Working with data has an interesting quirk: there is nearly always something worth noticing. Handing a team plenty of measures, reporting periods, categories and ways of cutting the numbers and trends, shifts, oddities and connections will show up all over the place. A few will be real surprises, a few will back up what people already believed and others will be accurate but of little use. The real difficulty, then, isn’t finding something. It’s choosing which of those things deserves our attention.
That choice becomes sharper once data enters a meeting room, because the group’s attention only stretches so far. Nobody can dig into every movement, debate every pattern or respond to every finding, however visible the analysis makes them. If we have no deliberate way of ranking what we’ve found, attention gravitates to whatever looks most dramatic, whatever moved most recently or whatever the most confident person in the room happens to raise. Those are not automatically the things that matter most.
What actually counts as an insight?
Gary Klein (2013), a research psychologist, has spent years studying insight and how people make decisions. He separates two things: recognising patterns we’ve already come across and uncovering something truly new. As he puts it, when describing how experts work “Intuition is the use of patterns they’ve already learned, whereas insight is the discovery of new patterns.” What defines an insight is that it shifts our understanding. We end up looking at the situation in a way we hadn’t before.
This difference is worth stressing, because many organisations throw the word “insight” around loosely. Almost any observation gets the label, even when it only repeats something we’ve watched happen month after month… and that isn’t an insight. Picture customer satisfaction slipping across six reporting periods, then slipping again this month. What we’ve spotted is a trend that’s carrying on. The newest figure may be significant, and it may make us more certain the pattern is genuine, but our understanding hasn’t necessarily moved.
A genuine insight might be learning that the fall is concentrated in a customer group we had assumed was very happy, or that it started right after an unrelated-looking change to one of our internal processes. In that case the data has done more than report that something is going on. It has rewritten the explanation we were relying on. Put simply, an insight is frequently the “Aha!” moment, the flash of a lightbulb.
New doesn’t always mean important
I do want to be clear… patterns, trends and confirmations are far from worthless. Learning that a known pattern is still holding can be hugely valuable, especially if we’re monitoring a serious risk, judging whether an intervention is working, or testing an assumption that a big decision rests on. On the flip side, being surprising doesn’t make a finding useful. A new pattern can be intriguing yet unreliable, unconnected to our goals, or too minor to matter. In the end, discovery and priority are two different things.
Sorting findings with the eisenhower matrix
At this point I introduce teams to the Eisenhower Matrix, which many people know from Stephen Covey’s The 7 Habits of Highly Effective People. It’s typically used to organise tasks by urgency and importance, yet it works just as well for organising what we spot in the data. It shifts the question from “What can we see?” to “What deserves our attention right now?”
Used on trends and insights, the matrix has us view every finding through two lenses: importance and urgency. These prompts can help you think each one through.
How important is this? Would it change how we see a strategic priority, put one of our assumptions to the test, uncover a genuine risk or opportunity, or touch the people and results we care about? If we ignored it entirely, what could follow?
How urgent is this? Is a decision around the corner, an opportunity close to disappearing, or a risk that grows the longer we leave it? Does it demand attention immediately, or would a more thoughtful investigation produce a better response?
Together, those two questions sort findings into four groups, and each group calls for its own kind of action.
Four kinds of findings
Urgent and important. These need concentrated attention straight away. They might point to an immediate risk, a serious slide in performance, or a fresh pattern that affects a decision we’re about to make. Findings like these deserve to interrupt the agenda, trigger more questions and end with a named owner for the next step. Important but not urgent. These are the easiest to let slide, and in my experience they’re sometimes home to the most valuable insights. A faint signal may not call for a quick reaction, but it could challenge a long-standing assumption or hint at a shift that grows in significance over time. They need protected time to explore. If we wait until they turn urgent, we may lose the chance to respond with care at a stage when the response could have been smaller and less rushed.
Urgent but not important. These tend to generate noise. A reporting deadline, a one-off spike or a worried request from a stakeholder can make something feel pressing without it being strategically meaningful. They may still need checking, answering or handing off, but they shouldn’t be allowed to reshuffle the organisation’s priorities just because they showed up loudly. We do have to weigh up whether any action is needed, and that call will depend on our professional understanding of the context and the data.
Neither urgent nor important. These can usually be noted, watched or put to one side. That can feel awkward for teams who put effort into the analysis, since finding something makes us want to share it. But not every observation deserves a slide, a discussion or an action. Good data practice includes knowing what not to raise and what to let go.
Deciding where a finding belongs takes conversation
Naturally, findings don’t come labelled “urgent” or “important”. Those are judgements made in context, and at times they’re fairly subjective. Sensible people can place the same finding in different quadrants because they read its implications differently. That discussion is a large part of what makes the matrix worthwhile, so when working with others, agree on the quadrant together. Doing so brings out what people think matters, what they feel can wait, and where their assumptions about risk, opportunity and organisational priorities pull apart.
Guard against urgency pushing out importance
The matrix also helps us push back against one of organisational life’s most stubborn habits: urgency squeezing out importance. When the pressure is on, our attention is pulled towards whatever needs an instant reply. Meanwhile the slower patterns, the emerging risks and the genuinely new insights sit quietly in the background, since they aren’t asking anything of us yet. Sadly, by the time they do become urgent, the story may have been unfolding for months.
The mark of a mature data culture
A mature data culture isn’t measured by how many trends and insights people can produce. It’s measured by whether people can tell confirmation from discovery, noise from consequence, and what needs action today from what deserves deeper thought before it becomes a crisis.
So next time your team gathers around a dashboard, report or set of findings, try holding back the instinctive question, “What do we notice?” Ask this instead: which of these findings changes what we understand, which matters most, and which needs something from us now? When we manage that, our observations start turning into judgement and that is the point where data begins to shape better decisions. For me, that’s what a data-informed culture is really about: not more findings, but better judgement.
FAQ
1. How do you prioritise data insights?
Judge each finding on importance and urgency. Importance asks whether it changes our understanding, challenges an assumption or affects outcomes we care about. Urgency asks whether a decision, opportunity or growing risk needs attention now.
2. How do you decide what to act on in data analysis?
Don’t act on whatever looks most dramatic, changed most recently or is raised most confidently in the room. Sort findings by importance and urgency, and discuss with colleagues which quadrant each one belongs in.
3. What are examples of insights in data?
If customer satisfaction has declined for six periods, another drop is a continuing trend. An insight would be finding the decline is concentrated in a group we thought was highly satisfied or that it began after an unrelated change to an internal process.
4. What is the difference between insight and intuition?
Gary Klein, a research psychologist, says intuition uses patterns we’ve already learned, whereas insight is the discovery of new patterns. An insight creates an unexpected shift in how we understand a situation.
5. Why do teams struggle to act on data?
Teams can find trends and anomalies almost anywhere, and attention in meetings is limited. Without an intentional way to prioritise, urgency tends to crowd out importance.