Why Most Reports Change Nothing
Somewhere in your business there is a report that gets produced on a schedule, sent to people, and acted on by nobody.
Possibly several. They were useful once, or seemed likely to be, and now they're produced because they're produced.
Each one costs time every month. The more expensive problem is that they make the business feel measured when it isn't.
The test
For any number you produce regularly, ask: what would I do differently if this were twice as high? Half as high?
If both answers are "nothing," you're not measuring. You're recording.
Recording has its place — for compliance, for history, for the thing you'll want to look back on. But it shouldn't be dressed as a management report, and it shouldn't consume a morning.
Why reports end up useless
They measure what's available, not what matters
Some things are easy to count. Website visits, emails sent, posts published, enquiries received.
Easy to count isn't the same as decision-relevant. The important things — why people don't buy, where work gets stuck, which clients cost more than they pay — are harder to get and get skipped.
So the report fills with what the tools happened to offer.
They have no comparison
A number alone means nothing. "47 enquiries" is not information.
Information is "47 enquiries, against 31 last month and 44 the same month last year." Now there's something to think about.
Most reports present current values with no baseline, no target and no trend, which means they can only be read by someone who already knows what good looks like — and that person doesn't need the report.
They aggregate away the answer
Monthly totals hide the thing you'd want to know. An average response time of six hours could be consistent, or it could be most within an hour and a tail of three-day failures.
The average and the distribution suggest completely different actions. Only one of them is in the report.
Nobody owns the response
A report arrives. Everyone reads it. Nothing happens, because nothing was supposed to.
A report without a named person who's expected to act on it is a newsletter about your own business.
They're built on data nobody maintains
The most common. The report is accurate about a system that's half-populated.
If your source field is optional and a third of records are blank, your source report is measuring which third of people filled in a field.
What a useful report looks like
It answers a question somebody asked. Not a dashboard of everything available — the specific thing someone needs to decide.
It has a comparison. Previous period, same period last year, or a target. Preferably two of the three.
It shows distribution where distribution matters. Response times, cycle times, order values — the spread tells you more than the average.
It's short. One page, or one screen. A twelve-page report gets skimmed, which means the important number gets missed.
It names the action. Not just what the numbers are. What should happen as a result.
It has an owner. Someone whose job it is to read it and respond.
The four numbers most small businesses actually need
Not a universal list, but a useful starting point. Most businesses can't produce all four, which is itself the finding.
| Number | Decision it informs |
|---|---|
| Enquiries by source | Where to spend marketing effort |
| Percentage replied to within target | Whether capture is working |
| Conversion rate, and reasons for losses | Whether you're attracting the right work, and what's blocking it |
| Time from enquiry to decision | Whether follow-up is working |
The second half of the third one — reasons for losses — is the most valuable and the most commonly missing. It requires a required lost-reason field, which requires a decision nobody made.
Before building any report
Write the question first. In a sentence. "Which marketing channels produce customers, not just enquiries?"
Check the data exists and is reliable. Before building anything. A report on a field that's two-thirds empty is worse than no report, because it looks authoritative.
Decide who acts on it, and what they can change. If nobody can act on the answer, the question isn't worth asking yet.
Decide what would be surprising. If you can predict the answer within ten percent, you don't need the report — you need a different question.
Retiring the ones you have
A useful exercise, and uncomfortable.
List every report produced regularly. For each:
- Who reads it?
- What decision has it changed in the last six months?
- What would happen if it stopped?
Stop producing the ones where the answers are nobody, none, and nothing.
Most businesses doing this find they can stop about half, which returns real time and makes the remaining ones more likely to be read.
The underlying point
Reporting sits last in the sequence for a reason.
Observe, map, document, simplify, automate, measure.
Measuring a process that nobody has defined produces numbers about variation. Accurate numbers, describing something that isn't a process yet.
That's why businesses that build dashboards first get accurate reports about chaos, and why it's so disorienting — the numbers are right, and they don't help.
If a number wouldn't change a decision, you're recording rather than measuring. Both are fine. Only one is worth a morning a month.
Want to know which of your numbers are worth producing?
A free systems audit covers your reporting and whether it's built on data you can trust.