Reporting and KPIs
How to trust the numbers in a management report
How to trust the numbers in a management report: the checks that catch a bad figure before it reaches leadership, so reconciled reporting ends the doubt.
Austin Levine
·
Aktualisiert
How to trust the numbers in a management report comes down to three checks: does every headline figure trace back to one owned source, do the totals reconcile across all the slides, and was the data refreshed and signed off before the pack went out. When all three check out, you can present the number without hedging.
You're on slide 3 of the management report and the cash figure reads 2.1 million. Six slides later, the balance sheet shows a different cash position for the same date. Someone on leadership notices. Now the meeting is about which figure is right, and the decision you came to get approved waits while everyone squints at two numbers that were both pulled from the same company.
This is the moment most finance leads dread, and learning how to trust the numbers in a management report is how you keep it from happening. You didn't build a sixty-page pack to lose the room over one line that won't tie out, but that's all it takes. One figure leadership can't square puts a question mark over every other page. Only 40% of executives say they always trust the data they work with (Talend), and a leadership review where two slides disagree is what that statistic looks like when it walks into the room.
The good part is that trust in a number is checkable. You don't have to rebuild the report to know whether a figure will survive scrutiny. You have to know what to look at before the pack leaves your desk, and you have to know what a reporting setup looks like when it stops producing figures that argue with each other.
Key takeaways
Trusting a management report comes down to three checks you can run in about twenty minutes.
Check that every headline figure traces to one owned source and that totals reconcile across slides.
Confirm the data was refreshed and signed off before the pack went out.
The numbers argue because the pack was built slide by slide instead of on one shared foundation.
Why don't the numbers in your management report match?
Here's the plain reason two reports show different numbers for the same month. Each slide in the pack was built by a different person, from a different export, at a different moment. Finance pulled revenue from the ledger on Tuesday. Sales pulled it from the CRM on Thursday. One counts a deal when the invoice goes out, the other counts it when the contract is signed. Both people did careful work. The figures still don't agree, because they were never built to.
Spreadsheets make it worse. When a number passes through three or four exports and a copy-paste on the way to a slide, small mistakes ride along with it. Research from Panko and colleagues found that 94% of business spreadsheets used in decision-making contain errors (2024). So the figure on slide 3 didn't get there through carelessness. It got there because the number had no shared definition and no single place it came from.
That's why "just double-check it" never solves this. You can check a figure ten times and it will still disagree with the other slide, because the two slides are measuring slightly different things and neither one is marked as the version everyone should use.
What does a management report you can trust look like?
A pack you can defend has four things behind it, and none of them are visible on the slide itself.
Every headline number traces to one owned source. Revenue, cash, and margin each come from one place that someone's name is attached to, not from whichever export was handy that week. There's a single source of truth for each figure, so two people asking the same question get the same answer.
Each number has one written definition. "Revenue" means one thing across the whole pack, and that meaning is written down once instead of living in five people's heads. When leadership asks how churn is calculated, there's one answer, not a debate.
The data was refreshed on a known schedule. You can say when the figures were last updated, so nobody presents last Tuesday's cash as today's.
And the totals reconcile across slides before the pack ships. Cash on the summary matches cash on the balance sheet. The numbers were tied out on the way in, not in front of leadership. 67% of organizations say they don't fully trust the data they use to make decisions (Drexel LeBow and Precisely, 2025), and a pack built this way is how you land in the other third.
How to trust the numbers in a management report: what do you check first?
You can pressure-test a pack in about twenty minutes if you check the right four things, in order. These are decision-level checks, not technical ones. You're asking questions, not writing formulas.
First, pick your three biggest numbers and trace each one to its source. Ask the person who built the slide where the figure came from and when. If the answer is "an export I keep on my drive," that number is fragile.
Second, reconcile the totals across the pack. Does cash on the summary match cash on the balance sheet? Does the revenue on the sales slide match the revenue finance is reporting? A management report that reconciles has already been checked this way before anyone sees it.
Third, confirm the refresh and the owner. Every trusted figure should have a last-updated date and one person responsible for it. A number nobody owns is a number nobody fixes when it drifts.
Fourth, ask who signed it off. One named person should be able to say "these figures are correct as of this date." If no one will put their name to it, leadership shouldn't be asked to bet a decision on it.
Check | A number you can present | A number you should question |
|---|---|---|
Where it comes from | One owned source, named | A private export, unclear origin |
Definition | Written down once, shared | Different on each slide |
Refresh | Known date, on a schedule | "Sometime last week" |
Reconciliation | Ties out across every slide | Only checked on the slide it's on |
Sign-off | One person stands behind it | No one owns the figure |
Run those four checks and you'll catch most of the figures that would have embarrassed you in the room. What you can't do with a checklist is stop the problem from coming back next quarter. For that, something upstream has to change.
What does your team do to fix this?
This is the one place worth going a level deeper. The reason the numbers argue is that the reporting was built slide by slide instead of on one shared foundation, so the fix is to put that foundation in.
In practice that means one governed data model that every report reads from, so revenue is calculated once and everyone pulls the same figure. It means each number's definition is written into that model instead of rebuilt by hand each month. It means the refresh runs on a schedule you can see, and access is set by role so the right people see the right rows. Do that once and the management report stops being a monthly reconciliation exercise, because the slides are reading from the same numbers to begin with.
This is the part that takes real work to get right, and it's the work we do for finance teams. Once one trusted model sits behind the pack, the twenty-minute check above mostly comes back clean, because the figures were never allowed to diverge in the first place.
When should you bring in outside help?
A few signs tell you the reporting has outgrown spreadsheets and manual fixes. Your month-end close drags past eight days, which is the point most finance teams call too long (the strongest performers close in under three, per SolveXia). Two reports show different numbers often enough that people have stopped trusting either. Your team spends more time assembling the pack than reading it. Or a leader has already caught a figure that didn't tie out, and you'd rather that never happen again.
Poor data quality costs organizations an average of 12.9 million dollars a year (Gartner), so the cost of leaving this alone isn't small, even if it hides inside slow closes and second-guessed decisions.
If any of that sounds like your last leadership cycle, start with our free Power BI Governance and Cost Scorecard. It's a five-minute check that scores where your reporting stands and flags your top three trust risks. If the numbers in your pack still won't reconcile after that, book a scoping call and we'll review the model behind your leadership team reporting with you.
Frequently asked questions
Why do two reports show different revenue for the same month? Because each report was built from its own export with its own definition of revenue, pulled at its own time. Both can be careful work and still disagree. The fix is one owned source and one written definition that every report reads from.
How long should it take to check a management report? About twenty minutes if you trace your three biggest figures to their source, reconcile the totals across slides, confirm each number's refresh date and owner, and check that one person has signed it off. It takes far longer if you're rebuilding numbers by hand in front of leadership.
What's a single source of truth in plain terms? It's one agreed place each number comes from, with one definition, that everyone uses. When revenue lives in one governed model instead of five spreadsheets, two people asking the same question get the same answer, and the management report stops contradicting itself.
Do we need Power BI Premium to get trusted numbers? Usually no. One owned data model, written definitions, scheduled refreshes, and access by role all work on standard Power BI. Premium or Fabric matters for very large data or heavier controls, not for getting a management report that reconciles.
Whose job is it to make the numbers trustworthy? Someone with authority over the figures, usually the CFO or a finance lead, has to name one owner and one definition per number and make it stick. Tools don't create trust on their own. A person deciding "this is the source, this is the definition" is what makes the number trustworthy.
Read next: Data governance framework: why most of them get ignored and Power BI row level security, explained.
