How to Close a Year Without a Scramble

Why this matters

The year-end scramble is not caused by year end. It is caused by twelve months of small deferrals arriving at once: the vendor whose information nobody collected, the card charges nobody could identify, the loan payment expensed in full, the jobs finished in the last week and invoiced whenever. None of those are hard on the day they happen. All of them are hard to reconstruct months later, by which point the person who knows the answer has forgotten, and the answer costs owner hours instead of office minutes.

The second cost is worse and less visible. A shop deep in a scramble sends its preparer a package it has not reviewed, gets a return built on it, and never has the one conversation where somebody asks whether last year actually went the way the owner thinks it did. The year closes and teaches nothing.

Nothing here is tax advice, and none of these steps involve you taking a position. The whole design is to get the shop to the accountant early enough, with a package complete enough, that the questions genuinely requiring their judgment land on their desk while there is still room to think about them.

Step 1. Set the delivery date, then build the calendar backwards

Ask your preparer when they want your package, then work backwards from it in weeks. A useful default for a small service shop, counting from the first week after year end:

• Week 1: the counts that can only be done at year end.
• Weeks 2 and 3: the ordinary monthly close for the final month, on its ordinary calendar.
• Week 4: the reconciliations that only happen annually.
• Week 5: the owner's cleanup pass, and treatment questions submitted.
• Week 7: the pre-review conversation with the accountant.
• Week 8: package delivered.

Skip the backwards build and the year-end work has no deadline of its own, so it competes with running the shop and loses every week until an external deadline supplies the urgency. That is the scramble, in one sentence.

Step 2. Do the counts in week 1, because they cannot be redone

Three things exist only at the year-end date and cannot be reconstructed afterward: parts and materials on hand, work completed but not yet invoiced, and work in progress on jobs that straddle the year.

Count parts on the last day or the first, not later. Pull the list of jobs completed in the final two weeks and confirm which are invoiced. Pull open jobs and note where each stands.

Skip this and you will be estimating in week 6 what was on the shelf in week 0, and the estimate will be defended rather than measured. How completed-not-invoiced work lands in the year depends on your accounting method, and that is your accountant's call, not yours and not your bookkeeper's - your job is to hand them an accurate count of it.

Step 3. Run the final month's close normally, on the normal calendar

Do not merge the final month into a special year-end procedure. It is a month, it gets the same document packet, the same reconciliations and the same review as any other. Merging it is how a normal month's errors get buried under year-end work and never separately found.

Skip this and you lose the ability to tell a December problem from a year problem, which is the distinction that tells you whether something is broken or something was seasonal.

Step 4. Reconcile what only gets reconciled annually

Some accounts do not earn a monthly reconciliation and do earn an annual one. Loan and finance balances against the lender's year-end statement, so you can confirm the interest and principal split was applied all year. Equity accounts, so owner contributions and draws are where they belong. Prepaid items, so anything paid for last year and consumed this year is sitting correctly. Accrued liabilities, so obligations that exist but have not been billed are visible.

Skip this and the errors here are the ones that compound quietly, because nothing else in the year touches these accounts. A loan payment booked entirely as expense understates profit by the principal portion every single month, and the principal share of an amortizing payment grows over the term, so the distortion gets larger the longer it runs.

Step 5. Clear the owner contamination, and make it stop

Two passes. The correction pass: identify every personal transaction on a business account and every business transaction on a personal one, and get them reclassified. The prevention pass: figure out which card, which habit, or which absent process produced them, and change it.

Skip the second pass and you will do the first one identically next year. The correction is clerical; the prevention is the only part with lasting value.

Step 6. Collect vendor information at first payment, not at year end

Your preparer will ask for information about vendors you paid during the year, and which ones are reportable is their determination, not yours. Your job is that the information exists.

Collect it before the first payment goes out, as a condition of setting the vendor up. A vendor who wants to be paid supplies their details immediately; the same vendor eleven months later, possibly having moved or closed, may not respond at all, and you have no leverage because you have already paid them.

Skip this and year-end vendor chasing is the single most reliable source of scramble, because it depends on other people responding on your schedule.

Step 7. Route treatment questions in week 5, and hold the pre-review

Everything the year raised that is set by rule rather than preference goes to the accountant in week 5, in writing, as one batch: large purchases, a new revenue model, a change in how someone is engaged, a write-off, a new jurisdiction. Then hold a scheduled conversation in week 7, before the return is prepared.

Skip the pre-review and the accountant meets your year for the first time while preparing a filing, which is the most expensive moment to discover anything, and the moment with the least room to do anything about it.

Worked example: 19 weeks down to 9

A two-crew shop that had delivered its package in week 19 the previous year ran the calendar above the following year, and missed it.

Vendor information. Prior year: 27 vendors needed details at year end, 19 were already on file, and 8 had to be chased. Of those 8, three never responded and two of the three had changed address. Under the new practice of collecting at vendor setup, the following year had 31 vendors, 31 on file, none chased. This single change removed the only part of the process that depended on other people answering their phone.

Owner contamination. Prior year: 14 transactions had to be reclassified as personal at year end, each requiring the owner to remember a charge from up to eleven months earlier. Following year: 2, both surfaced in the monthly open-questions list within the month they occurred and answered while the owner still remembered. The prevention pass had found the mechanism - one fuel card carried by a crew member with no restriction on use - and fixed it.

Counts. Week 1 identified 6 jobs completed in the final week and not yet invoiced. All six were counted and handed to the accountant as a number, not as a question the shop answered itself.

Where it slipped. Week 4's annual reconciliations found that the equipment finance payment had been recorded entirely as expense for 4 months of the year, with no split between interest and principal. Correcting it meant rebuilding four months of that account against the lender statement and restating those months, which pushed the pre-review conversation from week 7 to week 8 and the package from week 8 to week 9.

So the shop did not hit its 8-week target. It delivered in week 9, against 19 the year before: 10 weeks earlier, about 53% less elapsed time. Owner time on year-end work went from about 31 hours to about 9, roughly 71% less, and almost all of the remaining 9 hours was judgment rather than reconstruction.

Why the slip was the most valuable week in the process. That loan error had been running for four months and would have run for another year. It reached the surface only because week 4 exists as a scheduled reconciliation of accounts nobody looks at monthly. A shop still in scramble mode finds that error, if at all, when the preparer questions the loan balance, at which point restating four months is happening against a filing deadline instead of against a self-imposed one.

What would change this calendar. A shop with a genuine inventory position, seasonal work in progress that straddles the year, or work in more than one jurisdiction should add two to three weeks to the front, not the back. The counts get bigger and the treatment questions get harder, and both need to reach the accountant earlier rather than the delivery date getting later. A single-jurisdiction shop with no parts inventory and no straddling jobs can reasonably run this in six weeks.

The failure mode to watch for. The tempting shortcut is to deliver on time by sending a package with known gaps and a note saying the rest will follow. It preserves the date and destroys the point. The preparer starts, stops, and re-enters the work later with the context gone, and the gaps are always the items that needed the most thought. Deliver in week 9 complete rather than week 8 partial.

How to verify you got this right

• The reconstruction test. Count how many year-end items required somebody to remember something rather than look something up. Anything above a couple means the monthly process is deferring work into year end, and the fix is upstream in the monthly close.
• The vendor test. Before year end, pull the vendor list and check what share have their details already on file. Anything under complete means the collect-at-setup rule is not being enforced at the point of setup.
• The tie test. Confirm the year-end balances handed to the preparer match your own final statements exactly, per account. If the preparer's opening figures for the next year do not match your closing figures, you have a seam that will make every year-over-year comparison unreliable.
• The question test. Count how many treatment questions the preparer raised that you had not already sent them in week 5. Each one is an item your process missed, and it is the cleanest measure of whether the routing step is working.
• The retention test. The records supporting the return get archived together with it. The general instruction in IRS Publication 583 is to keep records supporting an item of income, deduction or credit until the period of limitations for that return has run, and your accountant will tell you where your own periods land.

References

• IRS Publication 583, Starting a Business and Keeping Records, on retaining records that support a filed return
• See related: The Month-End Document Handoff SOP, The Reconciliations That Actually Catch Things, How to Ask Your Accountant a Question Worth Their Time, The Owner's Tax Surprise: Plan Ahead