You ran a culture survey. What changes on Monday?
The report is not the result. The result is a clearer decision, a named owner, and a check on whether the experience of work actually changed.
The short answer
Begin by stating what the aggregate results can and cannot support. Select a small number of patterns, test them against real work, assign each next move an owner and observable evidence, and set a review date. Share the rationale without revealing individual responses. A survey becomes useful when its interpretation survives contact with Monday's decisions.
Pull the next thread
All articles →First, decide what the report actually says
A chart can show that one group reports more friction around decisions than another. It cannot tell you who caused the friction, whether the same thing happens in every team, or which fix will work. Avoid writing a confident story before checking those limits.
Look at who was included, which organizational levels met the response minimum, and when each wave was collected. If one level did not qualify, do not invent a comparison. If waves happened at different times, describe a historical sequence—not a same-day gap.
Choose the few patterns that could change a decision
A report with seven dimensions can tempt a leadership team to launch seven initiatives. That usually turns attention into a queue no one owns. Instead, choose the patterns with a clear connection to consequential work and a plausible next decision.
For each pattern, ask what would look different if the operating condition improved. If the answer is only 'people will feel better,' the action is not defined yet. If the answer names a decision made sooner, a handoff accepted with less rework, or capacity made visible before a commitment, you have something to test.
- Pattern: What do the qualifying aggregate results show?
- Hypothesis: What in the operating system might explain it?
- Check: Which real workflow could confirm or challenge that hypothesis?
- Decision: What will we change first if the check supports it?
Put a person and a measurement on each next move
A suggestion such as 'improve communication' is too broad to own. 'Publish the decision boundary for customer exceptions, test it in two live cases, and review time-to-decision in 30 days' is a move someone can accept or reject.
Name one owner, one piece of evidence, and one review point. The evidence need not be an elaborate metric. A short escalation log, a rework count, or a weekly list of stopped work may be enough to reveal whether the change reached the people doing the work.
Tell participants what will happen without exposing them
People who answered deserve to know what was heard, what leaders will investigate, what will change, and what will not change yet. Communicate at the level the report can safely support. Do not quote or infer an individual's response to make the story more vivid.
CNNCTD's assessment is designed for aggregate release only after each included level meets its minimum. Small answer groups are suppressed. Raw answers are not a manager's follow-up list, and the work should never become a hunt for who said what.
Review the work, then decide whether to widen the picture
At the review date, ask whether the new decision rule or handoff was used, what it changed, and what still stuck. If the evidence contradicts the original interpretation, update the interpretation. That is learning, not failure.
An organization may begin with one perspective and add another later. The new level joins the same client program, while prior released reports remain intact. This can sharpen the picture—but it should not erase the timing difference or manufacture a single timeless score.
Should we publish every survey result to the whole company?
Share the findings and decisions that can be communicated responsibly. Protect individual answers and small groups, explain the limits of the data, and tell participants what action or further investigation follows.
How many actions should we take from one report?
Use a small, reviewable set rather than a long list of aspirations. CNNCTD's review can refine proposed actions, but an action belongs in the plan only when an owner, evidence, and review point are clear.
Can we add another leadership level later?
Yes. A later module can add that level to the same client program. Earlier reports remain intact, and comparisons should identify when each wave was collected.
See the assessment
From findings to reviewed next moves
The fictional sample shows how CNNCTD connects an aggregate pattern to a hypothesis, an owner, evidence, and a review point.
Make the next move visible.
If your current survey ends in a deck instead of a decision, let's discuss what a more useful operating readout would require.