Skip to content

Prioritisation & strategy · Hard

How would you decide between investing in acquiring new users and improving retention of existing ones?

Short answer

Look at where value leaks most. If many new users churn early, acquiring more fills a leaky bucket, so retention usually comes first. If retention curves flatten at a healthy level, growth investment pays off. Use cohort retention curves, the cost of acquisition and the stage of the product to decide, and revisit the balance each planning cycle.

Detailed answer

Framework

  1. Read cohort retention curves. Does each cohort's retention flatten out, or keep declining towards zero?
  2. Compare economics. Roughly, what does it cost to acquire a user, and how much value does a retained user create over time?
  3. Consider stage. Early products need evidence of retention before scaling; mature products may be limited by reach.
  4. Identify the biggest lever and its confidence level.
  5. Decide the split, not just a single winner.

Worked example

Imagine a language-learning app.

  • Cohorts show a steep drop after the first week, and the curves keep falling without flattening.
  • Marketing has been increasing spend, and new sign-ups are rising.

Interpretation: the product has not yet found a durable habit for most users. More acquisition increases sign-ups but not the long-term active base. Prioritise retention: understand why users leave after week one (interviews, funnel analysis on lesson completion), and test changes such as better first-lesson difficulty or reminders users choose themselves.

Now imagine the curves flatten at a stable level after a month, and qualitative feedback is strong, but sign-ups have plateaued. Here, a growth push (referrals, new channels, localisation) is more valuable.

In practice, recommend a split, for example most of the team's effort on retention with a small growth experiment running alongside, and define the retention signal that would trigger a shift in investment.

Common pitfalls

  • Choosing based on which sounds more strategic.
  • Ignoring cohort data and looking only at total active users, which can rise even while retention worsens.
  • Treating it as all-or-nothing.