Decision Framing
Decision Framing
Replace the topic with a choice
“Analyse late deliveries” names a topic. HarborMart must choose among actions:
- accept the requested window;
- offer a later window;
- reroute fulfilment;
- add temporary capacity;
- leave the current process unchanged.
An analysis is decision-ready only when at least two feasible alternatives exist.
A six-part decision statement
At checkout, the fulfilment manager chooses an available delivery promise for one proposed order, to maximise expected contribution after service-failure cost, subject to capacity, safety and customer-commitment constraints.
| Part | Test question | Failure mode |
|---|---|---|
| owner | who is authorised to act? | “the business” owns nothing |
| action | what can change now? | predicting an immutable outcome |
| unit | one customer, order, item or day? | mixed denominators |
| horizon | when is the outcome realised? | target observed too late |
| objective | which consequences count? | optimising clicks while harming margin |
| constraints | what cannot be violated? | infeasible recommendation |
Worked decision table
Suppose an order yields £18 contribution if on time. A late delivery loses £7 through refunds and future attrition. Offering a later slot preserves £15 contribution and reduces late risk.
| Action | Late probability | Expected value |
|---|---|---|
| accept requested slot | 0.35 | |
| offer later slot | 0.08 |
Under only these consequences, accepting has higher expected value. If a late event instead costs £15, the values become £12.75 and £13.80, so the later slot wins. The prediction did not change; the decision changed because the loss did.
Threshold from consequences
Let be late probability, on-time contribution for accepting, late cost, and the value of a safer alternative. Accept when
Therefore the maximum acceptable risk is
With , and , . A universal 0.5 classification threshold would be economically arbitrary.
Prediction, intervention and policy
| Question | Object needed |
|---|---|
| Which orders will be late? | |
| Would adding a picker prevent lateness? | |
| Where should three picker-hours go? | effects plus costs and capacity constraints |
A risk model can prioritise review. It cannot by itself estimate the effect of the review or allocate scarce staff.
Value of information
Information is valuable only when it can change an action enough to improve expected outcomes.
If both a 10% and a 30% late-risk estimate lead to accepting the order, refining the estimate between those values has no immediate decision value. Near the 20% threshold, a modest improvement in calibration may matter greatly.
This is why model error should later be translated into decision regret, not reported only as RMSE or AUC.
Quick check
A campaign team asks, “Which customers are likely to churn?” What three facts are still missing before modelling?
Answer
Foundation
- Shmueli’s To Explain or to Predict? clarifies why predictive and explanatory goals imply different modelling choices.
- Kleinberg et al.’s Human Decisions and Machine Predictions shows why prediction must be embedded in payoffs and counterfactual decision evaluation.