Solution Finder
What is your customer trying to achieve? Pick the problem closest to what they actually said — this translates it into the right methodology, data requirements, and Lab.
You said: “Determine which customers are worth acquiring, not just cheapest to acquire”
Customer Lifetime Value Modelling
AudienceWhy this fits
Customer Lifetime Value Modelling compares acquisition cost against projected long-term value, not just first-purchase ROAS.
What it answers
Beyond the first purchase, what is a customer likely to be worth over time, and does that change which acquisition channels or segments are actually the best investment?
How it works
Historical customer purchase and retention behaviour is used to project forward value (e.g. 12- or 24-month predicted value), which is then compared against acquisition cost to reveal which acquisition strategies are genuinely profitable.
Data required
- Customer acquisition cost by channel or campaign
- First purchase value
- Purchase frequency and recency over time
- Retention/churn indicators
- Sufficient historical depth to observe repeat behaviour
Discovery questions
How much purchase history exists per customer cohort?
CLV projections need enough repeat-purchase history to be credible rather than purely assumed.
Is CAC tracked accurately by channel or campaign?
Without accurate CAC, comparing predicted value against acquisition cost is meaningless.
What decision will this CLV analysis actually inform — budget allocation, channel mix, retention investment?
Shapes whether the model needs to be channel-level, segment-level, or individual-level.
Implementation approaches
Vendors
Risks & limitations
Risks
- Short historical windows lead to unreliable long-horizon projections
- CLV models can be misused to justify overspending on acquisition if projections are too optimistic
Limitations
- Predicted value is inherently uncertain, especially for newer customer cohorts
- Category and business-model shifts can invalidate historical retention patterns