the perfect post title. (It’s not a perfect title. As the data shows, context matters).
Do high-volume campaigns perform better than more targeted ones? Do longer or shorter email subject lines perform better? What’s the best day of the week to send an email?
Vijay Jayaram from our Infrastructure team wrote this take about a less-common (for now) use case for Mixpanel: using it to measure data costs. This allows our internal teams to put a finer point on how much each individual customer is costing us with their data usage, which has allowed us to introduce a more equitable pricing structure that better serves our customers.
A quick and easy guide to understanding the various ways to slice and/or dice your customers for more powerful analysis.
Aniruddha from engineering takes a look in Mixpanel at the data our engineers’ commits are generating which show which languages we have evolved to use more of over time, and whether or not managers produce more code than individual contributors. (Obviously they do not).
Nothing quite warms our hearts here at Mixpanel quite like bringing analytical rigor to the task of helping people find love (or, as they call it over at Hinge, “good churn.”) This is the story of how Hinge looked at their data in Mixpanel and decided to move away from the swiping model, creating an “overnight” 20% improvement in new user retention.
For this and more content in the coming years, fill out the form below to subscribe to The Signal, and check out some of our other work from this past year including Data & the Modern CMO, as well as Product Benchmarks Reports on Media & Entertainment, Financial Services, and Retail & E-commerce.