No effect and no adequate evaluation differ
An inconclusive evaluation does not prove the program is ineffective. It may reflect poor measurement, limited data, or incomplete delivery. Conversely, years of operation do not establish that the program works.
Clarify which uncertainty is driving the decision.
Questions for a structured review
- What benefit is demonstrated, for whom, and with what uncertainty?
- What costs or burdens are occurring?
- Could a feasible change address the main problem?
- What alternatives are available?
- What happens to current participants if the service changes?
Do not make the transition invisible
Closing enrollment, referring participants elsewhere, preserving records, and communicating timelines all require ownership. A replacement service may have eligibility or capacity limits, so “refer elsewhere” is not a complete plan.
Review criteria can be set early
Before launch, identify conditions that would prompt reassessment. This makes later decisions less dependent on sunk costs or attachment to the original design.
Stopping an ineffective activity can free resources for something more useful. Continuing a valuable activity can also be justified despite imperfect evidence. The decision is stronger when the reasons, uncertainties, and consequences are explicit rather than hidden behind a simple success-or-failure label.
