Sign inJoin
5
3 Comments

Is this cohort analysis good or bad?

by Fiona Whittington

Since our official product launch in November, The Bit has been able to sign up over 200 users to our platform. I would love any help interpreting our cohort analysis, so I can better our user's journey using our platform. In your experience, is the analysis below concerning or normal?

Fiona Whittington's avatarFiona Whittington
on December 13, 2019
Say something nice to fionaw…
Post Comment
  1. 2

    Hey @fionaw look at you bein data-driven! 🤜🤛

    Ok so a couple of thoughts about analyzing your dataset:

    • Outliers: what did you do for onboarding the Nov10 cohort, and for re-engaging the Oct27th cohort 4 weeks later? What did you do for the Oct27 cohort that increased their engagement week-over-week? And likewise, what changed for the future cohorts?
    • Cohort filtering: it's tough to look at users all-in. How about filtering folks for specific marketing copy, or demographic (gender/age) - do the engagement percentages vary among the filtered cohorts?

    You're on a really good track with the data your team is collecting. Might make sense to pick the filtered-cohort that had the best engagement rate and do the same thing every week for a few weeks - you goal is to get a consistent cohort analysis (ie. week1 always 15% +/- something, week2 always 10% +/- something) because then you can work with growth person to predictably explode user numbers

    vdizzle's avatar vdizzle
    ·
    7 years ago
    ·
    Reply
    1. 1

      This is great, thank you!

      Fiona Whittington's avatar Fiona Whittington
      ·
      7 years ago
      ·
      Reply
  2. 1

    Cohort analysis can be a very powerful tool to understand user retention and drop offs.

    In this case, I think it depends a lot on how many users you've gotten so far, so the results are statistically significant.

    JaySignorello's avatar JaySignorello
    ·
    7 years ago
    ·
    Reply