During the close, accounting owns the numbers. FP&A owns what they mean.

Close week comes around and the question every new FP&A analyst asks, usually too quietly to ask out loud, is: what am I supposed to be doing right now? Accounting is heads-down. You’re refreshing a report that won’t have real numbers in it for another four days. Here’s the honest answer: during the close, accounting owns the numbers and FP&A owns what they mean. The two jobs meet in exactly one place, and it’s the accrual.

Get that division clear and close week stops being a fog. You’ll know what’s yours, what isn’t, and when your clock actually starts.

Two jobs that meet at the accrual

Accounting owns the general ledger. They record transactions, reconcile accounts, post journal entries, run the intercompany and sub-ledger closes, and sign off that the books tie out. When accounting says a number is right, that’s a statement they stand behind through audit.

FP&A owns the story. You take the closed actuals, compare them to budget and forecast, explain the drivers, update the forward view, and turn all of it into something management can act on. Accounting makes sure the marketing line reads $185,000. You make sure someone knows why, and what it means for next quarter.

The one task that sits on the seam is the accrual, and it’s worth understanding because it’s where FP&A does real work before the books close.

What you do before the books close

Pre-close is the back end of the month, roughly the last day or two before period-end. (Close calendars number days as “business day 1,” “business day 3,” and so on; the exact length varies a lot by company, so treat any specific day as illustrative.) This is when FP&A is most useful and most ignored: chase department heads for late invoices and open purchase orders, flag known one-offs, prep your reporting template so it’s ready the second actuals land, and submit accrual estimates.

An accrual records a cost incurred in the period before the invoice shows up, so the period’s P&L is complete. The matching principle in one line: book the cost when you used the thing, not when you pay for it.

Worked example. It’s late March and you support Marketing. A digital agency ran a campaign in March, but their invoice won’t land until mid-April. March’s P&L still needs the cost. The statement of work is $60,000 a month. The campaign only ran the back three weeks (call it 75% of the month), and the Marketing manager confirmed two of four planned ad sets actually went live (50% of scope). Your estimate:

$60,000 × 75% × 50% = $22,500.

You send $22,500 to your accounting counterpart with the backup (the SOW, the manager’s confirmation, the media-pacing report) by the pre-close deadline. Accounting books the entry: debit Marketing Expense $22,500, credit Accrued Liabilities $22,500. In many shops FP&A only prepares the accrual and accounting reviews and posts it; in smaller shops the analyst may post it directly. Either way, document the basis. If accounting can’t see how you got to the number, they can’t stand behind it.

On the first day of April the accrual reverses (most systems do this automatically; some shops reverse manually). When the real invoice arrives at $24,000, it hits April. March carries your $22,500 estimate; April carries the $1,500 difference, the “true-up.” That’s why estimate quality matters. A lazy round-number accrual just pushes noise into next month, and you’re the one who’ll have to explain the swing.

During close, wait — but wait actively

Once the books are open and accounting is closing, the work is theirs. Your job is mostly to wait, and the mistake is to wait passively. Monitor the close tracker. Answer the “is this material?” questions. Provide accrual support when accounting comes back with one. What you do not do is start analyzing.

Analysis on un-final numbers isn’t early. It’s wrong.

Build the variance deck on business day 2, watch accounting post three more journal entries on day 3, and every figure in your deck is now stale. The handoff signal, “actuals are final,” is the single most important line in the whole cycle. That’s when your clock starts.

After close, the real deliverable

Now load actuals and run the comparisons: budget-to-actual and forecast-to-actual.

March Marketing comes in at $185,000 against a $170,000 budget. That’s $15,000 unfavorable, 8.8%. The lazy version of the analysis is “Marketing was $15K over.” That isn’t analysis; it’s reading the number out loud. Decompose it instead: the $22,500 agency campaign wasn’t in the budget, partially offset by a $7,500 event that slipped from March into April. So the real sentence is: Marketing is $15K unfavorable, driven by a $22.5K unbudgeted agency campaign, partially offset by a $7.5K event timing shift into April. Underlying run-rate is on plan. That sentence is the deliverable.

Then it feeds forward. The agency campaign is now a recurring commitment, so add roughly $60K a month to your April–June forecast and flag the run-rate change to the Marketing leader and the CFO. The close just became a forward-looking decision input, which is the entire point of FP&A being in it.

Where it gets messy

Two things separate a useful analyst from a busy one. First, split timing from permanent. A cost that shifted from March to April is a different story than a cost that’s permanently higher, and only the second one should usually move your forecast. The $7,500 event is timing; the agency run-rate is permanent. Treat them differently.

Second, know your materiality threshold. If the company explains variances above, say, $10,000 (your real number will differ, so ask), chasing a $400 line to the penny is wasted close week. Accounting cares whether the pennies tie out. Your job is decision-useful, not exhaustive.

The genuine gray zone is reclasses and corrections: who books them, who decides “good enough,” who owns the final accrual number. There’s no universal rule, and the teams that handle it well treat it as a shared problem, not a border dispute. Practice varies, so confirm how your shop draws the line rather than assuming.

If this is your first close

Build a close tracker with owners and deadlines. Keep a running accrual and journal-entry checklist so nothing gets dropped under time pressure. Pre-build your variance template before actuals land. And on day one, ask your controller four questions: what’s our materiality threshold, when exactly is the “actuals final” signal, which accruals am I expected to estimate, and do I prepare or post them.

The whole month of work compresses into one habit: don’t touch the numbers until they’re done, then have something to say the moment they are.