Manual Retention and Repeat Purchase Impact Calculator
Retention and repeat purchase are related but not interchangeable. A customer may remain active without buying again in your chosen window, and one customer may place several repeat orders. Calculate the transactional scenario separately from the relationship-retention rate.
Use the worksheet in three steps: choose one fully matured cohort, copy the repeat-purchase formula with Finance-approved contribution and cost, then compare one conservative, one base, and one aggressive rate. Keep the separate retention-rate view for relationship health; do not mix its denominator into the order model.
Repeat-purchase impact inputs
| Symbol | Input | Unit | Valid range |
|---|---|---|---|
N |
Eligible, fully matured starting cohort | customers | integer ≥ 0 |
r0, r1 |
Baseline and scenario repeat-customer rate | decimal | 0–1 |
o |
Qualifying repeat orders per repeat customer | orders/customer | ≥ 1 |
M |
Contribution per qualifying repeat order | currency/order | ≥ 0 |
q |
Realization factor for returns, cancellation, quality and rework | decimal | 0–1 |
C |
Incremental program cost over the window | currency | ≥ 0 |
Every input is user-supplied with no default. Analytics owns N, r0, r1, o; Finance owns M, q, C. Copy the inputs/result into an approved cohort worksheet; the page does not save or export.
Shopify discusses repeat purchasing as part of retention, while Salesforce publishes a period-based customer retention formula. Neither provides a universal “good” rate for your cohort. Review Shopify and Salesforce.
Incremental repeat customers = N × (r1 − r0)
Incremental repeat orders = Incremental repeat customers × o
Realized contribution = Incremental repeat orders × M × q
Net impact = Realized contribution − C
Use only cohorts that have completed the observation window. Never compare a 180-day mature January cohort with an April cohort that has only 60 days of opportunity.
No finite break-even lift exists when N×o×M×q=0; report undefined. Missing inputs produce no result. If r0 + required lift > 1, label the threshold infeasible in this model.
Method and sources reviewed: 2026-08-16.
Worked example
Northstar uses N=2,000, r0=0.22, o=1.4, M=$45, q=0.85, and C=$6,000 for a 180-day window.
| Scenario | r1 |
Incremental customers | Incremental orders |
|---|---|---|---|
| Conservative | 23% | 20 | 28 |
| Base | 26% | 80 | 112 |
| Aggressive | 30% | 160 | 224 |
| Scenario | Realized contribution | Net impact |
|---|---|---|
| Conservative | $1,071 | -$4,929 |
| Base | $4,284 | -$1,716 |
| Aggressive | $8,568 | $2,568 |
For a stacked view:
- Conservative:
r1=23%; 20 incremental customers; 28 orders; $1,071 realized contribution; -$4,929 net. - Base:
r1=26%; 80 incremental customers; 112 orders; $4,284 realized contribution; -$1,716 net. - Aggressive:
r1=30%; 160 incremental customers; 224 orders; $8,568 realized contribution; $2,568 net.
Base: 2,000×(0.26−0.22)=80; 80×1.4=112; 112×$45×0.85=$4,284; net $4,284−$6,000=−$1,716.
Break-even repeat-rate lift is C ÷ (N×o×M×q) = 6,000 ÷ (2,000×1.4×45×0.85) = 5.60 percentage points. The model does not claim the program can cause that lift.
Separate retention view
For an account/relationship business, calculate:
Retention rate = (customers at end − new customers during period) ÷ customers at start
Do not feed that percentage into the repeat-order formula unless the customer and order definitions align. Report renewals, pauses, returns, win-backs and identity merges consistently.
Interpret the result and its limits
| Condition | Meaning | Action |
|---|---|---|
| Cohort not mature | Rate is right-censored | Wait or use a shorter predeclared window |
Positive only with q=1 |
Returns/quality decide the case | Use finance-approved realized contribution |
| Repeat rate rises but margin falls | Volume and economics conflict | Investigate discount and product mix |
Reducing q from 0.85 to 0.60 makes the aggressive net 224×45×0.60−6,000=$48; a small further change reverses it. Always stress-test contribution and cost, not only the rate.
The page stores no input and gives no benchmark/default. Values must use one currency, cohort, maturity window and identity rule. The scenario cannot prove causality, future retention, customer satisfaction, or Easy AI performance.
Frequently asked questions
Use these answers for cohort-accounting questions that remain after the repeat-purchase and retention definitions are locked.
How should returns recorded after the maturity window be handled? Apply one predeclared cohort policy consistently through q, then refresh the dated worksheet when late returns materially change the result.
Can subscription renewals use the repeat-order formula? Only when the renewal, order, and customer definitions fit the formula; otherwise model renewal separately.
How should win-back customers be classified? Declare the identity and cohort rule before calculation and apply it unchanged to baseline and scenario.
What to do next
Use the repeat-purchase definition to lock cohort and maturity rules.
Have Analytics lock cohort and identity rules, Finance approve contribution and cost, and Customer Operations define complaints/returns as guardrails. Pilot one lifecycle action and wait for the declared maturity window before deciding.