Share what happened, not just what you hoped

Useful learning includes results that are mixed, inconclusive, or different from expectations. This project-work image is illustrative stock photography, not a depiction of Northstar Supply or its employees.

Hands sketching interface ideas on paper during a design workshop

A concise results structure

Give the reader enough detail to judge the observation and its limits.

  1. Question: What did the test try to learn?
  2. What was tried: What change was agreed, and within what scope?
  3. What was observed: Give the measure, unit, count, denominator, source, and dates.
  4. Limitations: What was too small, different, incomplete, or uncertain to conclude?
  5. Recommendation: Should the team continue discussing the test, adjust it, stop, or gather more information?

Separate an observation (“8 of 10 sample handoffs recorded both details”) from an interpretation (“the prompt may have helped in this context”). A small before-and-after count does not establish why a result changed.

Worked example: the shift-handoff prompt

Fictional results, offered as a model for careful reporting.

Observation

Question and change

Would a brief prompt help record the next owner and open item? In the fictional test, a prompt was tried in one agreed handoff context.

  • Starting observation: 6 of 10 sample handoffs recorded both details.
  • Later observation: 8 of 10 sample handoffs recorded both details.
  • Source: fictional manual count; sample dates 7 and 14 May 2025.
Limitations

Interpretation and limit

The later sample contains two more handoffs with both details recorded. The example does not show that the prompt caused the difference or that it would work in another context.

  • The samples are small and do not include a comparison group.
  • Other conditions may have changed between observations.
  • The counts describe this fictional example only.
Recommendation

Next discussion

The proposer shares the record with the relevant work owner. Together they can discuss whether an adjusted test or more information would be useful; the figures do not automatically approve a wider change.

  • Name who should review the result.
  • State what decision or question is next.
  • Keep the recommendation proportionate to the evidence.

Choose a proportionate next discussion

These are discussion options, not automatic approvals or a substitute for the work owner’s decision.

What the evidence looks like A reasonable next discussion
The measure moved in the intended direction, and the scope stayed as agreed. Discuss whether a further bounded test would answer the remaining question.
The result is mixed, or the prompt was difficult to use consistently. Adjust the question or change with the work owner, then decide whether another observation is useful.
The change created an unwanted effect or crossed an agreed boundary. Stop the test and raise the issue with the relevant work owner; follow applicable approved procedures.
The sample is too small, inconsistent, or incomplete to interpret. Gather more appropriate information or report that no clear conclusion can be drawn.