Why more dashboards don't produce better decisions

Ask a commerce leader how many reports their organisation maintains and you will usually get a guess, not a number. The guess is always low. When teams actually count — the BI dashboards, the channel back-office exports, the weekly decks, the spreadsheets that quietly became systems of record — the estate is typically three to five times larger than anyone assumed.

None of it was built carelessly. Every one of those reports was created by a competent person solving a real problem, usually because the existing reporting did not answer their specific question. That is exactly why the estate grows: each addition is locally rational. The cost only shows up at the level nobody owns — the level where decisions are supposed to get made.

The estate grows because it works — locally

A category manager needs sell-through by size curve. The existing revenue dashboard doesn't break that out, so a new view gets built. A performance marketer needs blended CAC including marketplace ad spend. Finance's margin report treats that spend differently, so a second view gets built. Neither person is wrong. Both are now maintaining a definition of the business that is subtly, invisibly different from the other's.

Multiply that across four or five functions and eighteen months, and the organisation ends up with something that looks like coverage and behaves like fragmentation.

Reporting estate · growth

Four teams, four revenue numbers

Here is the failure that matters, and it is almost never framed as a data failure.

A pricing decision is scheduled. Commercial arrives with €4.18M for the quarter. Finance has €3.71M. Marketing's attributed figure is €4.40M. Operations has a fourth number, drawn from fulfilled orders rather than placed ones. Every figure is defensible. Every figure comes from the same underlying transactions.

The meeting does not produce a pricing decision. It produces a task: reconcile the numbers and reconvene. That reconciliation costs two analysts most of a week, and by the time it lands, the commercial window that prompted the question has moved.

One source · three definitions

The organisation was never short of analysis. It was short of an agreed basis for acting on any of it.

The gap is between the report and the decision

It is worth being precise about where the failure sits, because most remediation programmes aim at the wrong end.

The data pipeline is usually fine. The warehouse is usually fine. Dashboard quality is usually fine. What is missing is the layer that would turn a chart into a decision: a shared definition, a named owner, a threshold that says whether the number is acceptable, and a specific next action with a consequence attached.

Without that layer, a dashboard can only ever tell you what happened. Somebody still has to decide what it means, whether it is bad, whose job it is, and what to do — and in most organisations that translation happens in meetings, informally, inconsistently, and slowly.

What actually closes it

Closing the gap does not require more reporting. It requires three properties that reporting tools do not have, and were never designed to have.

1. One definition, held centrally

Revenue, margin, CAC, sell-through and stock cover each need a single canonical definition that every output inherits. Not a documentation page — a definition the system enforces, so a figure cannot be produced under a competing rule without that being visible. The moment two outputs can disagree, the organisation is back to reconciliation.

2. Ownership expressed as structure, not convention

"Who owns this?" should be answerable by the system, not by asking around. That means roles, responsibilities, and thresholds are modelled data — pricing owns margin floors by category, supply owns cover by SKU class, performance owns CAC by channel. When a signal breaches a threshold, routing is deterministic rather than social.

3. The recommendation is the artifact

The unit of output stops being a chart and becomes a proposed action with its evidence attached: 34 SKUs are trading below the category margin floor; the largest contributor is a promotional overlap in the marketplace channel; proposed action is to reprice this set; owner is pricing; expected recovery is 1.4 points. The chart is still there — underneath, as support. It is no longer the deliverable.

The closed loop

What changes when the gap closes

The visible change is speed, but speed is a symptom. The real change is that the argument moves. Teams stop debating whose number is right and start debating what to do about the number — which is the only debate that has ever created value.

Where to start

Do not begin by rationalising the reporting estate. That is a large, politically expensive project that treats the symptom.

Begin with one recurring decision that matters — weekly pricing, stock reorder, channel spend allocation. Write down its canonical definition, its threshold, and its owner. Then make the system produce that decision, with evidence, on a schedule, to that owner. When that loop is closed and trusted, take the next decision.

The reporting estate will shrink on its own once people stop needing to build their way around the gap.