Process standardization in a restaurant: before and after writing the manual

Step 1 · Map your critical tasks and put a number of minutes on each one

The first deliverable isn't a manual, it's one sheet listing the 12 to 15 tasks that wreck your shift when they go wrong, each with a target time in minutes and a named owner. Sit down on a Tuesday morning with your chef and your floor manager, walk the whole flow from receiving goods to closing the register, and actually time things: plating a hot dish, resetting a table, counting the walk-in. With those numbers in hand you stop arguing about perceptions. Outside references help you calibrate: The Restaurant HQ estimates 45 minutes for a two-person lunch and 90 for a six-top at dinner, so if your turnover runs two hours for two guests, the bottleneck is inside your house and that map will point at it before Friday. A spec sheet is finished when a brand-new cook produces the dish without asking anything, and you verify it by weighing three consecutive portions with under 5 % variance between them.

Step 2 · Write the spec sheet with gram weights, waste and cost per portion

Every sheet carries ingredients in grams, not spoonfuls or «to taste», the waste percentage of each input, the yield in portions and the resulting cost, which has to land at 32 % or below on the selling price. That 32 % is a ceiling, not a goal. Start with the 20 dishes that move 80 % of your sales and get them signed off before touching the rest of the menu. There's an upside almost nobody counts: UNEP measured 290 million tonnes of food wasted by food service in 2022, and written gram weights are your first line of defense against your share of that number. Every spec sheet needs its short video, and the rule is strict: 40 seconds, fixed shot, hands and product, no music and no voiceover. Diego F. Parra pushes hard on this step because paper teaches WHAT goes in and video teaches HOW it goes — the angle of the knife, the texture of the sofrito when it's ready, the height of the plating.

Step 3 · Shoot the 40-second video that replaces the hallway explanation

Shoot them on a phone during a normal shift, never in a staged session, and file them in a shared folder under the dish name. You verify the deliverable like this: a cook who has never made that dish executes it after watching twice, and the result passes the chef's tasting. And yes, there's a commercial prize: a library of 20 or 30 real clips feeds the restaurant's social channels for months with zero new production, which is exactly what the Masterestaurant method calls standardizing before spending on marketing. The weekly count turns your standard into cash-register information, and its deliverable is a 20-line sheet signed every Monday before 10 a.m. Don't count the 300 SKUs sitting in storage: count the 20 inputs that concentrate most of the spend — proteins, cheeses, spirits, oils — and assign each an explicit tolerance, 2 % on portioned product and up to 4 % on bulk.

Step 4 · Install a weekly count of your 20 A-items with a stated tolerance

What matters is what happens when that tolerance breaks. Without a count, you find out food cost ran to 39 % when the accountant closes the month, forty days late and with the money already gone; with Monday's sheet, the deviation is seven days old and you can still trace it back to the shift, the person and the dish that produced it. Your service standard gets written in minutes per stage and printed at the pass: greeting under 2 minutes, order taken in 5, starters in 8, mains in 18, check in 4 from the moment it's asked for. Without published numbers, «fast service» means one thing to the lunch server and something else to the night crew, and neither matches what the guest expects. Time one full evening with a stopwatch and the spread between tables will surprise you.

Step 5 · Set service times per stage and post them where the team can see

That discipline also protects revenue on the reservations side: OpenTable reports deposits cut no-shows by 57 % and prepayment by 44 %, and a policy like that only holds up if the team sustains a time standard that justifies charging in advance without embarrassing anyone. Mistake number one is the 60-page manual: nobody opens it, it ages in three months and it hands the owner a false sense of having solved the problem. Write one page per process, dated, and replace it when things change. The second mistake is standardizing without measuring, which produces handsome documents and zero correction. The third is leaving the criteria inside your most senior cook's head: you're paying an emotional bonus without knowing it, and the day he quits he walks out with the sofrito. The fourth, and the most expensive, is running campaigns before the process holds, because any ad that brings new people into an inconsistent operation only speeds up the first one-star review.

Step 6 · Four mistakes that sink a standardization, and how to dodge them

With 62 % of operators naming staff retention as their biggest obstacle according to the National Restaurant Association, a team turning over every four months NEEDS the criteria written down. With standardized processes you lose speed, you don't lose the dish. That's the real exam, and it's worth simulating before it happens on its own: pick a slow Tuesday, pull the chef off the pass for four hours and let the sous run it on spec sheets and videos. If food cost that shift stays inside range and mains leave in the agreed 18 minutes, the system holds. If the shift collapses, you just found out for free what you'd otherwise learn on a Saturday at 9 p.m. with 120 covers sold. Here's the paradox of the trade: writing the criteria down doesn't strip value from your good cook, it gives it back, because he stops being worth something for being the only one who knows and starts being worth something for his speed and his head, which is what a salary pays for rather than a hostage situation.

Closing · The seven-box list that confirms it landed

You know standardization is done when you can tick these seven boxes without cheating: the 20 spec sheets for your A dishes are signed with cost under 32 %; each one has its 40-second video in the shared folder; the weekly count sheet has arrived four Mondays running before 10 a.m.; service times are printed at the pass and met on 8 of every 10 tables measured; every process has a named owner; each page carries an effective date and a next-review date; and an employee with under 30 days on the job executes a critical task using the document alone. Miss a single box and you aren't standardized, you're documented, which is a different animal. Start this Monday with the count of your 20 A-items: it's the box that pays back fastest. The first difference is where judgment lives.

The four differences that actually move cash

In a memory-run operation it sits in the head of your most senior cook, and you are quietly paying an emotional premium, because the day he resigns he walks out with the recipe and the rhythm of the pass. With process standardization that judgment moves onto the spec sheet and into the 40-second video, so your best person stays valuable for speed and brains, not for being the only one who knows how much fat goes into the sofrito. Second comes the speed of correction. Without a standard, you learn food cost drifted to 39 % when the accountant closes the month, forty days late and with the money already spent. Weekly counts on 20 A-items plus a current spec sheet surface the same drift the following Monday, across seven days of purchasing, before it eats an entire quarter's margin. Marginal efficiency lives exactly there, in the gap between fixing seven days of damage and fixing forty.

The four differences that actually move cash — in practice

Third is what training costs you. A 45-day informal shadowing ramp means your head chef produced at half capacity for six weeks explaining the same thing three times. Video manual, operating checklist and a signed evaluation cut that ramp to 12 days and hand the head chef roughly 60 hours of real work back per hire. Multiply by your turnover and the project pays for itself, before counting the quality you stop losing along the way. The fourth difference is the one almost nobody connects, and the one that interests me most from the content side: a standardized operation is permanent footage. Once BOH labeling follows FIFO, the line stays clear and plating follows a reference photo, any phone held vertically captures usable takes during live service. Restaurant accounts that sustain growth on TikTok and Reels don't produce more. They simply run a kitchen that survives a camera switched on without warning.