Somewhere in your business there is a dashboard that was going to change everything. It got built with enthusiasm, shown in a meeting, praised — and then quietly stopped being opened around week three. Nobody decided to abandon it. It just lost every scheduling contest it entered.
This happens so reliably, across so many businesses and so many tools, that it's worth treating as a design fault rather than a discipline problem. If the same thing fails the same way everywhere, the fault is in the format.
Here are the four reasons, in the order they do damage.
1. Pull versus push
A dashboard requires someone to remember to go and look, decide what's changed since last time, and work out whether it matters. That's three separate acts of initiative before any value appears — and each competes against a diary full of things that are actively demanding attention.
A report requires nothing. It arrives, it says what changed, and the reader does no work at all until there's a reason to. The difference in sustained usage between the two isn't marginal; it's the difference between something that becomes a habit and something that becomes a bookmark.
This is not about willpower. It's about where the information sits relative to the person's existing routine. Anything that requires a new routine to be established will lose to anything that inserts itself into an existing one — and email, whatever its faults, is an existing routine in every business on earth.
If information needs to reach someone whose job isn't looking at data, it has to arrive. Dashboards are for people who are already interrogating numbers as part of their role. Reports are for everyone else — which, in a small business, is nearly everyone including the owner.
2. No interpretation
A dashboard shows that revenue is down 4%. It doesn't say why, and the "why" is the only part that leads to action.
So the reader has a choice: start digging, or note it and move on. Digging means filtering, cross-referencing, possibly exporting to a spreadsheet, and probably discovering that the answer requires data from a system the dashboard doesn't cover. Under time pressure, everybody notes it and moves on. Do that three weeks in a row and the dashboard has taught its reader that opening it produces questions rather than answers — at which point they stop.
This is the part that has genuinely changed in the last couple of years, and it's worth being precise about how. Interpretation used to require a person with system access and a free afternoon, which is exactly why it rarely happened. Given joined-up data, the interpretation can now be written automatically: that the fall is concentrated in one product, that three accounts stopped ordering, that the rest of the book is flat.
Note what that requires. It isn't a cleverer chart. It's the underlying data being joined up enough that the "why" is answerable at all — which is a plumbing problem, not a visualisation one.
3. Too much of it
Dashboards accumulate. Every stakeholder asks for their tile, nobody's tile ever gets removed, and after a year you have a wall of forty metrics with no hierarchy.
A wall of forty tells you nothing, because everything is equally prominent and therefore nothing is prominent. The eye has no route through it. Two people looking at the same dashboard will reasonably reach different conclusions about what the important thing is this week, which means the dashboard has stopped being a shared reference and become a Rorschach test.
Worse, it destroys accountability. When five things are shown, each has an owner and each owner knows it. When forty are shown, none do — responsibility is diffused across a grid.
The instinct when a dashboard isn't landing is to add to it. That instinct is precisely backwards, and it's how a moderately unhelpful dashboard becomes an actively counterproductive one.
4. Nobody agreed what the numbers mean
The quietest killer. A dashboard shows "sales" — but is that including VAT? Gross of discount? At order date or invoice date? Does it include the deposit taken on work that hasn't been delivered?
All those definitions are defensible and they produce different numbers. If two people in the business hold different definitions, the dashboard becomes something to argue with rather than act on. And arguing with a number is exhausting, so the argument gets avoided by simply not looking.
This is why definition work — one written sentence per metric — matters more than any tooling decision. It's unglamorous, it takes an afternoon, and skipping it undermines everything built afterwards.
What to build instead
Not "nothing" — dashboards are genuinely useful for the second step. Once a report has raised a question, somewhere to dig is exactly right. The mistake is making the dashboard the primary surface.
The shape that works:
- A short report that arrives on a schedule. A handful of numbers, each with a comparison rather than a bare total, and a sentence of interpretation each.
- One exception worth chasing. Not a metric — a specific thing that has slipped, with enough context to act on it.
- Somewhere to dig, second. Explicitly secondary, reached when the report raises something, and not expected to be visited otherwise.
Everything about that ordering is about not requiring initiative. The report earns attention by arriving. The dashboard earns its keep by being useful when attention is already there.
A weekly report should be readable in about ninety seconds by someone who has a full day planned. That constraint is not a stylistic preference — it's what determines whether it gets read at all after the novelty period. If it takes five minutes, it will be skimmed; if it's skimmed, the exception gets missed; if the exception gets missed, the report stops earning its place.
The counter-arguments worth taking seriously
"Our team does look at the dashboard." Some do, and where a role genuinely involves daily numbers — a stock controller, a credit controller, a sales manager with a target — a dashboard is the right tool and this article doesn't apply. The failure is specific to people for whom looking at data is not the job.
"A report can't show trends properly." True, and it doesn't need to. A report needs to show that something changed and roughly by how much. The moment somebody wants to see the shape of six months, they should be clicking through to exactly the dashboard we've been disparaging.
"Email is dead / nobody reads email." Everybody reads email. What people don't read is long email. That's a length problem, and it applies equally to every other channel.
How to tell if you've got it right
One test, six weeks in: has anyone forwarded it?
A report that gets forwarded to a colleague with a line of comment above it has become part of how the business talks about itself. That is the thing you were actually trying to build. It never happens to a dashboard, because you can't forward a login — and that single asymmetry explains most of the difference in how long each survives.
A second test, less obvious: has anyone complained about a number in it? Complaints are a good sign. They mean somebody read it closely enough to disagree, which means it matters. Silence at week six usually means it's already become wallpaper.