Customer lifetime value prediction by acquisition cohort
The short answer
Base a customer lifetime value prediction for new customers on how last year's equivalent cohort spent. Take the average 12-month spend of customers acquired in the same quarter last year, adjust it for differences in this quarter's mix of channels and first products, and multiply by the new cohort's size.
Why it matters
Finance needs to know what this quarter's acquisition spend will return, and first-order revenue alone understates it. A cohort-based forecast puts a dollar figure on the next 12 months and shows early whether new customers are tracking above or below last year.
The question to ask
Copy this into Lexi as written.
How to work it out
1. Build last year's curve. For customers acquired in the matching quarter last year, total their spend in each month after the first order, out to 12 months.
2. Adjust for mix. Split last year's curve by acquisition channel and first-purchase category, then reweight it to match this quarter's new customers.
3. Track and correct. Compare the new cohort's actual spend with the curve each month, and update the forecast once three months of data are in.
What the answer looks like
Illustrative output on sample data
Illustrative example: how Lexi answers this for a beauty retailer that acquired 18,000 customers this quarter. It applies last year's cohort curve, adjusted for this quarter's channel mix.
- Last year's cohort: $142 average 12-month spend. $158 from search and email, $94 from a marketplace promotion.
- This quarter: 40% came through the marketplace, against 25% last year.
- Mix-adjusted average: (60% × $158) + (40% × $94) = $132.40.
- Forecast: 18,000 × $132.40 = $2.38 million, against $2.56 million unadjusted.
What the data shows
The shift toward promotion-led customers takes about $170,000 off the 12-month forecast.
What it means
Acquisition volume is up, but the customers are worth less. A forecast built on the average alone would overstate next year.
What to do
Plan on $2.38 million and recheck after the first three months of repeat orders.
If this quarter's marketplace customers buy again faster than last year's, the gap narrows. Would you like the cohort curve by channel?
Frequently asked questions
What is the difference between historical and predicted customer lifetime value?
Historical lifetime value adds up what a customer has already spent. Predicted lifetime value estimates what they will spend over a future period, such as the next 12 months, based on how similar customers behaved. Finance uses the predicted figure to judge acquisition spend, since historical value says nothing about customers acquired last week.
Why can last year's cohort mislead a forecast?
Cohorts change shape when acquisition changes. A new channel, a deeper welcome offer or a different first product can bring in customers who behave unlike last year's. Reweight the old curve to the new mix, and check early repeat purchases before trusting the full 12-month figure.
How can I improve customer lifetime value for a new cohort?
Focus on the second purchase, since every customer who makes one carries forward into later months of the curve. Time a relevant follow-up to the usual gap between first and second orders, steer acquisition toward channels whose customers repeat, and cut welcome discounts that attract one-time buyers.