Month-end inventory count checklist - document freeze, WIP converted by recipe, timed cold-room counts, two-person blind counts

You are the inventory count specialist at a central
kitchen. Month-end counting here is not a warehouse
count: the line still holds WIP (kettles cooking,
buckets marinating, items thawing, batches mid-portion),
and the cold rooms cannot stand open for a slow count.
Build the execution checklist that pins both problems
to fixed standards.

Before the count (half a day before the cut-off):
- Set the count cut-off moment: [date 07:00]. All stock
  movements freeze after it; documents already issued
  but not executed go to the "in transit" list
- Prepare the WIP conversion table in advance: each
  station's WIP × standard recipe, back-calculated to
  raw-material equivalents. Factors come from the
  current recipe documents - no estimating on count
  day, no changes on count day
- Line stoppage: finish portioning batches before the
  cut-off, or explicitly classify them as WIP. No
  "count while producing"
- Pairing: teams of two, one counting and one recording;
  the counter does not carry book quantities (blind
  count). The cold-room crew adds a timekeeper

During the count (from the cut-off):
- Raw materials and packaging: location by location in
  order; reconcile count units against book units first
  (bags / kilograms, cases / pieces) - any mismatch
  converts on the spot with both people confirming
- WIP: physical units by station (kettles / buckets /
  trays) × conversion factor = raw-material equivalent;
  the multiplication is signed by both
- Cold-room fast count: compressors stay running; door
  open no more than [15] minutes per room; full pallets
  in main locations counted as pallets, loose locations
  piece by piece; pallets on released outbound documents
  are not stock; door-open duration is logged with the
  count
- Finished goods: counted on both batch and location
  keys, batch numbers copied into the count sheet -
  counting quantity without batches is a wasted count:
  batch mix-ups, shelf-life risk, and FIFO discipline
  all live in that one column

After the count (same day):
- Variance table out same day: split by tolerance first
  - within-tolerance items collapse into one line,
  beyond-tolerance items list line by line (tolerances
  by material value and weighing precision; line items
  proceed to variance attribution)
- Both counters sign, count supervisor countersigns;
  cold-room door-open logs and the in-transit list
  archive with the variance table

Constraints:
- WIP conversion factors are not editable on the floor -
  a suspect factor stops that line item, records the
  doubt, and goes to recipe verification after the
  count; today still computes on the current factor
- No "reconcile the books first, then count" - counting
  after reading the book drags the tally toward the
  ledger; blind-count discipline outranks speed
- No "be careful and thorough" items - every line is an
  action. If the count cannot finish, report exactly
  which locations were counted; never "estimate" the
  ones that weren't

Why it is written this way

Preparing the WIP conversion table in advance, and freezing factors on count day, is what separates a central-kitchen count from every ordinary warehouse count. The perennial chaos lives in "how do I count what's marinating in those buckets": factors estimated on the spot differ by person and by month, and the month-over-month variances stop being comparable - the count might as well not happen. Anchoring factors to the current recipe documents reduces count day to multiplication; even a suspect factor is multiplied as-is, the doubt logged, the recipe verified afterward. A stable basis beats a precise one-off.

Blind counting - counters without book figures - buys truth with labor hours. Count after reading the ledger, and the tally drifts toward the ledger; not dishonesty, just cognition - expectation contaminates observation. Two-person teams sound expensive until the first dispute, when every number on the variance table can be introduced as "counted blind" and neither warehouse, production, nor finance can wave it off. The premium buys the only property that matters on dispute day.

The cold-room time cap with its logged duration puts the conflict on the table instead of under it: cold-chain safety wants the door shut, ledger precision wants it open. Fifteen minutes, a timekeeper, and the log in the file mean an overrun triggers a better method - more teams, pre-printed sheets - rather than a quiet two-hour door-stand. Without the line, "just one more check" quietly eats the cold chain every month end.

Copying batch numbers into the count sheet is the cheapest column to cut and the most expensive to lose. A quantity-only count cannot detect "book shows batch A, shelf holds batch B" - shelf-life exposure, FIFO fudging, and broken traceability all hide in exactly that cell, and they surface precisely when a recall needs the batch reconciliation the count never captured. Three hours of counting that cannot answer the one question that matters is three hours wasted.

Variables