On my second opening the training was beautiful. Two-week curriculum, printed binders, station-by-station walks, an experienced trainer from headquarters. Opening week went clean. Ticket times were tight. Guest feedback was strong. The corporate trainer flew home on day 12.
By week six the standard had drifted. Ticket times were up 40 percent. Plating was inconsistent. Two of the strongest openers were doing things differently than they had been taught, and the new hires who came in after opening were being trained by them, so the drift was compounding.
The training had not failed. The training had been dependent on the trainer. When the trainer left, the training left with them. That is the failure mode I see in almost every opening, and it is the failure mode this article is about.
Why standard training decays
Three reasons, in order of severity:
- The trainer was the standard-holder. The local team was watching the standard, not owning it. When the trainer left, the local team had to invent their own version of the standard from memory.
- Nobody trained the trainers. The opening team knew how to do the work. Nobody taught them how to teach it. So the new hires that arrived in week four got a diluted version of what the opening team got.
- The training was a phase, not a rhythm. Training happened before opening. After opening, it was assumed to be done. But new hires kept coming in and the operating environment kept evolving, and there was no formal mechanism to update the training in real time.
Fix these three and the training holds. Skip them and it will not, no matter how good the pre-opening curriculum was.
The 21-day pre-opening curriculum
The minimum useful pre-opening training window is 21 days of paid, in-space time. Anything shorter and the team is learning on paying guests. Break the 21 days into three phases of a week each.
Days 1 to 7: SOP walks
Every SOP in the brand, walked in the actual space with the actual equipment. Two per morning, two per afternoon, four per day. By day seven the team has walked every single SOP at least once. This is not menu training. This is procedure training: how a line check runs, how a temperature log gets filled in, how an order gets received, how the cleaning schedule works, how a comp gets documented.
Days 8 to 14: menu execution
Every dish, cooked and plated by every station, at increasing pace. Day 8 is slow and careful. By day 14 the kitchen is firing tickets at 80 percent of expected peak volume. Timing is being measured on every dish. Plating is being photographed and compared to standard on every plate.
Days 15 to 21: integrated service
Front of house and back of house together. Full simulations. Tickets fired in real time, expo'd, run to tables, checks presented. Three soft opens fall inside this window: team-only at day 15, friends and family at day 18, public reservation at day 21.
Fig. 1 · The 21-day training arc.
Train the trainers before you train the team
The mechanism that makes training hold past week six is train-the-trainer. Here is how it works:
Every SOP is taught twice. First by the opening crew to the local team. Then, within 48 hours, by the local team back to the opening crew, for verification. Anyone who cannot teach the SOP has not learned it. The people who can teach are the people who will hold the standard when the opening crew leaves.
This adds about 30 percent to the time budget for training. It is the single highest-leverage 30 percent you will spend in the whole opening. Without it, you have taught the team how to execute the standard. With it, you have taught the team how to teach the standard to the next wave of hires, which is the entire game.
The people who will hold the standard in week seven are not the ones who executed it best on opening day. They are the ones who could teach it best on day 14. Identify those people early. Promote them into training roles.
The general manager designs the training
The single biggest predictor of whether training holds is whether the general manager designed it. Training that is handed to the GM by corporate, in a binder, to execute against, decays fast. Training that the GM built from the corporate materials, with the specific local team in mind, holds.
The reason is simple. The GM is the person who has to live with the training outcome. If they designed it, they know why every piece is there, and they can defend it in week seven when someone asks why a particular step matters. If they inherited it, they cannot defend it, and it drifts.
Practical version: give the GM the brand training materials at T-75. Ask them to build the 21-day plan by T-45. Review it with them, do not rewrite it. Their name is on the plan. Their judgment is in it. They will teach it like they mean it, because they built it.
The post-opening reinforcement phase
Training does not end at opening. This is the mistake almost everyone makes. The 21-day pre-opening plan is one third of the actual training program. The other two thirds happen after the doors open, in the first 40 to 60 days.
Weeks 1 to 2: daily post-shift huddles
After every shift, 15 minutes, whole team. What went well. What broke. One thing to fix tomorrow. The GM runs it. The training crew from existing units attends but does not lead. This is where the local team starts owning the standard.
Weeks 3 to 4: recorded refresher walks
Every SOP walked again, at speed, with the local team leading. Recorded on phones, filed to a shared drive. When a new hire comes in during week eight, they watch the video before their first shift. The video is the trainer, and the trainer is the local team, not corporate.
Weeks 5 to 6: cross-training rotations
Every station gets a second person trained on it. If your line cook on the grill quits in week nine, someone else can step in without a scramble. Cross-training is the resilience layer that keeps the operating rhythm from breaking when one person leaves.
The one metric that tells you if training worked
Everyone tracks a lot of numbers during an opening. Cover count, labor as a percent, food cost, guest complaints. All useful. None of them tell you if the training held.
The metric that tells you: ticket time from fire to plate, measured in week seven and compared to week two.
If ticket time in week seven is within 10 percent of ticket time in week two, the training held. The team is running the same operating rhythm they were trained on. If ticket time in week seven is 25 percent slower than week two, the training decayed. The team is running a different, slower version of the standard, and every week the drift will compound.
This one metric will tell you whether you need to run a mid-cycle training reset. Which is a thing, and if you have to do it, it is much better than pretending everything is fine.
What went wrong on that second opening
Going back to the story I opened with. The training was beautiful because the trainer was excellent. The trainer taught. The team executed. The trainer left. The team drifted. Almost none of the team could have taught what the trainer taught, because the trainer never let them try. The whole training program lived in one person's head.
What I do differently now: the corporate trainer arrives with a stack of empty training scripts. They fill in the scripts with the local team during the training week. The local team then teaches the scripts back. By the time the corporate trainer leaves, the scripts are in the GM's hands and the local team has taught every one of them at least once. When the next hire arrives in week eight, the local team pulls out the script and teaches. The standard holds.
The point
Pre-opening training that holds past week six is not about better content. It is about a different design principle. The people who will hold the standard in month two are the people who were trained to teach in month one. Build train-the-trainer into every phase. Let the GM design the plan. Do not treat training as a phase that ends at opening. Track ticket time in week seven against week two.
The units where training holds are the units where the plan assumed the trainer was going to leave, and built the transfer into every step. The units where training decays are the ones where the plan assumed the trainer was going to stay. Cadence beats charisma. So does design for hand-off.