Stop Watching New Hires and Start Measuring Them (The First-Week Metrics That Actually Matter)

I've opened more than twenty concepts. I've watched confident GMs greenlight trainees who then torched a Saturday service because they didn't know the menu cold. I've watched nervous trainees get held back an extra week when they were actually fine. In both cases, the manager was doing their best with zero objective data. That's the real problem, not bad judgment, but no measurement at all.

Here's what I've come to believe after thirty years of this: if you're not measuring something specific during week one, you're just crossing your fingers. And in a business where Cornell's research puts the cost of a single hourly turnover at nearly $5,864, crossed fingers are expensive.

So let me give you three benchmarks. Not a philosophy. Not a framework. Three concrete things you can actually check.

The first one is an 85% floor on menu fundamentals. That means a quiz, not a conversation, not a verbal rundown, an actual quiz with a score. Somewhere north of 85% correct on the core menu items, descriptions, and preparation basics. Why 85? Because below that threshold, the server doesn't know the menu well enough to answer a guest's follow-up question with any confidence, and a server who hedges or stumbles on a menu question loses the table fast. The downstream effect is real. When a guest asks "what's the difference between these two sandwiches?" and your new hire says "I think one has peppers?" that's a comp waiting to happen. An 85% quiz score isn't a guarantee, but it's a line in the sand that correlates with basic confidence on the floor. And a quiz is objective. Your gut is not.

The second benchmark is the five-dish cold description. No menu, no cheat sheet, no peeking. Pick five dishes, any five, and ask the new hire to describe them to you the way they'd describe them to a guest. What's in it? How's it cooked? What does it taste like? If they can do five of those from memory with real specificity, they've done the work. If they can't, they haven't. This one matters because early turnover almost always traces back to the first few shifts. The new hire goes out on the floor, blanks on a dish, feels embarrassed, and decides this job isn't for them. Or worse, they fake it, the table notices, and the manager hears about it. Either way, you're losing them in week two or week three. The five-dish check gives you a real signal before that happens, not after.

The third benchmark is the allergen scenario. One scenario, correctly answered. Not "do you know our allergens?" Not a list recited back to you. A real scenario. Something like: a guest tells your new hire they have a tree nut allergy and asks if the house salad is safe. What does the server do? The right answer isn't "yes it's fine" or "I think so." The right answer is "let me check with the kitchen right now." That's it. If your new hire gives you that answer, they've internalized the right instinct. If they guess, you have a problem. The downstream on this one isn't a comp or a bad review. It's a liability. An allergic reaction that traces back to a server who didn't know to double-check is a lawsuit and, more importantly, someone's health. You need that scenario answered correctly before someone hits the floor. Period.

Now here's the part nobody talks about. You don't need a dedicated trainer to run this. You need one page.

A simple scorecard, literally one page, that the GM fills out by end of day three. At the top: the new hire's name, start date, role. Then three rows. Row one: quiz score, pass/fail with the actual number. Row two: five-dish cold description, names of the dishes attempted, a yes/no on whether they passed each one. Row three: allergen scenario, which scenario was used, pass/fail. At the bottom: one line for the GM's notes and a box to check: ready to solo, needs one more shift, not ready.

That's the whole thing. It takes five minutes to fill out. It creates a paper trail. And it takes the decision out of pure gut territory without making training feel like a corporate HR exercise.

The scorecard works because it changes the conversation. Instead of "I think she's ready," the GM says "she scored 91, nailed four of five dishes, passed the allergen scenario, she's good." Or "he scored 78, struggled on descriptions, we need another day." Those are conversations you can actually act on.

One last thing. If you're running a staff where the training documentation lives in someone's head and the readiness call is made by feel, you're not measuring turnover at all, you're just experiencing it. The first week is where you lose people or you keep them. A scorecard won't fix a bad hire. But it will stop you from pushing unprepared hires onto the floor and then wondering why the first three months are always a mess.

Measure the week. The rest gets easier.

Have a great day! — Terry Psaltakis
Your AI Restaurant Guy

For more on putting this into practice, see how ShiftTrained approaches wine training.