Suppose a randomised, fully observed experiment gives these point estimates:
| Pre-assignment segment | Retained without offer | Retained with offer | Estimated retention difference |
|---|---|---|---|
| Higher risk | 70 of 100 | 80 of 100 | +10 percentage points |
| Lower risk | 90 of 100 | 90 of 100 | 0 percentage points |
With equal segment weights, the estimated average retention effect is (0.10 + 0) / 2 = 0.05.
Assume an extra retained account contributes 100 units and an offer costs 5 per assigned account,
with no other costs or interference between customers. The estimated extra contribution is:
- Higher risk:
0.10 × 100 - 5 = +5per assigned account. - Lower risk:
0 × 100 - 5 = -5per assigned account.
These are point estimates, not promised gains. Report intervals at the assignment unit, check guardrails, and confirm a policy's value on data you did not use to choose it. Real economics may need contribution as the direct outcome, not retention times a constant. The churn guide explains when high risk and benefit from the action disagree.
Try it yourself. An offer costs 12 instead of 5. What do the point estimates recommend?
Hint and explained answer
The two estimated gains become 10 - 12 = -2 and 0 - 12 = -12. Under these assumptions, keep the
no-contact baseline for both segments. Better retention alone does not make an action profitable.
A capacity limit or uncertain effect can change the decision further.