Wine Training for Servers That Actually Sticks
Hey Team!
Wine training is where a lot of restaurants waste a full afternoon and get almost nothing back. You know how it goes. GM pulls the staff together on a Tuesday, opens a couple bottles, talks through the wine list for thirty minutes, maybe forty-five if someone asks questions. Everyone nods. A few people take notes on their phone. You feel good about it. Then Saturday night hits, a guest asks your server what pairs with the salmon, and that server says "I think the Pinot Noir?" with an upward inflection that tells the table they have no idea.
That moment right there is a training failure. But it's not a failure of effort. It's a failure of method.
Here's the thing about how memory actually works. A single exposure to information, no matter how good the presentation, fades fast. We're talking 50 to 80 percent forgotten within 48 hours. That's not my opinion, that's Hermann Ebbinghaus and his forgetting curve from the 1880s. Nothing about human memory has changed since then. What that means for your wine training is brutal: by the time your server is standing tableside on a busy Saturday, most of what they heard on Tuesday is gone. Not because they weren't paying attention. Because that's just how brains work.
The thing that actually works is spaced repetition. Small doses, repeated over time, with enough gap between exposures that the brain has to work to retrieve the information. That retrieval effort is what makes it stick. Five questions a day for two weeks will outperform a ninety-minute seminar every single time. Not a little bit. By a lot.
Think about what five questions a day actually looks like in practice. Monday morning your server gets: "Name one white wine on our list that pairs well with fish." Tuesday: "What's the difference between our Pinot Noir and our Cabernet Sauvignon in terms of body?" Wednesday: "A guest orders the ribeye. Which red would you recommend and what would you say about it?" These aren't trick questions. They're the exact conversations your servers need to have at table. You're just rehearsing them before the shift instead of hoping muscle memory shows up on its own.
At Fat Tommy's, we added a walnut-crusted salmon as a seasonal special a few months back. Beautiful dish. But we had four different white wines on the list and two of them were terrible with it because of the oak. Our servers needed to know that. Not in the abstract, not as a bullet point on a printed sheet they'd lose. They needed to know it the way you know your own phone number. So we built a short question set around that specific dish. "Which of these two whites works with the walnut crust and why?" We asked it five days in a row in slightly different ways. By the end of the week, every server could answer it cold.
That's the other piece people miss. Generic wine knowledge is almost useless. Knowing that Pinot Noir is lighter-bodied than Cabernet Sauvignon is a fact. Knowing that the Pinot Noir on *your* list, from *your* specific producer, is a gorgeous match for your seared salmon because it won't fight the delicate fat content of the fish — that's a selling tool. The training has to be anchored to your actual menu, your actual bottles, your actual price points. Otherwise you're just running a wine appreciation class.
Cab versus steak is a perfect example of this. Every trained server has heard that Cabernet Sauvignon pairs with red meat. Fine. But can they explain *why* to a guest who's curious? The tannins in a big Cab bind to the proteins in a fatty cut. It's a textural thing as much as a flavor thing. When a guest orders the ribeye and your server says "the Cabernet's going to be great with that — the tannins cut right through the richness," that guest feels taken care of. They trust the recommendation. They order the bottle. The check goes up. That's the difference between a server who absorbed wine knowledge and one who sat through a seminar and forgot it.
Now, the two-week spaced repetition approach only works if the questions are actually good. Vague, multiple-choice trivia won't do it. The questions need to simulate the tableside moment. "A four-top orders two steaks and two salmon. How do you navigate a bottle recommendation?" That's a real question. That's what the shift actually looks like.
This is exactly the problem I was trying to solve when I built ShiftTrained. Not wine training specifically, but the broader failure of one-time lecture training across every category. Upload your menu, get quiz questions your staff takes on their phones, five minutes before a shift. The repetition happens automatically. The GM doesn't have to host another seminar. And because it's on their phones, staff actually do it. We see people at Black Barrel voluntarily retaking quizzes on their own time because they want to know the list cold. That's not me pushing them. That's what happens when training feels useful instead of obligatory.
The bottom line is this. Your wine list is a revenue tool. Every bottle your staff sells confidently is money that wouldn't have come in otherwise. But that confidence doesn't come from a Tuesday afternoon tasting. It comes from five good questions a day for two weeks, tied to your actual menu, asked in a way that mimics the real conversation. Run that for one full menu cycle and then watch your wine sales on a Friday night. The difference will be obvious.
Have a great day! — Terry
Your AI Restaurant Guy
