If your dashboard has more than seven tiles, nobody reads it. We walk through the cull we did with a Series B client last month.
When we first saw the Series B client’s dashboard, it had fourteen tiles. Revenue, ARR, MRR, MoM growth, net new MRR, expansion MRR, churn MRR, CAC, CAC payback, LTV, LTV:CAC, pipeline created, pipeline closed, and a gauge showing headcount against plan. The person who built it was proud of it. Everyone had asked for their number, and everyone’s number was on the screen.
Nobody was looking at it.
The CEO had bookmarked the tab but stopped opening it two months after launch. When asked why, she said: “By the time I find the number I care about, I could have just asked someone.” The CFO had a separate spreadsheet for the three numbers he actually used in board prep. The head of sales had never opened the dashboard once, the pipeline tiles used the CRM’s own definition of “created” rather than the one the sales team actually tracked.
Why dashboards accumulate tiles
The fourteen-tile dashboard was not a design failure. It was a political outcome. Every team lead had been asked what they wanted on the screen. Every team lead named their metric. Nobody said no, because saying no to a VP’s metric is a harder conversation than adding another tile.
This is the standard path. A dashboard starts with three tiles. Someone asks for a fourth. A new team joins and asks for two more. A board presentation goes badly and a new metric gets added to prove the point will not happen again. Within eighteen months, the dashboard reflects every request anyone ever made of it, and it no longer answers any of them clearly.
The problem is not the tiles themselves. It is that adding a tile has no cost. Removing one has a political cost. So the ratchet only turns in one direction.
What “nobody reads it” actually means
When we say nobody reads a fourteen-tile dashboard, we do not mean nobody opens it. We mean nobody arrives at it with a question and leaves with an answer. The dashboard has become a display, not a tool.
A useful dashboard fails the moment you have to scan more than two rows to find the number that prompted you to open it. Human working memory holds about four things at once. A screen with fourteen items is not four items you can act on, it is fourteen items you have to filter, which is the same cognitive load as opening four separate reports. People stop doing that. They call the analyst instead.
The irony is that the dashboard was built specifically to replace the analyst call. When it fails at that job, the analyst gets more calls than before, not fewer.
The cull: what we ask
We do not start the cull by looking at the dashboard. We start by asking three questions of the people who are supposed to use it.
The first question: what number do you check first when something has gone wrong? Not what number you think you should check. What you actually check, in practice, when you are trying to understand a bad week. The answers are almost always the same three or four metrics across the whole company, even when the company thinks it has fourteen important ones.
The second question: which tiles have you never clicked on? We watch people use the dashboard live and ask them to describe what they are looking for as they look. The tiles they skip without noticing are not useful. They may be important metrics, they are just not being used here, on this screen, by this person, in this context.
The third question: if you had to brief someone on how the business is doing in sixty seconds, which numbers would you use? This is the list the dashboard should serve. Everything else is either a separate view for a specialist, a drill-down behind one of those numbers, or noise.
What we cut, and why
For the Series B client, the sixty-second brief always used the same five numbers: net new MRR, churn rate, CAC payback, pipeline coverage, and cash runway. Those five stayed.
MRR and ARR were cut because every person who looked at them immediately scrolled to net new MRR instead. The totals were context, not decisions. We moved them to a drill-down behind the net new MRR tile, still there if someone needs them, not competing for attention with the number that actually changes behavior.
LTV and LTV:CAC were cut because the company was too early-stage for LTV to be reliable, and the CFO knew it. He was not using those tiles to make decisions; he was showing them to investors. They belong in an investor update, not the operating dashboard.
The headcount gauge was cut because it had no decision attached to it. Nobody could name a concrete action they would take based on reading it that they were not already planning. It was a metric someone had asked for because it felt like something a well-run company should track, not because anyone was actually going to use it on this screen.
Pipeline created and pipeline closed were merged into a single pipeline coverage tile, the ratio the sales lead actually used in weekly reviews. Two tiles became one number that told the same story more directly.
The rebuild took two days. The result was a seven-tile dashboard.
What changed
Three weeks after the rebuild, the CEO had opened the dashboard eleven times. In the three months before the rebuild, she had opened it twice.
The CFO stopped maintaining his separate spreadsheet for board prep. The three numbers he had been tracking manually were now on the screen, in the right format, without him having to copy them out of anything. He told us the bigger change was that he trusted the numbers on the dashboard in a way he never had when there were fourteen of them, because with fewer tiles, it was easier to check each one against its source.
The head of sales opened it for the first time once we changed the pipeline tile to use the definition his team actually tracked. That had nothing to do with the number of tiles and everything to do with the metric being right. But it would not have been fixed in the fourteen-tile version, because nobody was close enough to those tiles to notice they were wrong.
The rule we now apply on every build: if you cannot explain, in one sentence, the decision a person makes differently because they saw that tile, the tile does not belong on the main view. It may belong somewhere else. It does not belong here.
If you are building a new dashboard, or wondering why an existing one has stopped being used - book a thirty-minute call. We can usually tell within the first ten minutes whether the problem is too many tiles, the wrong tiles, or the right tiles pointing at the wrong data.