The production general ledger: why actuals drift from the budget

The budget is a plan; the ledger is what happened

A production budget is a set of intentions: these people, at these rates, for these days, in these categories. The general ledger is the other side — the bookkeeping record of every dollar the production actually spent, posted to accounts. As long as the two agree, cost reporting is arithmetic. They almost never agree for long, and the difference between them is where shows lose money quietly.

Drift isn't a failure of bookkeeping; it's structural. The budget is built before facts exist, and production generates new facts every day. The discipline isn't preventing drift — it's catching it early enough to do something about it, and knowing exactly which account each difference lives in.

Where actuals actually drift

Most drift comes from a handful of recurring places. Wages drift when the day ran longer than planned: overtime, 6th and 7th day premiums and meal penalties never appear in the bid, and fringes ride on top of them. Fringes themselves drift when the budget applied a blended percentage but payroll applied per-person rates with caps and flat funds.

Purchases drift twice: once at commitment (the purchase order) and once at cost (the vendor invoice), and the two differ on tax, freight, damage waivers and rates. Crew remittances carry kit rentals, box rentals, car allowances and per diems the budget never itemized. Petty cash gets logged days after it's spent, and wrap journals reallocate costs that changed shape after the fact.

• •Overtime and premium pay, plus the fringes riding on them — bid at straight time, paid at loaded time.
• •Blended fringe rates vs. per-person payroll calculations with caps and flat-rate funds.
• •Purchase orders vs. invoices: commitment vs. cost, differing on tax, freight and waivers.
• •Kit rentals, box rentals, car allowances and per diems added on remittances.
• •Late petty cash entries and end-of-show journal entries that re-shape the numbers after the fact.

Reconciliation is a habit, not an event

A wrap-time reconciliation is an archaeology project. The working method is a weekly cost report: compare the budget, the commitments and the actuals line by line, and attach a reason to every difference that matters. A variance with a reason is management information; a variance without one is a leak.

The other half of the habit is classification. A cost that lands in the wrong account hides: it overstates one category, understates another, and defeats every report that runs off the account structure. Getting the account right at entry time — on the purchase order, on the remittance, on the timecard — is cheaper than reclassifying a shoebox of invoices later.

Keep the trail

Every difference between plan and actual should be traceable to a document: the purchase order behind the invoice, the timecard behind the remittance, the fringe calculation behind the fund report. That trail is what a tax incentive audit walks — and what your own variance meetings walk a week later.

Retain the tie-out: a final ledger where each budget line, its commitments and its actuals agree, with the residual differences explained. It's the last thing an auditor asks for and the hardest thing to reconstruct after wrap.

How Clarice handles this

Clarice keeps the budget and the ledger as one connected record instead of two documents that meet at wrap. Budget lines carry a running copy of their numbers — estimate and running total side by side — so the working budget can move with reality without overwriting the bid. Vendor remittances, invoices and purchase orders link back to the budget lines they pay for, and journal entries post to a chart of accounts mapped from the budget's own categories.

Because the links exist, the comparisons assemble themselves: the reconciliation view shows each line's estimated, committed and actual cost together, daily cost reports roll up the spend as it happens, and an audit export packages the trail — budgets, invoices, remittances, timecards, journal entries — without a spreadsheet rebuild.