Source: Longitudinal user enrollment and retention data from the piano learning app Skoove.

Data Range: Monthly start-date cohorts tracked over a six-month duration from January 2021 to December 2024.

Methodology: This is a longitudinal cohort analysis. We grouped 1.1 million users by their enrollment month and tracked the retention of each specific group at monthly intervals. To normalize for year-specific anomalies, monthly retention rates were averaged across the four-year study period. The percentages shown represent the relative likelihood of persistence compared to the December cohort, which served as the lowest annual baseline (0%).

Tools: Data extraction via Mixpanel; analysis performed using Python/Pandas; visualization designed with Adobe Illustrator / Figma.

Key Insight: The period of highest initial motivation (the New Year "Fresh Start") correlates with the lowest rates of sustained habit formation. Conversely, learners who begin in April-June are over 60% more likely to stick with the habit for six months compared to December starters.

Posted by DataPulse-Research

Share.

6 Comments

  1. schoolforapples on

    I get why people who start in January are more prone to quit, but why are people who start in April-June more likely to continue?

  2. I suspect that the “percentages in relation to December” hide some some absolutely minuscule random fluctuations in the data. Would be good to know the absolute numbers and also the sample size. Until I see that, I call BS on this one!

  3. Zestyclose-Barber-24 on

    Thats such a weird correlation, and its also very strong. Who were those people who were tested?

  4. So this is to be read as the excess rate over and above December’s retention rate? I want to read this as saying 0% of December students survive 6 months, but that can’t be right.

  5. I started 10 days ago. This is not looking good 😂

    Nah, I’d expect January to be the worst one, I’m actually surprised about December.

    That said, I might not sign up at something like Skoove for some weeks, so maybe I’m ok.