The meeting had been going fine until the sales numbers came up. The regional manager — let's call him Diego, since that's not his real name — presented a total for the quarter. Three slides later, the finance lead presented her own total for the same quarter, for the same region. The numbers were off by about eleven percent.
Nobody in the room could say, with any confidence, which one was right. Diego's number came from the CRM. Finance's number came from a report pulled out of the accounting system. Both were technically the sales total. Neither was wrong, exactly. They just weren't measuring the same thing, and nobody had noticed until two people said two different numbers out loud in the same room.
I've sat through versions of this meeting more times than I can count, across completely different industries, and the pattern is always the same: two honest people, two careful reports, two numbers that don't match — and a room that goes quiet because nobody wants to be the one who's wrong.
Why this happens more often than you'd think
It's tempting to assume someone made a mistake. Usually, nobody did. What actually happened is that two reports were built by hand, independently, at different times, by different people, each making small decisions that felt obvious in the moment and were never written down.
Diego's CRM total counted every deal marked "closed won" as of the day he pulled the report. Finance's number only counted revenue that had actually been invoiced and recognized by quarter-end, under accounting rules Diego has never had reason to learn. A deal Diego closed on the 29th showed up in his number and not in hers. Neither total is false. They're answering two different questions that both happen to be called quarterly sales.
Add in a cutoff time that's slightly different, a currency conversion applied in one spreadsheet and not the other, a discount that one report subtracts and the other doesn't — and you get exactly what happened in that meeting. Nothing dramatic. Just two slightly different definitions of the same word, built independently, never compared until the moment they were both on a screen at once.
The damage goes further than the one number
The immediate problem is obvious: the room can't agree on a number. The bigger problem is what happens afterward, and it's less obvious because it doesn't happen in the meeting — it happens quietly, over the following months.
Once people have seen two "official" totals disagree, they stop fully trusting either one. Not loudly — nobody announces that they no longer believe the reports. It shows up as a second-guessing reflex: someone double-checking a number before using it in a client conversation, a manager asking for "the real number" as if the one in the report might be fake, a new habit of pulling your own version just to be safe. Every one of those is a small tax on time and confidence, and it applies to every number that report produces from then on, not just the one that was wrong that day.
This is the part people underestimate. A single mismatched number doesn't just cost you an awkward meeting. It costs you the easy trust in every future report, because once the room has learned a number can be wrong without anyone catching it, trust but verify quietly becomes verify, then maybe trust.
What "one source of truth" actually means
People use the phrase "single source of truth" like it's a slogan you adopt. It isn't. It's a specific, boring, unglamorous decision: for a given number — sales, active customers, overdue invoices — someone writes down exactly what counts, what doesn't, and when the clock starts and stops. Then that definition runs the same calculation, against the same data, every single time, and everyone who needs that number looks at the same result instead of rebuilding it themselves.
For Diego and finance, that would have meant sitting down once — not during a tense meeting, but beforehand, calmly — and agreeing: does sales mean closed-won in the CRM, or invoiced and recognized? Does it include deals closed after quarter-end but backdated? Is it in local currency or converted, and at what rate, on what day? None of those answers is objectively correct. They're just decisions that need to be made once, written down, and then automated so nobody has to remember them or rebuild them from memory every time the question comes up.
That's the real function of a dashboard built around defined figures: not that it's prettier than a spreadsheet, but that the definition lives in one place, runs the same way every time, and produces the same number for everyone who looks, whether that's Diego, finance, or the CEO three months from now who has never heard this story.
The fix isn't more meetings. It's fewer definitions.
The instinct after a meeting like that is usually to schedule a reconciliation meeting, and then another one next quarter when the numbers drift apart again. That treats the symptom. The actual fix is deciding, once, precisely what each important number means — and building the one calculation that everyone, regardless of department, points to.
It's not a technology problem first. It's a definition problem that technology then makes permanent, so it stops needing to be solved again every quarter by whoever happens to be in the room.
If your team has ever had a version of that meeting — two totals, one quarter, an uncomfortable silence — that's usually the moment to stop reconciling after the fact and start defining before the fact.