Prime Cost Is Too High: A 7-Day Triage Plan for Restaurant Operators

You looked at the P&L on Monday and prime cost is 67%. You think it was 61% three months ago. Maybe it was. The real problem is you cannot tell whether this is a one-week blip, a quiet two-month drift, or a seasonal pattern that you have been missing for years. The information is buried in three different systems, the reporting is inconsistent across managers, and the next payroll prints in nine days.

Most operators in this situation do one of three things. They cut a labor shift. They blame the new line cook. They do nothing and hope next week is better. None of these are diagnostic. Below is a 7-day plan that produces real answers and a defensible action list before payroll prints again.

Day 1 — Audit the last 4 weeks of invoices

Pull every invoice from the last four weeks. Sort by vendor, then by category. The goal is not to find one bad invoice — it is to spot the categories that are climbing.

Look for three patterns:

• Same item, rising price. A protein category that is 7% more expensive this month than four weeks ago, with no menu price adjustment, is silently eating margin.
• New vendors you did not approve. When line cooks order on the fly, sub-vendors creep in. Sub-vendors are almost always more expensive per unit than your primary distributor.
• Unit-of-measure changes. A 5-gallon container that became four 1-gallon containers is a price increase disguised as a packaging change.

By end of day one you should have a one-page list of categories where COGS has moved more than 4% in either direction, with the dollar impact at current volume.

Day 2 — Pull labor by daypart

Open your POS and your labor system side by side. Pull the last four weeks of sales by daypart and the last four weeks of labor hours by daypart. Calculate sales per labor hour for each daypart.

Most operators discover one of two things on day two. Either there is a daypart where labor is fine in aggregate but disastrous in one specific shift band — usually the one between the lunch and dinner crews. Or there is a manager who is over-scheduling on Wednesdays in a pattern that goes back six months because that is what the schedule template says, regardless of actual volume.

Document the worst daypart by sales-per-labor-hour. That is your first cut target — but you are not cutting yet. You are diagnosing.

Day 3 — Identify your top 5 COGS offenders

Take the four-week invoice data from day one and rank every category by absolute dollar spend, not by percentage growth. The most expensive categories are where any percentage savings will compound the fastest.

For a typical full-service operator, the top five categories are usually:

1. Center-of-the-plate proteins
2. Produce, fresh
3. Dairy and eggs
4. Beverage alcohol
5. Cleaning and chemicals (yes, really — this drifts more than people think)

For each, calculate your current cost as a percent of the revenue it drives. If your protein category is more than 28–32% of food revenue from protein-driven items, you have a price, portion, or vendor problem. You will find out which one in day four.

Day 4 — Renegotiate one vendor category

You cannot fix everything in 7 days. Pick one category — the largest dollar opportunity from day three — and run a structured negotiation cycle.

The script is simple:

1. Pull the last 12 weeks of invoices for the category. Calculate your average weekly volume.
2. Get a quote from a competing vendor for the same SKUs at that volume. This takes 24–48 hours.
3. Show the competing quote to your current vendor. Tell them you need their best price at current volume, and you would prefer to keep the relationship if the numbers work.
4. Ask about volume rebates, payment-term discounts, and category expansion (if you are using them for one category, would they sharpen the pencil if you moved a second category to them?).

Most operators recover 4–9% on a category in this single cycle. On a category that is $4,000 per week, that is $20,000–$45,000 of annualized margin. From one negotiation.

Day 5 — Adjust one scheduling block

Take the worst daypart from day two. Build a counter-schedule for the next two weeks based on demand, not habit. Specifically:

• Cut one labor hour from the daypart, in the band where sales-per-labor-hour was lowest.
• Cross-train one position so the cut does not show up at the guest level.
• Have the manager-on-duty document any service breakdowns during the trial.

This is a test, not a permanent change. The point is to see whether the labor hour was protecting service or padding the schedule. Most operators find out it was padding.

If service breaks down, you put the hour back. If it does not, you move on to the next daypart with a real data point.

Day 6 — Inspect waste and comp logs

Two systems are almost always under-used: waste tracking and comp logs. Pull both for the last 30 days.

For waste, look for:

• Pattern by daypart. Heavy waste at end-of-night usually means over-prep. Heavy waste at lunch transition usually means par levels are wrong.
• Specific items. One protein that wastes 12% of par every week is your most expensive guest. Either the par is wrong, the storage is wrong, or the menu is over-featuring it.
• Manager differences. Same items, same volume, different waste percentages by manager-on-duty. That is a training issue you can fix without changing anything else.

For comps, look for:

• Comp-per-cover trending up. This is one of the earliest signals of service drift. It often shows up before guest reviews do.
• Specific server patterns. One server who comps 3x the team average is either getting consistently bad food (a kitchen issue), getting consistently bad guest reactions (a service issue), or running an unauthorized loyalty program.

By end of day six you should have two or three specific waste or comp issues that have a name and a number on them.

Day 7 — Set a weekly KPI rhythm

The goal of the previous six days is not just to recover this quarter — it is to make sure you never run a 7-day triage again. Day seven is when you install the rhythm.

Set up a weekly KPI meeting, 30 minutes, every Tuesday morning. The agenda is the same every week:

1. Last week’s prime cost, with the variance from target named and explained.
2. Top three labor outliers by daypart.
3. Top three COGS outliers by category.
4. One waste or comp signal flagged for follow-up.
5. One vendor cycle in motion.

The meeting is short on purpose. Anything that takes longer than 30 minutes belongs on a different agenda. The point is rhythm, not depth — depth comes between meetings.

The 7-day triage checklist

• [ ] Day 1: Audit 4 weeks of invoices, flag categories moving >4%.
• [ ] Day 2: Pull labor by daypart, rank by sales-per-labor-hour.
• [ ] Day 3: Identify top 5 COGS offenders by absolute dollars.
• [ ] Day 4: Renegotiate one vendor category with structured competition.
• [ ] Day 5: Test a one-hour labor cut in the worst daypart.
• [ ] Day 6: Inspect waste and comp logs, find 2–3 specific issues.
• [ ] Day 7: Install the weekly KPI rhythm.

Prime cost recovery is not magic and it is not a single decision. It is a stack of small, defensible moves made on a consistent rhythm. The operators who do this for 90 days in a row pull two to four points of prime cost back without cutting anything that matters to the guest. The operators who skip the rhythm cut the wrong thing and get the same problem again next quarter.