ENTR
Product SenseJunior≈ 30 min

Lumio — AI Filters

D30 retention dropped from 45% to 31% three weeks after launching 20 AI photo filters.

The scenario

Lumio is a mobile photo-editing app. Three weeks ago you launched 20 AI-powered filters as a headline feature. Initial installs spiked 30%, but D30 retention has fallen sharply. App Store reviews mentioning "filters" have tripled, sentiment is mixed.

Constraint

You have a 3-day sprint with 1 designer to act on this.

Your task

Diagnose the retention drop and propose one fix you can ship this sprint.

1. Clarify: What do you need to know before diagnosing?

Pick the 3 questions you'd ask first.

Model approach (strong picks)

  • Strong pick
    Is the retention drop isolated to users who used the AI filters, or is it app-wide?
  • Strong pick
    What do the App Store reviews say — are users complaining about quality, discoverability, or something else?
  • Strong pick
    Did retention start dropping immediately at launch, or after a delay?

2. Hypothesize: What are the most likely causes of the drop?

Pick the 3 hypotheses you'd investigate first.

Model approach (strong picks)

  • Strong pick
    Filter quality is inconsistent — results look bad on certain photo types, killing trust.
  • Strong pick
    Too many filters with no guidance — users are overwhelmed and give up.
  • Strong pick
    The filters are buried in the UI — most users never find them after the first session.

3. Investigate: What do you do in the next 48 hours?

Pick the 3 highest-leverage moves.

Model approach (strong picks)

  • Strong pick
    Segment D30 retention by users who used ≥1 filter vs. those who never used one.
  • Strong pick
    Read the App Store reviews mentioning filters — tag themes (quality, UX, performance).
  • Strong pick
    Check the filter usage funnel: how many users open the filter panel, apply a filter, save a photo.

4. Propose: What do you ship this sprint?

Pick the one fix you'd commit to.

Model approach (strong picks)

  • Strong pick
    Reduce visible filters from 20 to 5 curated "Starter" picks, with a "See all" option — lower overwhelm, preserve depth.
  • Strong pick
    Add a short onboarding card on first open that highlights the top 3 filters with example photos.

Related questions

Solve this case study