Customer lifetime value can rise even when customers have not become more valuable. An older customer group may simply have had more time to place repeat orders. Without matched observation windows, calendar time can look like improved retention.
That makes customer lifetime value analysis easy to misread in a young or fast-changing product business. A skincare customer may reorder every eight weeks, while a gift buyer may not return for a year. Ask not only “What is our LTV?” but “Which customers, measured for how long, using which definition of value?”
Start with a fixed-window value, not a guessed lifetime
The familiar LTV formula multiplies average order value, purchase frequency, and customer lifespan. When history is limited, the lifespan assumption often does most of the work.
A safer starting point is a fixed-window measure:
180-day customer value = total value generated by a customer cohort during its first 180 days ÷ customers in that cohort
“Value” might mean net revenue, gross profit, or contribution after variable order costs. Pick one and label it. Net revenue answers a sales question. Contribution is usually more useful for deciding how much acquisition or retention spending the relationship can support.
A fixed window does not claim to know the full future. It creates a fair checkpoint that can be repeated for several customer groups.
Build cohorts around the first purchase
A cohort is a group of customers who share a starting point. For this cohort analysis, group buyers by the month or quarter of their first completed purchase. Google Analytics describes cohorts as users connected by a common characteristic and lets analysts examine their behavior over time. Your order system can support a simpler version in a spreadsheet.
For each cohort, record:
- first-purchase period;
- number of verified new customers;
- first product or product family;
- acquisition source when reliably known;
- discount status;
- net revenue, refunds, and variable costs;
- cumulative value at 30, 60, 90, and 180 days;
- percentage that placed a second order.
Do not mix first-time and returning customers in the starting count. Remove canceled orders and document how guest checkouts, duplicate emails, and merged records are handled.
Match the observation window before comparing results
Suppose a home-fragrance business reviews two groups of 100 first-time customers:
| Cohort | Time observed | Cumulative contribution | Contribution per customer |
|---|---|---|---|
| January | 180 days | $6,400 | $64 |
| April | 90 days | $3,900 | $39 |
At first glance, January appears far stronger. But January had twice as long to collect repeat orders. The comparison confuses maturity with performance.
Now compare both groups at day 90. January had produced $4,100, or $41 per customer. April produced $3,900, or $39. The matched-window difference is about 5.1%, not the dramatic gap implied by $64 versus $39.
That smaller difference may still reflect product mix, discounting, availability, seasonality, or customer experience. It is now a useful question rather than a calendar artifact.
Separate observed value from projected LTV
Keep three figures distinct:
| Measure | What it says | Best use |
|---|---|---|
| Observed 90-day value | What the cohort actually generated in 90 days | Comparing newer groups quickly |
| Observed 180-day value | What a mature cohort actually generated in 180 days | Retention and payback review |
| Projected LTV | What the business estimates over a longer relationship | Planning scenarios with explicit assumptions |
A projection can extend a mature pattern, but it should show the customer group, forecast horizon, expected repeat orders, margin basis, and uncertainty.
Google Analytics’ user-lifetime documentation also distinguishes initial, most recent, and lifetime interactions and notes that reporting identity affects the data. That is a useful reminder: a dashboard value is shaped by who the system recognizes as the same user. Compare the analytics view with completed-order records before treating it as financial truth.
Read the repeat-order curve, not only the final number
Two cohorts can reach the same 180-day customer value by very different paths.
One may recover most value on the first order. Another may start with a discounted trial, then place three profitable replenishment orders. Their totals match, but cash timing, inventory needs, and retention opportunities differ.
Plot cumulative contribution per customer at the same checkpoints. Then add repeat customers as a separate line: the percentage of the original cohort that has placed another completed order by each checkpoint.
A flattening curve may be normal for a seasonal or long-lasting product. For a frequently replenished product, it may signal stockouts, an inconvenient reorder path, a formula or packaging change, weak follow-up timing, or customers who did not find enough value to return. The guide to repeat purchase rate by customer cohort helps examine that behavior without treating every non-returning buyer as lost too soon.
Segment only where the decision changes
One blended LTV is easy to report and hard to act on. Segmenting can reveal meaningful differences, but dozens of tiny groups create noise.
Start with one dimension tied to a decision:
- first product, when deciding which product introduces the healthiest relationship;
- acquisition source, when reviewing marketing limits;
- full-price versus discounted first order, when evaluating promotions;
- subscription versus non-subscription, when planning replenishment;
- season, when gift buying changes repeat behavior.
Give every segment the same maturity window. Label small groups directional and wait for more evidence.
When acquisition spending is the decision, connect the matched cohort to the same cost scope. The LTV-to-CAC ratio guide explains why contribution-based value is more useful than revenue alone, while the customer acquisition cost guide shows how to count genuinely new customers and included costs consistently.
Run a monthly maturity review
Create one table with cohort rows and 30-, 60-, 90-, and 180-day columns. Each month:
- add the newest first-purchase cohort;
- update older cohorts only through the checkpoint they have reached;
- compare groups at equal maturity;
- note changes in price, discounts, product mix, availability, fulfillment, or tracking;
- investigate one material difference instead of reacting to every movement.
Keep the original customer lifetime value definition and formula beside the worksheet. The new table does not replace LTV. It makes the estimate easier to challenge and improve.
Practical takeaway
Choose one recent quarter of first-time customers. Calculate observed contribution per customer at 30, 60, 90, and 180 days. Compare only cohorts that have reached the same checkpoint, then separate the repeat-purchase curve from the dollar-value curve.
If LTV rises, ask whether customers ordered more, margins improved, or the measurement window simply got longer. A useful lifetime-value report makes time visible. That prevents an older cohort from winning by age and gives the business a clearer view of what actually changed.
Explore more Kerno Resources for plain-language guides to customer economics, marketing decisions, and healthier growth for physical-product businesses.
Frequently asked questions
How long should an LTV observation window be?
Use a window long enough to include a normal reorder opportunity for the product, then apply it consistently. Many businesses benefit from showing several checkpoints rather than choosing one permanent window.
Should refunds and discounts reduce customer value?
Yes. Start with completed net sales after refunds and discounts. For contribution-based value, also subtract the variable product and order costs included in your written definition.
Can a new business calculate customer lifetime value?
It can calculate fixed-window customer value from observed orders. A full-lifetime estimate should be labeled as a projection because limited history cannot prove a long customer lifespan.
Why can analytics and order-system LTV disagree?
They may identify customers differently, cover different dates, include different revenue events, or handle refunds and guest purchases differently. Document both definitions before comparing them.
