Most functions now measure themselves thoroughly. Very little of that measurement is designed to produce a decision. The worked example below is from support operations; the pattern is not.
A support operation reviews AHT, CSAT, TTR, transfers and SLAs. An engineering team reviews defect counts, cycle time and change failure rate. A clinical analytics team reviews readmissions, length of stay and adverse events. Different business groups, same challenge.
We have become very good at producing accurate numbers. But an accurate number is not yet an insight, and an insight is not yet a decision. A metric can tell us what changed without telling us where the change sits, why it matters, or what we should do next. That is the distance between an accurate number and a good decision. Bridging it requires more than better reporting. It requires analysis to find the signal, interpretation to establish meaning, and judgement to decide what happens next.
Four stages between data and decision
It is not reporting against storytelling, but a chain in which each stage answers questions the one before it cannot.
| STAGE | QUESTION | PRODUCES |
| Reporting | What happened? How much changed? When did it change? Against which baseline? | Evidence |
| Analysis | Where is the change? What differs? What moves with it? | Signal |
| Storytelling | What does it mean? Why does it matter? What might explain it? What does it change? | Meaning |
| Decision | What should we do? What should we test? How will we know? | Decision |
Reporting is not the problem. Stopping at reporting is. Accurate evidence is the foundation. Analysis finds what deserves attention. Storytelling connects that signal to its business meaning. A decision turns that understanding into action.
The same number, read two ways
Reporting. Transfers increased from 12% to 16%.
Storytelling. Transfers increased to 16%, but 70% of the increase is concentrated in two product queues. Those queues also show higher TTR and more repeat contacts. The pattern suggests the increase is localised rather than operation wide, with skill coverage, routing or case complexity as possible explanations. Before adding capacity or launching broad based coaching, validate the transfer reasons, skill coverage and case mix in those two queues.
The reporting version ends with a number. The storytelling version changes the management question and identifies what to investigate before committing resources.
The same pattern, across domains
The metrics change. The reasoning does not. The same progression applies whether we are examining support performance, software quality or clinical outcomes.
| DOMAIN | REPORTING STOPS AT | QUESTIONS THAT MOVE US TOWARD A DECISION |
| Support operations | Transfers rose from 12% to 16% | Where is the increase concentrated? Is it associated with skill coverage, routing or case complexity? |
| Software engineering | Defect escape rate rose 18% | Which components or releases account for the increase? Are these new defects, regressions, or concentrated in specific change types? |
| Clinical analytics | Readmissions rose two points | Which patient cohorts, conditions or facilities account for the change? Does the evidence point toward case mix, discharge process or care pathway? |
| AN INSIGHT IS NOT A BETTER WRITTEN OBSERVATION Observation. Transfers increased from 12% to 16%. Weak insight. Transfers have increased significantly. This adds emphasis, not understanding. Real insight. 70% of the increase is concentrated in two queues, so the deterioration is localised rather than systemic. Insight narrows where to look. Hypothesis proposes what might explain it. Evidence determines whether that explanation survives. |
Classify by ability to act, not only by category
Categorising problems is useful, but it does not tell us where action can happen. A second classification often matters more: how much influence does the team receiving the analysis actually have over the outcome?
| DEGREE OF INFLUENCE | MEANING | WHAT IT IMPLIES |
| Controllable | The team can change the outcome directly. Routing, response time, guidance, defects it owns | Own it. Act, measure, close the loop |
| Influenceable | Needs another function or partner to act. Upstream quality, a third party, another team’s backlog | Engage with evidence. Track the dependency and the outcome |
| Not controllable | Outside the team’s influence. Policy, market conditions, patient factors | Report it. Keep it out of the improvement target |
A category tells a decision maker what the problems are. Influence tells them where action belongs. If two thirds of an outcome is driven by factors outside a team’s control, asking that team to work harder is unlikely to move the number. The recommendation may instead be to change a process, engage another function, address an upstream dependency, or revisit the target itself.
Finding the driver explains the problem. Finding who can influence the driver makes the analysis actionable.
Reporting and storytelling have different jobs
Both are necessary. Reporting establishes a trusted evidence base. Data storytelling connects that evidence to meaning, implications and action. The problem begins when we expect one to do the job of the other.
| REPORTING | DATA STORYTELLING | |
| Core question | What happened? | What does it mean, and what should we do? |
| Starting point | The metric or performance question | The business question or decision point |
| Role of the data | Establishes the evidence | Evidence supporting an argument or interpretation |
| Time horizon | Backward looking | Connects past evidence to future decisions |
| Typical artefact | Dashboard, scorecard, pack | Briefing, recommendation, decision note |
| Skill it depends on | Accuracy and data engineering | Domain judgement and reasoning |
| Interpretation | Left to the reader | Makes the reasoning explicit |
| Uncertainty | Usually omitted | Stated, with evidence separated from hypothesis |
| Failure mode | The number is wrong | The number is right, the conclusion is not |
| What done means | The number is verified | A decision is informed, or a test is defined |
Reporting asks whether we can trust the number. Storytelling asks what we should conclude from it. Good decision making needs both.
Six rules for Turning Reporting into Decision-Ready Storytelling
- Reporting and storytelling are stages, not alternatives. Reporting establishes the evidence. Storytelling gives that evidence meaning and connects it to what should happen next
- Before telling the story, find where it lives. An average can show that something moved. Its distribution tells you where to investigate. A change spread across the system suggests a different
- problem from one concentrated in two segments.
- An insight must change the next question. Storytelling is not making an observation sound more compelling. A useful insight changes what you investigate, discuss or decide next.
- A hypothesis is not a finding. Good storytelling separates what the evidence establishes from what might explain it. It makes uncertainty explicit rather than turning a plausible explanation into a conclusion.
- Find the lever, not only the driver. Categorising the problem explains what is happening. Classifying it as controllable, influenceable or outside the team’s control tells us where action belongs. Knowing who can influence that driver makes the story actionable.
- Never ask for a chart on its own. Ask for the evidence, the interpretation, the hypothesis and the action it informs. A chart communicates data. Storytelling connects the data to a decision.
| THE DISTANCE, CLOSED We have become very good at producing accurate numbers. The harder skill is knowing what those numbers justify us saying, what they do not, and what we should do next. Reporting creates visibility. Analysis creates understanding. Storytelling creates alignment. Decisions create value. That is the distance between an accurate number and a good decision. |