The Monthly Close, End to End: What Actually Happens Between TB and Board Pack

Most descriptions of the monthly close are either a checklist (“upload TB, reconcile accounts, review variances, sign off”) or a diagram with arrows. Neither tells you what actually happens to a number between the moment it lands in a trial balance and the moment it sits on a slide in front of the board. Here is that path, stage by stage, the way Rexfin builds it.

The upload isn’t the finish line, it’s the start of a quarantine

A trial balance lands as a file. Rexfin fingerprints it (a hash of the workspace and the file’s own bytes) so the same TB can’t quietly get re-ingested as new, and a tampered upload gets rejected outright. Until someone confirms the chart-of-accounts mapping, the upload sits in a pending state, withheld from every read path, including read-only views. That’s not a reaction to anything suspicious in the file: every upload goes through the same holding pattern, clean or not, because a number nobody has mapped yet isn’t a number anyone should be looking at.

Mapping confirms, and a few gates fire whether you notice or not

Confirming the mapping isn’t just clicking “accept.” A handful of checks fire at that moment: does each department’s contribution sum to the reported total, and (once a full year of history exists) does the accumulated year tie to a filed annual figure. Both are fail-closed: if either doesn’t hold, the confirm stops and says why instead of going through quietly.

The board-vs-actual grid, and what it does and doesn’t catch on its own

Once mapped, the numbers render against budget in a variance grid, with every cell honestly labeled by its trust tier (actual, cited, or plan) so nobody mistakes a fallback estimate for a confirmed figure. A library of integrity rules also runs automatically: statistical distribution checks, round-number bias, balance-sheet tie-outs, ratio and margin bounds, trend breaks. It’s a wide net, and it catches a lot of the patterns that usually signal a data problem. It is not a substitute for a controller actually reading the variance column. No rule library catches everything, and automated checks narrow where a human needs to look; they don’t replace the look.

That’s exactly why the variance-commentary step is mandatory, not optional. A material variance can’t move from “open” to “in review” without someone writing down why it happened. That comment is timestamped, attached to the specific cell, and becomes part of the permanent record, which matters later if the explanation turns out to be wrong and needs revisiting during audit prep or a restatement.

Sign-off is two distinct people, not two clicks by one person

Preparer and reviewer are separate roles by default: the person who uploaded and mapped the TB is not the person who signs off on it, unless the workspace is explicitly configured for a single operator. Every transition, exception resolution, and mapping change writes an append-only event. Nothing in that trail gets edited or deleted after the fact, which is the difference between “we have a process” and “we can show you the process.”

Sealing the close: the part built to be checked, not trusted

When a period closes, Rexfin freezes the exact cells that fed every calculation into a signed manifest and hashes it. Every later render of that close (the board pack, a time-travel view, an auditor pulling it back up months later) recomputes that hash and checks it matches. If it doesn’t, nothing renders. Not a warning banner next to a number that looks fine; nothing served at all. That’s a stronger guarantee than “we kept a PDF”: it means a closed period can be mechanically verified as unchanged, not just assumed to be.

Closes can be reopened, but only through an explicit, reasoned action that mints a new sealed version rather than editing the old one quietly. The prior version stays exactly as it was, timestamped and dated, so nobody can rewrite what a board already saw.

Who this is actually for

This is built for controllers and group finance teams running a real close cadence, not a spreadsheet macro dressed up as a process. If your close currently depends on one person remembering which tab has the latest numbers, or a sign-off that means “I skimmed it,” the value here is less about speed and more about being able to answer, months later, exactly what a closed period said and why. For the piece of this that boards actually see, see the board reporting cycle; for what happens when a closed number turns out to be wrong, see restatement response. The full set of use cases this fits into is at the finance team hub.