Skip to article

GapBridge Institute

← All articles

GapBridge Institute Articles

How to turn evaluation findings into actions someone owns

Translate evidence into a short decision meeting, a named owner, a realistic next step and a review date.

Three participants work together on laptops around a workshop table.

An evaluation can be thorough and still end with nobody sure what to do on Monday. Plan for use before writing the final report. Who has authority to change the programme? Which decision is coming up? What evidence will be good enough to act, and what would justify collecting more? Ask the people who will carry out the work as well as the people who approve it.

Keep the finding separate from the decision

Suppose a fictional outreach programme finds that some site reports arrive late more often when they require two approvals. That is a finding, not proof that the extra approval caused the delay. The sites might also be handling more work, or their approval processes might operate differently. Put the finding and those plausible alternative explanations on the same page. Then present the options: investigate further, try a limited change or keep the current process. Give the team enough context to compare them without relitigating the entire evaluation.

Use the meeting to agree a next step

Invite the people who understand the workflow and those who can authorise a change. Start with the decision rather than a summary of every chart. In the fictional example, the group might test a single approval for a defined set of low-risk reports, provided existing safeguards allow it. It might instead need to examine where delays occur before changing the process. A decision to wait is useful when people agree what further evidence is needed and when they will revisit it.

Record who will act and how you will know

Use a short action log. Record the finding, the agreed step, one accountable owner, what support they need, and a review date. For a pilot, track both the intended improvement and a sign of unintended harm. In the report approval example, the team could compare on-time submission with the number of returned reports. Faster reporting will not be much of a gain if fewer errors are caught. Agree in advance what would justify keeping, revising or stopping the change.

Close the loop by telling contributors what the team decided and what remains uncertain. A detailed report can preserve methods and limitations; a short decision note should make the next step easy to find. At the review date, look at what happened and update the log. That makes the evaluation useful beyond the meeting where it was presented.

Further reading: CDC Program Evaluation Framework Action Guide: Step 6, Act on Findings.