February 2026 · 7 min read
Your Dashboards Are Wrapping Paper
A dashboard that doesn't change behavior is decoration. The difference between a metric and a decision tool is whether anyone does something different on Monday because of what it showed them on Friday.
We built a dashboard once that got a standing ovation in the demo. Animated charts. Drill-downs by segment, region, and product. Executive summary at the top, detail views for each function. The CMO called it “exactly what we’ve been asking for.” The VP of Sales said it was the best visibility they’d ever had into pipeline health.
Six weeks later, the usage log showed that three people had opened it. One of them was me, checking whether anyone had opened it.
The dashboard was beautiful. It answered every question anyone had asked. And it changed nothing about how the business operated, because nobody’s Monday morning was different because of what it showed them on Friday afternoon.
This is the story of most dashboards in most companies. They are built to answer questions. They should be built to change behavior. The gap between those two goals is the gap between analytics as a service function and analytics as a performance driver.
The ceremony of visibility
There is a ritual in data teams that goes like this. A stakeholder says “I need visibility into X.” The analyst builds a dashboard. The stakeholder reviews it, requests changes, approves the final version, and never looks at it again. The dashboard goes into a folder with forty others. The analyst moves on to the next request.
Everyone involved did their job. The request was fulfilled. The deliverable was shipped. And the business has one more piece of wrapping paper: something that makes the data look presentable without changing what happens underneath.
The problem is not the dashboard. The problem is that “visibility” is the wrong goal. Visibility assumes that if people can see the number, they will act on it. They will not. They are busy. They have twelve other tabs open. The number has to reach them, interrupt their default behavior, and make the alternative action easier than the current one. A dashboard sitting in a BI tool does none of those things.
What changes behavior
I started paying attention to which metrics moved people and which ones just existed. The pattern was consistent across every team I worked with.
Metrics that change behavior share three properties.
They are connected to something the person controls. Revenue is a lagging indicator that a sales rep cannot directly influence. Meetings booked this week is a leading indicator they can. When you show a rep their meeting pace against their historical average, they do something about it. When you show them the company’s quarterly revenue, they glance at it and move on. The metric has to land in someone’s zone of control, not just their zone of interest.
They surface at the moment of decision, not after it. A weekly pipeline report that arrives Friday afternoon is a retrospective. A daily alert that flags deals with no activity in ten days arrives when the rep can still do something. Timing is not a distribution problem. It is a design problem. The question is not “how do we get this report to more people” but “when does this person make the decision this number should inform, and how do we get it there at that moment.”
They create a gap that demands a response. The most effective metric I ever deployed was not a dashboard. It was a coverage ratio by segment that showed up in the Monday pipeline review with red/yellow/green formatting. Green meant your pipeline covered your target at historical conversion rates. Red meant it did not. There was no action button. There was no workflow attached. But sitting in a room with your peers while your segment is red and theirs is green is a behavioral trigger that no amount of dashboard design can replicate.
None of these require sophisticated technology. They require thinking about the metric as a behavioral intervention instead of an information delivery.
The leading indicator shift
Most analytics teams over-invest in lagging indicators because lagging indicators are easy to produce and impossible to argue with. Revenue happened. Bookings closed. Churn occurred. The number is the number.
The problem is that by the time a lagging indicator tells you something, the window for intervention has closed. You are writing the post-mortem, not preventing the incident.
Leading indicators are harder. They require a theory about what predicts the outcome. Meeting volume predicts pipeline generation. Pipeline velocity predicts close rates. Engagement recency predicts renewal likelihood. Each of these is a bet, and some of the bets will be wrong.
But a wrong leading indicator that gets corrected after one quarter is more valuable than a correct lagging indicator that arrives too late to act on. The leading indicator teaches you something about causation. The lagging indicator confirms something about history.
The shift I push for in every analytics organization I work with is a rebalancing of the portfolio. Not abandoning lagging indicators. Supplementing them with a smaller set of leading indicators that are tied to specific behaviors and reviewed at the cadence where intervention is still possible.
This is uncomfortable for most teams because leading indicators can be challenged. Someone can argue that meeting volume does not predict pipeline. With lagging indicators, there is nothing to argue about. The preference for certainty over usefulness is how you end up with 200 dashboards that nobody uses.
Designing for action, not consumption
When I build an analytics product now, the first question is not “what does the stakeholder want to see.” It is “what decision does this metric inform, who makes that decision, and when do they make it.”
If the answer is “it doesn’t inform a specific decision, it’s just good to know,” that is a signal to not build it. Not every request deserves a dashboard. Some requests deserve a conversation about what the stakeholder is trying to change, followed by a metric that supports that change, delivered in the format and cadence that makes the change more likely.
Sometimes that is a dashboard. Sometimes it is an automated alert. Sometimes it is a single number in a Slack message every morning. Sometimes it is a red cell in a spreadsheet that the team already reviews weekly. The format follows the behavior, not the other way around.
The 200-dashboard graveyard I mentioned in the decision systems post did not happen because the data team was bad at building dashboards. They were excellent at building dashboards. It happened because nobody asked whether the dashboard was the right intervention for the behavior they were trying to change. The tool was the default, so the tool got built, and the behavior stayed the same.
The accountability question
There is an organizational reason this pattern persists. Most analytics teams are measured on output: dashboards shipped, reports delivered, requests closed. They are not measured on outcomes: did the metric change the behavior, did the behavior change the result.
This is rational from the analyst’s perspective. Measuring output is clean. Measuring outcomes requires proving causation in a complex system, which is hard and often impossible. But it creates a function that optimizes for delivery speed rather than decision impact, which is how you get a team that is genuinely busy, genuinely responsive, and genuinely not moving the numbers.
The fix is not to measure analysts on business outcomes. That is too indirect and too unfair. The fix is to change the intake process. Instead of “I need a dashboard that shows X,” the request becomes “I want to improve X by doing Y, and I need a metric that tells me whether Y is working, delivered at the moment I decide how much of Y to do.”
That is a harder conversation. It requires the stakeholder to articulate their theory of change, not just their desire for data. It requires the analyst to push back on requests that have no behavioral thesis. And it requires leadership to accept that fewer, better-targeted analytics products will outperform a large catalog of unused ones.
The line
A dashboard that nobody opens is not a failed dashboard. It is a successfully delivered product with no behavioral thesis. The build was clean. The delivery was on time. The impact was zero.
If you want analytics to drive performance instead of just reporting on it, stop asking “what do people need to see” and start asking “what do people need to do differently, and what metric, delivered when and where, will make that more likely.”
The forecasting post said most forecasts are reports pretending to be predictions. This is the companion: most dashboards are wrapping paper pretending to be tools. The difference between decoration and a decision system is whether someone’s Monday changes because of what the data showed them.