Prime Cost Recovery Playbook: When You’re Above 65% and Need to Get Back

By The Pragmatic CFO. Last verified: 2026-09-03.

Prime cost above 65% for two months in a row means you are not making money on a dollar of new revenue. You are losing it. On a full-service P&L, once you clear roughly 32% for rent, utilities, R&M, insurance, marketing, and G&A allocations, a 65% prime cost leaves 3 points for you. Add another point of drift and you are underwater on the marginal dollar. Every additional cover costs you 30 to 40 cents to serve after variable and semi-variable costs. That is the math. Ignore it and you finish the quarter wondering where the cash went.

This is a playbook for the operator whose prime cost has crept above the ceiling and wants to walk it back. Six levers, in order.

Prime cost, defined properly

Prime cost = COGS (cost of goods sold, food + non-alcoholic beverage + alcohol) + fully loaded labor (wages + payroll taxes + workers’ comp + benefits + PTO + training), as a percentage of net sales.

Concept-specific healthy ranges (from the RBL master benchmark table):

• QSR: 55% to 60% healthy. Above 62% is a red flag.
• Fast casual: 55% to 60% healthy. Above 62% is a red flag.
• Full-service (casual, polished casual): 60% to 65%. Above 65% is a red flag.
• Fine dining: 62% to 68%. Above 70% is a red flag.

Why 65% is the structural ceiling for casual and 60% for QSR and fast casual: the remaining costs of running the box (occupancy, utilities, R&M, insurance, controllable expenses, above-store allocations) do not compress much regardless of concept. Rent is rent. A full-service concept structurally carries a higher labor load because of table service, so the ceiling is higher. A QSR strips out the labor and the check average is lower, so the ceiling has to be lower or the box does not clear.

The sequence: fix the leaks before you take price

Order matters. Take price first and you paper over the diagnosis. Do the diagnosis, then take price if you still need it.

Lever 1 (Week 1): tighten receiving and yield

Pull the last 30 days of invoices against inventory usage. Look for cost variance (invoice cost vs. theoretical cost per unit) and plate cost variance (theoretical recipe cost vs. actual food cost by category).

Common leaks:

• Cases short on delivery, credited late or not at all
• Substitutions accepted at higher per-unit cost without a menu-mix adjustment
• Portioning drift on high-cost proteins (2 oz to 2.2 oz on a steak spec is 10% of that plate’s cost)
• Waste log not run, so unfavorable yield gets buried in “usage”

Expected recovery: 0.5 to 1.5 points of COGS in 30 to 60 days if you had drift. If your COGS is running exactly to theoretical, this lever gives you nothing. If it is running 2 points over theoretical, this is your first and cheapest fix.

Lever 2 (Week 1 to 2): kill wage leakage on the schedule

Not a schedule redesign (that comes later). Just the leakage. Pull the labor detail report and look for four things:

• Punch-in and punch-out compliance. Employees clocking in 10 minutes early because “the app was slow.” That is 40 hours a week on a 50-person store, at $18 fully loaded, that is $720 a week or $37,000 a year.
• Unauthorized overtime. Anyone hitting time-and-a-half without a signed OT authorization. This is 100% preventable and 100% margin.
• Salaried-manager work not accounted for. When your GM is behind the bar for four hours on a Friday, that shift is uncovered in your labor model and your hourly labor % looks better than reality.
• Missed meal breaks or missed break penalties in break-mandate states. California is the obvious one; noncompliance turns into a wage claim and a check.

Expected recovery: 0.5 to 1.0 point of labor % in the first two payroll cycles.

Lever 3 (Week 2 to 4): reprice the top 10 items by menu mix

Not a blanket price increase. Item-level pricing based on contribution margin.

Pull menu-mix for the last 90 days. Sort by units sold. Take the top 10 items (this is usually 60% to 75% of your covers). For each, calculate:

• Current food cost % on that item
• Current contribution margin per unit
• Contribution margin per labor minute (as covered in the labor cost article)

Now cross-check against elasticity: which of these items have you not repriced in 12+ months? Those are your candidates. On each, take price to bring contribution margin back to concept norm. In casual, that is usually a $0.50 to $1.50 move per item. In fast casual, $0.25 to $0.75. Do it once, in a menu print cycle, not as a series of little bumps that erode guest trust.

Expected recovery: 0.5 to 1.5 points of prime cost. The mix matters: repricing the top 10 items reprices the majority of your revenue.

Lever 4 (Week 3 to 6): re-engineer the top 5 dogs

Dogs are items that (a) sell in real volume and (b) have below-average contribution margin per labor minute. They are your worst items to make. They are usually not your worst-selling items; a worst-seller is easy to kill because nobody misses it.

For each dog, three choices:

1. Kill it. Cleanest. Frees kitchen bandwidth. Simplifies prep.
2. Replace with an item that hits the same menu category at better contribution margin. Guests still see a lasagna on the menu, just a version that food-costs and labor-costs better.
3. Reduce the portion or restage the plate. Smaller portion at the same price, or a different garnish spec that removes labor minutes.

Expected recovery: 0.3 to 1.0 point of prime cost. Smaller than repricing because you are only touching 5 items, but it also frees kitchen throughput, which cascades into service time and labor.

Lever 5 (Week 4 to 8): renegotiate the top 3 vendor lines

Pull 90 days of purchases by vendor and by line item. Rank by spend. The top 3 line items are almost always some combination of: (i) primary protein (beef, chicken, or seafood), (ii) dairy (cheese in particular), (iii) produce.

For each of the top 3:

• Ask for a CPI-linked contract with a monthly true-up, not a fixed price that gets renegotiated when the vendor “has to.” Fixed pricing works for the vendor when the market drops. It hurts you when it rises.
• Get a competing bid from a secondary distributor. Even if you do not switch, having the number gives you a real conversation.
• Look at group purchasing organization (GPO) options. For a 5- to 25-unit operator, a GPO like Buyers Edge or Dining Alliance typically brings 3% to 8% off primary distributor cost after their fee.
• On produce, ask whether a local direct-from-farm option beats broadline pricing on the top 5 SKUs.

Expected recovery: 0.5 to 1.5 points of COGS. This lever takes longer to negotiate but rolls in as pure margin once the new pricing hits.

Lever 6 (Week 6+): shrink or split the menu

Only if levers 1 through 5 have not closed the gap. Menu reduction is a bigger operating change and takes more coordination.

The move: cut the bottom 20% of menu items by unit sales. That segment usually contributes 3% to 6% of revenue and consumes 15% to 25% of kitchen prep bandwidth and pantry SKUs. Concentrating prep on fewer items reduces waste, tightens yield, and lets you schedule less BOH labor.

Expected recovery: 1 to 3 points of both COGS and labor. The compounding effect is real, but do not do this lightly. You will get pushback from long-time guests on their favorite item. Manage the messaging.

Worked example: FSR concept from 68% prime to 63.5% in 60 days

A polished-casual concept, one location, $2.4M annual revenue, running the following P&L (pre-fix):

Levers pulled in order (60-day window):

• Lever 1 (receiving and yield): Portioning drift on ribeye (2.4 oz over spec on a 12 oz plate) and undercredited case shortages. Fixed with a scale-check protocol and a Friday invoice reconciliation. COGS down 0.9 points.
• Lever 2 (wage leakage): Two employees consistently punching in 12 minutes early; one manager working 6+ bar shifts a month uncoded. Fixed via biometric clock and coding manager bar shifts to hourly. Labor down 0.6 points.
• Lever 3 (top 10 reprice): $0.75 to $1.50 taken on 8 of the top 10 items in the next menu print. Guest complaints minimal. Prime down 1.4 points.
• Lever 4 (dogs): Killed the seafood pasta (14 labor minutes, contribution $0.68/labor minute) and reengineered the meatloaf (portion trim, 3 labor minutes saved). Prime down 0.6 points.

Total pulled: 3.5 points of COGS + 0.6 points of labor = 4.1 points on food cost accounting arithmetic, but the labor reprice loop also helps. Actual measured prime after 60 days:

Prime moved from 68% to 63.5%. On $2.4M revenue, that is $108,000 of annualized margin recovered without touching vendors or menu size. Levers 5 and 6 stayed in reserve for the next cycle.

What NOT to do

• Cut FOH labor blindly during service. You will drop check average and Yelp scores in one weekend and take three months to rebuild. Address FOH scheduling, not FOH headcount, first.
• Drop a supplier without a backup lined up. Empty walk-in on a Friday is a bigger problem than a 4% price gap.
• Take a blanket 10% menu price increase. Guests notice, elasticity bites, and you lose the diagnostic value of item-level analysis.
• Try to do all six levers at once. You will lose focus, miss what worked, and blame the wrong thing when the number does not move.

Cadence: how to know each lever is working

• Weekly prime cost report. COGS and labor as % of net sales, versus prior week and versus budget. If prime does not move within two weeks of pulling Lever 1 or 2, you have a measurement problem, not an operating problem.
• Menu mix report, weekly. Watch the top 10 items for volume drop after a reprice. A 2% to 5% unit drop on a repriced item is normal and margin-accretive. Anything larger is a signal to review.
• Vendor scorecard, monthly. Track top 3 vendor pricing against a baseline. Every 60 days, get one competing bid.
• Break the labor number into FOH and BOH weekly. Watch which side moved.
• Escalate to your COO, accountant, or fractional CFO if: prime cost does not move at all after 30 days of Lever 1 and 2 (measurement problem), or if guest count drops more than 5% for two weeks after a reprice (elasticity problem).

Prime cost is one number. It hides at least six different problems. Fix them in the right order.

Methodology footer

Concept-specific prime cost bands referenced in this playbook match the RBL master benchmark table published 2026-08-22. The worked example is a composite from operator engagements, not a specific chain filing.

External references: RBL master benchmark table (2026-08-22); Dailypay restaurant turnover data (dailypay.com/resource-center/blog/qsr-and-restaurant-turnover-rates) for turnover context on Lever 2.

Last verified: 2026-09-03.