Business Context

You’re the analytics lead for NovaCard, a consumer fintech app (similar to Cash App + debit card) with 6.5M MAUs, 1.8M weekly transacting users, and ~$420M annual gross profit. NovaCard monetizes via interchange plus a $4.99/month “Instant Transfers” fee that enables immediate cash-outs to external bank accounts.

Last month, Finance increased the Instant Transfers fee from $4.99 → $6.49 for new subscribers immediately and for existing subscribers at renewal. The change was rolled out globally over 48 hours with no A/B test (urgent margin pressure). Two weeks later, the Head of Growth reports that “retention is down” based on a dashboard showing a drop in 30-day active rate. Finance argues the fee hike increased revenue per user and that any retention change is seasonal.

You have one week to provide a recommendation to the GM: keep the new price, roll it back, or introduce mitigations (e.g., grandfathering, discounts, feature bundles). You must quantify whether the fee increase is hurting customer retention, identify which segments are impacted, and estimate the net business impact (revenue vs churn).

Metric Scenario

Key complication: NovaCard users can be active without paying the fee (they can still use the debit card and standard transfers). So “app retention” may move for reasons unrelated to the fee. The fee is tied specifically to Instant Transfers subscribers, and the price change only affects users who are exposed (new subs and renewing subs).

Stakeholders ask:

• Did the fee increase cause higher churn among Instant Transfers subscribers?
• Is any churn concentrated in specific cohorts (low-income, high-frequency cash-out users, certain countries)?
• Are users downgrading to free and still transacting (good) vs leaving NovaCard entirely (bad)?

Data Available

Requirements

1. Define the primary retention metric you will use to evaluate impact of the fee increase (be explicit about numerator/denominator, time window, and who is “at risk”).
2. Propose a measurement strategy to isolate the fee effect given the lack of a clean experiment (e.g., cohort-based analysis around renewal dates, difference-in-differences using non-exposed users, or exposure-based comparisons).
3. Provide a decomposition plan to find the driver(s): renewal churn vs voluntary cancellation vs payment failures; downgrade-to-free vs full platform churn.
4. Specify guardrail metrics and how you’ll interpret trade-offs (e.g., revenue per subscriber up but retention down).
5. Recommend a decision framework (what magnitude of retention decline would justify rollback) and 3 concrete actions you’d take depending on findings.

Constraints:

• Price change is already live; you cannot re-run the past.
• Some users renew on different billing cycles (monthly/annual).
• Countries have different taxes and displayed prices; FX moved ~3% during the period.