Most of the operators I talk to have the same complaint about talent. They cannot find good people, they cannot afford to train them, and the ones they do find leave. Every conversation eventually gets to the moment where someone says, half serious, that if only they had a proper training budget, everything would be different.
I want to push back on that. The last three general managers I promoted at the Zareen's group came up from inside the units. Two started as line cooks. One started at the counter. None of them ever attended a class on management. What they got instead was about 90 days each of deliberate on-shift development, from operators who were already on the schedule, using shifts that were already happening. Total incremental cost, close to zero.
The framework is small and boring. Shadow. Teach-back. Stretch. Solo. In that order. It costs nothing. It works. And it beats most of the formal programs I have seen at larger chains, because it develops people on the actual work rather than on a simulation of the work.
Why the "no budget" excuse is almost always wrong
The industry loves to blame budget. In practice, the operators I know who complain loudest about training budget also do not use the free tools sitting in front of them. The pre-shift is a training moment. Every 15-minute post-shift debrief is a coaching moment. Every ordering cycle is a management moment. Every schedule build is a decision-making moment.
What is actually missing is not money. It is deliberateness. Nobody has decided that Ana, the prep lead who has been running mise for a year, is going to spend Wednesday nights this month shadowing the closing manager. Nobody has decided that Jae-min, the sous chef, is going to run the monthly food-cost meeting in November. Those decisions cost zero dollars. They also do not happen unless somebody puts them on paper.
Fig. 1 · Shadow, teach-back, stretch, solo.
Step one: shadow, and do it twice
Shadow is exactly what it sounds like. The person you are developing spends a full shift alongside someone doing the job one level above them. They watch, they take notes, they stay out of the way. At the end of the shift you spend 15 minutes debriefing. Three questions. What did you see them do that you would not have done. What did you see them do that you did not understand. What decision did they make that surprised you.
Two shifts, not one. This is the part everyone gets wrong. The first shadow shift is always a distortion, because either the shift is a slow Tuesday and nothing hard happens, or the shift is a wild Friday and the person you are shadowing goes into autopilot mode and cannot explain anything. You need one of each. Slow so they can talk. Busy so the person shadowing sees what the job actually is under load.
Cost of two shadow shifts. Zero. The senior operator was already working. The junior operator was already scheduled. The only thing you added was intention.
Step two: teach-back, the step everyone skips
This is the most important step in the ladder. It is also the one operators skip because it feels awkward. Do not skip it.
After the person has shadowed the task twice, they teach it back. Not to you. To a third person who does not know it. The new hire on prep. The dishwasher who is angling for a promotion. Whoever. Someone who will actually ask questions, because the questions are where the real learning happens.
What teach-back exposes is the gap between "I watched somebody do this" and "I understand this well enough to run it." Those are two very different states. When a prep lead has to explain to a new dishwasher why the walk-in gets a first-in-first-out rotation instead of a nearest-item rotation, they either understand the food safety and shrink implications or they do not. If they do not, they will bluff, and then the new dishwasher will ask a follow-up question that pops the bluff, and you will find out in real time exactly where the person's understanding stops.
Training is what you do to somebody. Teach-back is what they do to somebody else. The difference is where the learning lands.
Teach-back also happens to be free labor for you. The new hire is getting trained by the senior person. The senior person is getting developed by the act of teaching. Two people are learning. One shift is producing double the human capital it usually does. And you are learning, in real time, whether the person you are developing is ready for the next step.
Step three: stretch, one size larger than what they have done
A stretch assignment is a piece of the next job, dropped into the current job, with a real deadline. The word "real" is doing all the work in that sentence. Fake stretches do not develop anybody. If the person knows that you will fix it if they mess it up, they do not learn anything, because they never really own it.
Good stretch assignments have four traits. One size larger than the current job. Real deliverable. Real deadline. Small enough that if it goes poorly, the unit survives.
Some examples that I have used, at Hana and at Zareen's:
- Prep lead: own inventory for one week. Not just count. Order too. Reconcile the P&L impact. Present to the general manager on Monday.
- Shift lead: build the schedule for one week. With the demand curve you show them. Publish it. Own the phone calls when someone calls out.
- Sous chef: run the monthly food-cost meeting. Prepare the deck. Walk the general managers through the numbers. Take the questions.
- Assistant general manager: present the monthly P&L to the district manager. No safety net. The GM sits in the room and does not speak.
All four of those are already work that has to happen. You are not creating extra tasks. You are just letting the person one level below do them on a defined cycle, with you close enough to catch a real disaster but far enough that they own the outcome.
The trap: never taking your hands off
The most common failure mode in the stretch phase is the operator who assigns the stretch but keeps doing the work themselves anyway. They sit in the pre-shift and "help." They rewrite the schedule "just a little." They comment on the P&L slides at 11 pm the night before the meeting. Every intervention like that empties the stretch of its value. The whole point is that the person feels the weight of the decision. If you carry it for them, they never feel it.
Step four: solo, and you are not in the room
The last step is the tell. Same task, but you are physically not there. Reachable by phone if the building is on fire. Not in the room. Not watching. Not correcting.
Then, same day, a 30-minute debrief. Three questions. What worked. What did not. What would you do differently next time.
The solo phase is where the promotion decision actually gets made. Not on tenure. Not on technical skill. Not on how much the person wants it. On whether they can hold the seat when you are not backfilling. I have had strong technicians fail this step and never advance, and I have had quiet, unassuming line cooks pass it clean and become the best general managers in the region. You do not know until you take your hands off.
A vignette from the Zareen's Palo Alto turnaround
The general manager I promoted at Zareen's Palo Alto in month eight of the turnaround started the year as a shift lead. Trilingual, English and Spanish and Urdu, a strong operator with two years of tenure, but not on anybody's list as a promotion candidate. The reason was that the existing culture had never given her anything to prove.
We ran the ladder over about 90 days. Two shadow shifts with the outgoing GM in November, including a Sunday brunch that went sideways in a way that turned out to be perfect for the diagnostic. Teach-back on the closing walkthrough to a new prep hire in December, which surfaced two SOP gaps the outgoing GM had been quietly working around for years. Two stretch assignments in January, running inventory for two weeks and building the schedule for one. Solo run for a full week in February, with me at another location.
The solo week beat the shift-lead-average P&L by about four points of labor and two points of food cost. That was the decision. She got the promotion in March. She is still there, still running that unit, and the last three assistant managers she promoted were developed on the same ladder she came up on.
What I got wrong the first four times
I want to be honest about the mistake, because it took me years to see. For a long time I ran what I thought was on-shift development but was really just "let the ambitious person help me do my job." I was doing the shadow phase without knowing it. I was skipping teach-back entirely. My stretches were fake, because I was always in the room. And I never made a clean solo call, because I did not have the discipline to leave.
The result was predictable. I would promote people who had watched me do the job and could not do it themselves. Then I would blame the person for not being ready, when really I had never let them prove that they were.
The fix was structural. The ladder goes on the calendar, in writing. The shadow shifts are named. The teach-back sessions have a recipient. The stretch assignments have a deadline and a deliverable. The solo shift has a date on which I am physically at another location. When any of those pieces are informal, the whole ladder collapses back into "the ambitious person helps me." When they are written down, the person actually gets developed.
What this looks like at scale across a region
The ladder above works at a single unit. What changes when you run it across 21 units, as we did at Hana, is that the coordination becomes the whole game. You cannot personally coach every prep lead through their teach-back cycle. You do not need to. What you need is a district-level rhythm where each general manager runs the ladder on two or three people in their own building, and reports up once a quarter on who is where in the sequence.
The quarterly bench review takes 20 minutes per unit. Every general manager brings a one-page grid. Each person on the development track lists which step they are on, what their stretch assignment was, and when the solo shift is scheduled. That is it. No slides. No jargon. Twenty minutes a quarter times 21 units is seven hours of my time per quarter, and it is the single highest-leverage thing I did in the Hana Group years. It replaced almost every external recruiting cycle we would otherwise have run for shift lead and assistant manager roles.
The point
You do not need a training budget. You need 90 days, a name on the schedule, and the discipline to take your hands off the work at the right moment. The best operators I have promoted came up on the shifts that were already on the calendar, using tasks that already had to happen. All I did was decide, in advance, who was going to do them.
Talent development is not a program. It is a set of decisions about who runs what next. Make those decisions on purpose, and your bench builds itself.