In plain words
A cohort simulation plays a plan out customer by customer. Each simulated customer buys, then renews or leaves at each step according to the renewal rates. Do that for a whole cohort, many times over, and you get a spread of results.
A single projection gives the average. The spread shows how far a real cohort could land from it, both because a few hundred customers never behave exactly like the average and because the inputs themselves are estimates. The result is a band, and a chance of reaching the goal.
Formula
- Run the cohort many times, each with its own random draws; report the median and the range that holds 80% of runs
Worked example
200 runs of a 200-subscriber cohort, with the acquisition cost uncertain by 20% and the renewal rates by 10%. After twelve months the median run has earned $21.82 per subscriber, and eight runs in ten fall between $9.23 and $30.68.
98.5% of runs pay back within the year. The median does so in month 5, and the unlucky edge of the band in month 6.
The example order
A $40.00 bag of coffee, shipped free, every cost counted
2 declared kinds cost nothing on this option
| Step | Grain | Amount |
|---|---|---|
| Revenue | – | $40.00 |
| Cost of goods | per unit | −$12.00 |
| Pick and pack | per order | −$1.18 |
| Packaging | per parcel | −$0.50 |
| Shipping label | per parcel | −$7.50 |
| Card fee per order | per order | −$0.30 |
| Refunds | percent | −$2.00 |
| Card processing | percent | −$1.16 |
| Contribution margin | – | $15.36 |
| 2 kinds costing nothing | – | $0.00 |
Computed by the margin engine in Offer Suite's own code, in whole cents, from a $40.00 bag of coffee with a $12.00 unit cost.
A simulation is not more accurate than its inputs
It turns the uncertainty you describe into a range. It cannot discover uncertainty you left out. If the renewal rates are guesses, the band is a careful picture of those guesses.
Using it
Read the bad edge of the band, not the middle. A plan that pays back early on average and months later when unlucky needs cash for the later date. Use a fixed seed so the same inputs always give the same answer, and widen the uncertainty on any input you have not measured.
In Offer Suite
For a subscription offer, Offer Suite's simulator runs seeded cohorts and reports an 80% band with the chance of payback. Renewal rates are supplied by the brand: they are not yet read from its own orders, and the simulator refuses to guess them.
Source: Offer Suite's offer simulator.