How to read a diagnostic
Read scores alongside coverage, different perspectives and the evidence still needed to explain them.
A diagnostic helps focus an investigation. Its value depends on understanding whose answers it represents, what was asked and what still needs checking. A tidy chart cannot supply evidence that was never collected.
The InnovationFlow Snapshot reflects one respondent's perspective. A full engagement can bring together relevant participants, qualitative discussion, factual checks and organisational evidence. The methodology explains the boundaries of interpretation.
Read the answers before explaining the pattern.
- Perception: positive
- A respondent says priorities are clear. This describes their experience; it does not establish that everyone shares it.
- Behaviour: limited use
- They report that recent continuation decisions rarely referred to those priorities. This suggests examining how decisions were made.
- Capability: unknown
- They do not know who can stop work. This limits the interpretation; it does not demonstrate that nobody has authority.
A useful next question: take one recent continuation decision and establish who made it, which criteria they used and how the decision was communicated.
The pattern could reflect a weak process, poor visibility or different meanings of “priority”. Choosing among those explanations requires more evidence.
Check coverage, perspective and comparability.
Coverage: unknown and not-applicable responses are distinct from a low rating. Both affect what the result covers. In this Diagnostic they are not scored as zero. Read the stated scope and missing information alongside the scores.
Perspective: two respondents can legitimately describe different parts of the same organisation. A difference between roles becomes useful when you understand their responsibilities and access to information. One person's P–B–C contrast cannot establish a leadership-versus-team disagreement.
Comparability: before comparing cohorts or repeated assessments, check the questions, scope, participant mix and time periods. A change in the respondents can change the picture even if the underlying practice has stayed the same.
When available, look at the spread of answers as well as their average. Do not turn a small numerical difference into a precise claim about maturity, improvement or financial value without evidence supporting that interpretation.
Turn one signal into an investigation.
- State what the answers actually show, including missing coverage.
- Write two plausible explanations instead of immediately choosing one.
- Identify the people and records that could distinguish them.
- Agree the next check, its owner and the decision it will inform.
The output is a better-supported question and a route to answering it. Findings can become firmer as the evidence develops.