Skip to content

Weekly Stock-Take Variance Report

Example prompt: "Every Sunday evening once the head chef has entered the weekly stock-take, compare the closing stock against last week's closing plus what we bought from suppliers minus what the till says we sold, flag the lines where the numbers are clearly out, and email the head chef a short report so we can talk about it on Monday."

The Problem

In an independent restaurant the weekly stock-take is the only meaningful read on whether the kitchen is portioning honestly, the bar is pouring fairly, and whether anything is walking out the back door — and in most kitchens it gets entered into a spreadsheet on Sunday and nobody runs the variance calculation because the calculation needs three sheets cross-referenced and an evening that nobody has. The result is that wastage and over-portioning are caught at the next quarterly accountant review, which is twelve weeks late, and the line that started over-portioning the steak in week three has done it for ten weeks before anyone notices. The variance report is a half-hour spreadsheet job done well or a five-minute spot-check done badly — and the half-hour version is the one that catches the slow drift on the dish where the chef has been generous with the protein.

How GloriaMundo Solves It

We build a workflow that fires every Sunday evening once the head chef has entered the weekly stock-take into the 'Stock Take' tab of the Google Sheet. An integration step reads the closing stock for each line item against last week's closing stock for the same line. A second integration step reads the supplier invoices logged in the 'Purchases' tab for the week, summing the units bought of each line. A third integration step reads the POS sales export the till supplier emails on Sunday night, pulling the units sold of each menu item, and a code step maps the menu-item units back to ingredient-line units using the recipe-card ratios held in a 'Recipes' tab. The same code step computes the expected closing stock — opening plus purchases minus sold-at-recipe-cost — and flags any line where the actual closing is materially different from the expected closing. An LLM step writes a short report grouped by section — protein, produce, dairy, dry goods, bar — calling out the lines that are out and offering a plain-language read of why each one might be out, without claiming to know. The report goes to the head chef as a Gmail email Sunday night with a Slack DM headline so the Monday morning conversation runs off the same numbers in both heads.

Example Workflow Steps

  1. Trigger (scheduled): Sunday at 21:00, after the weekly stock-take is expected to be entered.
  2. Step 1 (integration): Read this week's closing stock from the 'Stock Take' tab in the bookings Google Sheet, line by line.
  3. Step 2 (integration): Read last week's closing stock from the same tab so the opening stock for the week is known.
  4. Step 3 (integration): Read the 'Purchases' tab for the week and sum the units bought of each line item from the supplier invoices logged there.
  5. Step 4 (integration): Open the POS sales export email received from the till supplier between Sunday 18:00 and Sunday 21:00, parse the CSV attachment, and pull the units sold of each menu item for the week.
  6. Step 5 (code): Map the menu-item units back to ingredient-line units using the recipe-card ratios held in the 'Recipes' tab, compute the expected closing stock per line as opening plus purchases minus sold-at-recipe-cost, and produce the per-line variance.
  7. Step 6 (llm): Write a short report grouped by section — protein, produce, dairy, dry goods, bar — calling out the lines where the variance is materially out and offering a plain-language read of why each one might be out, without claiming to know.
  8. Step 7 (integration): Send the report as a Gmail email to the head chef and the owner with the Sunday-evening timestamp and the link to the source tabs.
  9. Step 8 (integration): Post a short Slack DM to the head chef with the headline — number of lines out, total estimated value of the variance, and the section most affected — so the email is opened before the Monday-morning conversation.

Integrations Used

  • Google Sheets — the 'Stock Take' tab for the closing counts, the 'Purchases' tab for the week's supplier invoices, and the 'Recipes' tab for the menu-item-to-ingredient-line ratios
  • Gmail — the POS export email parsed for the units sold per menu item, and the Sunday-evening report to the head chef and owner
  • Slack — the head chef's DM with the headline so the email is opened before Monday morning

Who This Is For

Owner-operators and head chefs at independent restaurants and gastropubs running a weekly stock-take in a Google Sheet, logging supplier invoices in the same workbook, and receiving a Sunday POS sales export from the till supplier — anyone who has had the accountant flag a six-month food-cost drift and only then started asking which line on which section was the cause.

Time & Cost Saved

The variance calculation done by hand is forty-five minutes to an hour of cross-referencing three tabs and a CSV; in most kitchens it is done at the quarterly accountant review and never weekly. The cost of catching a portioning drift at week four instead of week sixteen is meaningful — twelve weeks of over-portioned protein on the busy dish is a real number against the food-cost line. The workflow does not replace the head chef's read of the report — they are the only one who can say whether the variance on the lamb shoulder is the new starter on the section or last Friday's wastage of the slow-braise that did not sell — but it puts the lines in front of them every week with the calculation already done, in a format that reads from one page.