Somewhere in most growing businesses there is a recurring meeting nobody enjoys. A figure is presented. Someone says that is not the number I have. There is a pause, some scrolling, and then either a decision gets deferred or somebody volunteers to go and check.
The instinct in that moment is to find the accurate number. That hunt can consume a week, and it usually ends without agreement, because the premise is wrong. In our experience the four systems are not producing one right answer and three errors. They are producing four correct answers to four different questions that nobody ever wrote down.
The same word, four meanings
Consider a business selling a service on invoice terms. Ask each system what revenue was last month.
The CRM reports the value of deals marked closed-won during the period. It is measuring commercial commitment. It counts a deal the moment a salesperson updates a stage, which may be days before anything is delivered or invoiced.
The billing system reports the value of invoices raised. It is measuring what has been billed. A deal closed on the 29th and invoiced on the 2nd lands in a different month here than it does in the CRM.
The accounting system reports recognized revenue, net of credits, adjusted for deferrals, aligned to the period the service was actually delivered in. It is measuring what an accountant would defend.
The spreadsheet reports whatever rule its author encoded, possibly years ago, possibly with an exclusion that made sense at the time. Often the author has left the business and the rule left with them.
None of these is wrong. Each is a faithful answer to its own question. The failure is that the business has one word, revenue, pointing at four different questions, and no document anywhere establishing which question is meant in which context.
Why this gets worse rather than better
Ambiguity compounds. Every new tool you adopt inherits it and adds a fifth interpretation. Every dashboard built on top of an undefined metric propagates the confusion faster and more attractively than the spreadsheet did.
Worse, the ambiguity gets absorbed by people rather than fixed by process. A controller learns which figure to use for the board and which for the bank. An operations manager knows the CRM number runs high and mentally discounts it. That knowledge is real and valuable, and it is entirely undocumented. It walks out of the building at the end of any notice period.
By the time a business notices, the cost is not a wrong decision. It is that senior people have stopped trusting reports and started ringing each other instead. That is a slow, expensive way to run a company, which never appears on a budget line.
What actually fixes it
Not a better dashboard. The fix is three deliberate decisions, and they are organisational rather than technical.
Decide what each metric means. One written definition per metric, in a sentence a new employee could apply without asking anyone. Not “revenue” but “revenue recognized in the period per the general ledger, net of credit notes, excluding intercompany.” The hard part is not the writing. It is the argument the writing forces, usually two departments discovering they have meant different things for years. That argument is the deliverable.
Decide where it lives. One source of record per metric. When the definition and the system disagree, the system is wrong and gets fixed. Without a designated source, every disagreement becomes a negotiation.
Decide who owns it. A named person, not a department. The owner approves changes to the definition, and changes get dated. If “active customer” quietly changed meaning in March, that should be discoverable in seconds rather than reconstructed from memory.
Those three decisions, recorded in a single spreadsheet, resolve the majority of recurring reconciliation disputes in a small or mid-market business. Not because the systems start agreeing, they will not and they should not, but now you know what each one is for and what the difference between them ought to be.
The gap is the signal
Here is the reframe that makes this practical. Stop trying to make the numbers match. Start establishing what the difference between them should be.
If the CRM figure normally runs 8-12% above billing because of timing, that is a known, explainable, monitorable relationship. When it runs 30% above, you have a real signal, deals being marked closed prematurely, or an invoicing backlog. You have converted a monthly argument into a monthly control.
Businesses that do this well are not the ones whose systems agree. They are the ones who can explain, without hesitation, exactly why their systems differ and by how much.
Where to start this week
Open a spreadsheet. Five columns: metric, definition, owner, source of record, last changed. Write your ten most-used metrics down the left. Try to complete the definition column without ringing anybody.
Whatever you cannot fill in is not a gap in your documentation. It is a gap in your company’s agreement with itself. This is the most useful thing you will discover this quarter.