Your QuickBooks P&L says one number. The board deck built from the same data says another.

Nobody touched a transaction. Nobody made an error, at least not one anyone would point to.

The gap is structural. Any time you regroup, remap, or reformat numbers outside the system of record, you create a place for the two versions to quietly stop agreeing, and you usually don't find out until someone on the board asks why.

The Two Reports Were Never the Same Document

QuickBooks builds its P&L from its own chart of accounts, in its own order, as of the moment you run it. The board deck is built from an export, at a moment you chose, reshaped into groupings a board actually wants to read.

Those are two different documents pulled from one ledger. They should tie out to the dollar. But "should" is doing a lot of work, because nothing forces it.

Every Remapping Step Is a Place Drift Can Hide

The moment you map "Contract Labor" and "Outside Services" into one line called "Professional Services," you've made a judgment call. Judgment calls are fine. Unverified ones are where dollars go missing.

Same with a combined "Other Income" line, a rolled-up cost of goods section, or any custom grouping built for a board audience instead of a bookkeeper's audience. Each step is useful. Each step is also a seam, and a seam is where a transaction can get counted twice, left out, or dropped into the wrong bucket.

This isn't carelessness. It's the nature of building a presentation layer on top of a ledger. The more useful the view, the more transformation it took to get there, and the more places drift could have entered along the way.

The Account That Got Recategorized in March

Charts of accounts move over the year. An account gets renamed, merged into another, or shifted from operating expense to cost of goods sold for reasons that made sense at the time.

Your saved view, built back in January, doesn't know any of that happened. It's still mapping the old structure. Now your prior-year comparison is quietly comparing two different definitions of the same line, and nothing about the output looks wrong. It just isn't right.

Timing Gaps and Rounding Differences Nobody Chased Down

Export the transaction list on the 3rd, and QuickBooks can show different totals by the 5th if a journal entry posted late or a prior period got adjusted. Neither number is wrong. They're from different moments.

Every accountant has also seen a small gap between two versions of a report and decided it wasn't worth chasing. Sometimes that's a rounding artifact. Sometimes it's a transaction that landed in the wrong account, and the small gap is the visible tip of it.

You can't tell the difference by looking. You can only tell by tracing it.

"I Double-Checked It Visually" Isn't a Control

Scanning a P&L for numbers that look reasonable catches typos. It doesn't catch a real expense sitting in the wrong grouping, because a wrong number in a plausible spot looks exactly like a right number.

Visual review is a habit. It gives you confidence, not proof, and confidence is not what you want to be selling a board.

The Fix Is a Line-by-Line Tie-Out, Before the Pack Ships

The only way to know a report pack matches its source is to reconcile it against that source, line by line, before anyone outside your team sees it. Take the QuickBooks trial balance or P&L export, compare it against the report you built, and look at every line that doesn't land exactly.

Some of what you find will be rounding. Some will be timing. Some will be an account that exists in one version and not the other. All of that is fine, as long as you know which is which before you hit send.

This is what tie-out means in ClosePack. You drop in the trial balance or P&L export from QuickBooks (or wherever the numbers came from), and it reconciles line by line against the report pack you built, showing what ties, what's off by rounding, and what's a real difference worth a second look. QuickBooks stays the system of record. ClosePack just proves the pack still agrees with it.

Try it on sample data at closepack.io/demo. No signup required.