How to onboard a new kitchen hire in the first two weeks

The cost of getting this wrong is not the recruiting fee. An undertrained cook produces waste, slow tickets, and remakes for months while everyone works around them, and nobody connects those costs back to a training decision made in July.

The most expensive employee in a kitchen is the one who was never properly trained and never left.

The price of doing this badly has moved, and most operators are still working from an old number. The National Restaurant Association estimates that average hourly earnings of restaurant employees are up 41% since February 2020. Every hour of training, every hour of a trainer's reduced output, and every hour a cook spends slow on a station has been repriced since the last time you thought about what a hire costs.

Why do new kitchen hires quit in the first two weeks?

Because nobody was responsible for them.

The usual first day: a hurried tour, a locker, and then go with Marco, he will show you. Marco has his own station and a full prep list. He shows them what he can while doing his job, so the new person spends the shift standing slightly in the way, learning that they are an inconvenience.

By day three they have a decision to make and nothing pulling them to stay. No relationship, no station they own, no idea whether they are doing well. People do not leave because the work is hard. Kitchen people expect hard. They leave because they are invisible.

The fix costs no money. Name the trainer, tell them they are the trainer, and take enough off their plate to make it real. An hour of somebody's prep list is cheaper than rehiring.

Tell the new hire what happens on day fourteen too. A person who knows a review is coming behaves like someone with a future here.

What should week one actually contain?

One station. Not a tour of all of them.

Depth beats breadth in week one. A cook who owns the fryer completely after five days is useful. A cook who has seen every station once is useful nowhere, and now believes they were trained.

Week one, in order:

• Day one: safety, sanitation, where everything lives, and shadowing the station without touching a ticket. Send them home on time.
• Day two: the five highest volume items on that station, made to standard, trainer watching every one.
• Day three: they run the station during the slow period, trainer beside them. First checkpoint at the end of the shift.
• Days four and five: they run it, trainer within reach, increasing volume.

The written standard matters more than people expect. Not a binder. One laminated page per station with portion, build, and plate for the five items that station sends most. Spec drift is a cost problem before it is a quality problem, and it starts with each new hire learning the recipe from whoever was standing there.

That is one of the quiet ways prime cost creeps, because both halves of it move when a station is run by someone still figuring it out.

What does week two look like?

Volume and speed, then a second station only if the first one is solid.

The day seven checkpoint is the one people skip and the most important. Ten minutes, three questions: what is confusing, what slows you down, what do you need. Then fix one thing they name before day nine. Fixing something a new hire asked for is the strongest retention move you have, and it usually costs a container or a shelf.

In week two the trainer steps back deliberately rather than gradually. Tell the new hire that today they are on their own and you will be watching, then actually watch. Being observed and told you did fine is how confidence gets built.

Do not schedule their first full Friday until they have handled a busy Thursday. The gap between a moderate night and a real one is where undertrained cooks decide the job is not for them. Which shift counts as a real one is a forecasting question, and if you are guessing at volume, staffing to a forecast instead of a feeling comes first.

Day fourteen is a real conversation with a decision in it. Are they on their station, are they meeting standard, what is the next station, and when. Say the plan out loud. A cook with a named next step stays longer than one told they are doing great.

How do you know the training worked?

You measure it, and the POS tells you before anyone reports it.

Compare ticket times for the station on the nights that person works against nights they do not. Then remakes and voids on their items, and waste on the station. That needs no new software, only knowing that the POS holds more than the sales total and pulling the report by employee and item.

Give it four weeks before judging anything. Week one and two data reflect training, not the person.

The comparison is the point. A station that runs slower on Tuesdays is telling you something about who works Tuesdays, and it is not always the new person. Sometimes it exposes the trainer.

What do the first two weeks actually cost?

Price it once and the three checkpoints stop feeling like overhead. Use your own loaded hourly costs in place of these, meaning wages plus payroll taxes plus whatever benefits you carry.

Two weeks of onboarding, done properly:

That is what a kitchen hire costs before they have produced a full shift of value. It is not a recruiting fee, it does not appear on any invoice, and it is already on your labor line whether or not anyone named it.

Now lose them on day nine. You have spent roughly a thousand dollars of that $1,452, the station goes back to being covered, and you start the whole number again with the next person.

Do that three times in one year on one station and the cost is $4,356. The Association puts the pre-tax margin of a typical independent restaurant near 5%, which means $4,356 of profit is what a restaurant doing roughly $87,000 in annual sales produces in a year. One station, three failed starts, and the equivalent of a small restaurant's entire year of profit.

Then add what the survivor costs. An undertrained cook who stays produces remakes, waste, and slow tickets for months. Six remakes a week at $7 of food cost is $42, which is $2,184 a year, and that ignores the ticket time and the comped desserts the floor hands out to cover it.

Against those numbers, an hour off somebody's prep list is not a cost. It is the cheapest line item in the whole exercise.

Is the turnover just what this industry is?

No, and the belief that it is comes packaged with a statistic that has been dead for twenty years.

The claim is that 90% of restaurants fail in their first year. It is false, and it was falsified in a peer reviewed journal. H.G. Parsa and colleagues, publishing in the Cornell Hotel and Restaurant Administration Quarterly in August 2005, found the 90% figure had become widely accepted with no data supporting it. The measured first year failure rate for independent restaurants was 26.16%. Second year 19%. Third year 14%.

The disproof in the paper is the part worth carrying. If 90% of restaurants failed in year one, simulation shows a market of 1,500 units shrinking to 254 units in twenty years, a loss of 84% of existing restaurants. That is the opposite of the observed industry, which has grown continuously.

Why this belongs in an article about training: fatalism is a management style, and it is expensive. An operator who believes the business is a coin flip and that cooks leave because cooks leave will not spend an hour building a station standard. The real first year number is about one in four, which is a serious operating fact and a solvable one. Week one attrition is the same shape. It is not weather. It is a design decision about whether anybody was responsible for the new person.

There is one honest data point on pay and staying. California's AB 1228 set a $20 an hour minimum for fast food workers at chains with more than 60 locations nationwide, effective 1 April 2024. Economists disagree sharply about its employment effect, and anyone quoting you a confident job loss number from that law is taking a position rather than reporting a result. But both camps report the same finding on one point: the separation rate fell. Pay moves retention. It does not replace being trained by someone whose job that was.

Where does this break down?

Two places, and both are common enough to name.

When the trainer is the problem. A strong cook is not automatically a trainer, and a resentful one teaches the new hire on day one that this station is a burden. If two consecutive hires leave from the same station under the same trainer, you have learned something about the trainer, not about the labor market.

When the standards page describes a restaurant you do not run. A spec copied from a book or from your last job produces a cook trained to something the line does not actually do, and now they are wrong twice: once against the page and once against the person next to them. Write the page from five plates you weighed off your own pass.

What to do this week

Write the one page standard for your busiest station. Five items, portion, build, plate, with a photo of each. That is one afternoon and it outlives every hire you make.

Name a trainer for your next hire in advance and tell that person now. Take one thing off their prep list on training days.

Put three checkpoints on the calendar the day you make the offer. Ten minutes each, all three actually held.

At day seven, ask what slows them down and fix one answer within two days.

Pull the station report four weeks in. If ticket times and remakes look normal, the training took. If not, the problem is the training rather than the hire, and you know it before it costs you a season.

Questions about your first two weeks?

Email me at eric@seod.com with what a new cook's first shift looks like in your kitchen right now, in a few sentences. I will send back what I would change about day one and day three, and the one page standard I would build first for your menu.

Sixteen years running multi unit restaurants and more than $54M in annual P&L means I hired and lost a lot of cooks, and the ones I lost were usually lost in week one. Judging a change on a fixed window applies elsewhere, including how to measure the first ninety days of any marketing.

Hiring is also easier when people are applying, which starts with being findable. That is what how restaurants actually get found on Google in 2026 covers.

Otherwise, more on running the operation is here.