Reports nobody reads: pruning your legacy reporting

Almost every company that has been running for a few years carries a collection of reports produced weekly or monthly for one reason only: they always have been. Someone asked for them once, that person changed roles or left, and the ritual carried on. The decision this article deals with is awkward to raise and liberating to take: which reports to stop producing, how to test safely that they really are surplus, and what conditions the survivors have to meet.

Where a report nobody reads comes from

Zombie reports are not born zombies. Nearly all of them started by answering a real need: a one-off request from the leadership team that hardened into habit, a tracker built during a crisis that has long passed, a "provisional" version someone duplicated under a different name until nobody remembers which one is official. Over time the need disappears and the report does not, because nobody dares to kill it.

The asymmetry is the heart of the problem: retiring a report feels risky — what if someone needs it? — while producing it feels cheap — "it's only twenty minutes". Under that logic the list can only grow. Every new report gets added; none ever gets removed.

Producing it is not free, even when it looks free

The visible cost is the time spent gathering, reconciling and formatting. It is real, but it is the smallest of the three.

The second cost is maintenance: every report drags along its own definitions, its sources and its manual steps. When something changes in a source system, every report that draws on it needs checking — including the ones nobody opens. The more live reports, the more surface there is to maintain.

The third cost is the most expensive and the least visible: noise. When many reports circulate, the one that matters competes for attention with the ones that don't. And when two of them show different figures for what looks like the same thing — because each calculates it its own way — the cost is no longer time: it is credibility. The conversation shifts from "what do we do about this number" to "which of the two numbers is right".

An honest method: stop sending it and listen

Asking people does not work. If you ask "can I stop sending you this?", almost everyone says they would rather keep receiving it: giving up information, even unused information, feels like losing something. The reliable answer is not in what people say but in what they do.

The method is simple: stop sending the report and watch who complains. With conditions, so you do it properly:

• Delete nothing. You pause the distribution, not the ability to produce it. The data stays where it is and the report can come back tomorrow if needed.
• Set a deadline. Decide up front how many silent cycles are enough to retire it for good: several editions that nobody misses are an answer.
• Take the mandatory ones out of the game. Anything tied to a legal, tax or contractual obligation is untouchable, read or unread. This method is for voluntary internal reporting.
• Go report by report. Pausing everything at once turns an experiment into an incident.

And when someone does complain, don't treat it as the method failing: it is the method's best outcome. The complaint tells you who uses the report and, if you ask a little further, what for. It is common to discover that from a document of many pages, the person uses two or three figures. That conversation is worth more than the whole report: you have just learned what they actually need.

What the surviving report has to earn

Pruning clears space for the question that matters: what makes a report worth having? Being read is not enough on its own. A report earns its place when it can answer these:

• Which decision does it feed? Not "it covers sales", but what gets decided differently depending on what it says. If the answer is "it's interesting information", be careful: interesting is not a use.
• Who receives it, by name? A report addressed to "management" or to a list of twenty people has no recipient: it has an audience. Audiences don't decide.
• Does its frequency match the decision's? Producing weekly what feeds a monthly decision means manufacturing three editions nobody will use.
• Can every figure be traced? If, the day someone questions a number, its origin cannot be reconstructed, the report generates questions instead of answering them.

Pruning is a habit, not a project

If you do one big clear-out and change nothing else, in a few years you will be back where you started: reports accumulate by the same mechanism as cupboards. Two small habits prevent the relapse. First: every new report is born with a review date, at which it has to justify itself again to continue. Second: whoever requests a new report states which decision it will feed — not as bureaucracy, but as the same question the survivors already passed.

The goal is not to have few reports for minimalism's sake. It is for every report in circulation to have a real reader and a decision behind it — and for the time freed from producing the dead ones to go into improving the live ones.

Three questions before retiring a report

1. Is there a legal, tax or contractual obligation behind it? If there is, it stays, whoever reads it.
2. Do we know who uses it and for which decision? If nobody can answer, pause the distribution and listen.
3. Is only a small part of it actually used? Keep that part, with a name and a recipient, and retire the rest.

If pulling this thread exposes wider problems — figures that don't reconcile across reports, different definitions for the same metric, sources nobody can trace — the issue is no longer pruning. That ground is exactly what a data & systems assessment covers: before deciding which reports are surplus, knowing whether you can trust the ones that stay.