Ask most operators about their bench and you will get a name for the general manager seat and a pause for everything else. That pause is the whole problem. The general manager seat almost never turns over first. The seat that turns over first is the Saturday closing lead, the morning host, the sous chef who just got a better offer down the street. If you cannot name the first and second successors for those seats without checking a spreadsheet, you do not have a bench. You have a hope.
I learned this the way most operators learn it. By missing a shift because I could not fill a seat.
The Sunday morning that built the chart
A few years into my Hana Group run, we had a Sunday morning where the opening lead at one of our Sam's Club kiosks called out at 5:15 a.m. with a legitimate emergency. The store opened at 8. Prep for the case had to start at 6. The closing lead from Saturday was not answering the phone. The general manager was two hours away. The area director on call, me, was three hours away.
We opened at 9:20 that morning. We lost the entire breakfast rush at a Sam's location that lives on breakfast. The financial hit was modest, maybe four figures. The lesson was expensive.
When I dug into it Monday morning, I found that we had a first successor named for that opening lead seat. The problem was that person had transferred two months prior and no one had updated the chart, because the chart lived in a shared drive nobody opened. The second successor column had never been filled in because we had never taken the two-deep discipline seriously below the general manager level.
By Tuesday I had rewritten the bench chart for the region. By the end of that quarter it was a living document reviewed monthly. Six months later we had another opening lead call out on a Sunday morning. We opened on time.
A bench you do not open is not a bench. It is a wish list stored in the wrong place.
What a two-deep chart actually looks like
Not a slide. Not a paragraph in a strategy deck. A one-page table. Every seat that can shut a shift down, one row per seat. Two columns for successors. One column for cross-training status. One column for date last reviewed.
Fig. 1 · The one-page two-deep bench chart. Red cells are the work.
The magic of a chart that looks like this is not the format. It is the shame of the empty cell. A blank second successor is a red rectangle staring at you. The chart makes the risk unignorable, and unignorable risk gets worked.
What counts as a seat
Not every position needs to be on the bench chart. If a seat cannot shut a shift down when it is empty, it does not belong there. The test is simple. Ask two questions:
- If this person does not show up tomorrow, does the shift open on time and hit its numbers? If yes, the seat is not a bench seat.
- Is this seat's knowledge held in a checklist, or in a head? If it is in a head, the seat is a bench seat, regardless of title.
In a five-unit Bay Area group, that usually gives you between twenty five and forty seats across the region. In a twenty-one-unit multi-format group like Hana, closer to ninety. Manageable in either case, once you accept the size of the actual problem.
Who goes in the first column
The first successor is not the most senior person. It is the person who can step into the seat tomorrow and run it at a passing bar. That is a functional test, not a tenure test. I have had opening leads whose first successor was a nine-month teammate, and general managers whose first successor was a peer general manager from a nearby unit rather than the assistant general manager inside the same one.
Three tests I run for the first-column name:
- Have they actually held the seat, even for a shift? Not shadowed. Held it. If the answer is no, they are the second column, not the first.
- Can the current seat-holder trust them with the closing count? Trust is expensive to fake. Ask the primary directly.
- Do they want the seat? Not eventually. Now. The person who does not want the seat will not run it well in a pinch.
Who goes in the second column
The second column is where the future of the region gets built. It is also the column most operators skip because it feels theoretical. It is not. It is the ninety-day insurance policy for the seat, and it is the development track for the person you are naming.
The person in the second column should be someone who could hold the seat with about ninety days of targeted cross-training. Not the most obvious next-in-line. Sometimes the most obvious person belongs in the second column for a different seat, one they are actually growing toward.
The second column is also where you name people from other units. Cross-unit succession is one of the strongest levers a multi-unit operator has and almost nobody uses it. The assistant general manager at one unit can be the second successor to a general manager at a different unit an hour away. It builds their range. It builds regional depth. It also builds the kind of internal talent market that reduces external hiring costs by real percentages.
Cross-training that is not fake
The most common way bench charts fail is that the second column is full of names but the cross-training column is full of "planned." Planned is not trained. Trained is done, and done means the person has run the seat, live, at least once.
Real cross-training in an operating unit has three parts:
Shadow shift
The second-position person works one full peak shift alongside the primary, watching, taking notes, asking questions between fires. No responsibility. Just presence. Usually about six hours.
Assist shift
The second-position person handles specific responsibilities of the seat, with the primary present. Maybe they run the closing count. Maybe they own the ticket rail for the second half of service. The primary is the safety net, not the operator.
Run shift
The second-position person runs the full seat for a live peak shift. The primary is on-site but stays off the pass. They do not intervene unless something is on fire. Afterwards, they debrief. This shift is the graduation. Until this shift happens, the second-column name is a plan, not a bench.
The cadence I have used across Hana and Zareen's is one full cross-training cycle per second-column person per quarter. That is a real time commitment. It is also less than one percent of the labor cost of a full unit, and it is the reason my regions have survived every general manager resignation without missing a beat.
Cross-training that never turns into a live shift is not training. It is a promise you did not keep.
The quarterly bench review
Every ninety days, I sit with each general manager for forty five minutes with their bench chart on the table. We walk it, one seat at a time. Three questions per seat:
- Is the first name still ready? What changed this quarter?
- Is the second name still developing? What did the last cross-training shift teach us?
- Which cell is the biggest risk on this chart right now, and what are we doing about it in the next thirty days?
Most quarters, three to five cells move. A name gets promoted into the primary. A new hire replaces a second who left. A cross-training cycle completes and the second becomes ready. If a chart does not move at all over ninety days, either the general manager is not developing the bench, or the chart is not honest. Both are worth a conversation.
Two ways bench charts fail
I have seen every version of this go wrong. Two failure modes come up more than the rest.
The chart that impresses the board and lies to the region
Slick slides. Every seat filled. Every succession pathway diagrammed. Nobody in the region has ever seen it. When a real seat opens up, the chart does not match reality, because it was built for a room, not a shift.
The fix is to make the chart operational. Print it. Post it in the manager office. Reference it in every general manager one-on-one. If the chart is only useful in a boardroom, it is a piece of theater, not a bench.
The chart that names people who do not know they are on it
Nothing insults a strong teammate more than finding out, in a moment of crisis, that they were the named backup for a role they were never told about and never trained for. It happens more than anyone admits. It humiliates the operator, it burns the teammate, and it destroys the credibility of every future bench conversation in that region.
Every name in the second column deserves a real conversation with their general manager. Something like this. You are on the bench chart for the sous chef seat. That means over the next ninety days we are going to give you three cross-training shifts. It also means if something happened, I would ask you to step in. Is that a seat you want to grow into? Then listen. Sometimes the answer is no, and that saves you from a bad bench pick.
What the bench actually costs
The most common objection I hear to real two-deep bench work is that it costs too much time. Let me put a number on it. Across a twenty-one-unit region with roughly ninety bench seats, running one cross-training shift per second-column person per quarter, the total labor cost was around 0.6 percent of regional payroll. That is well under half a point.
The cost of one unplanned general manager departure with no bench, at a $2M unit, ran between $18,000 and $45,000 in lost margin, external hiring, and onboarding time, depending on how badly the transition went. One prevented departure paid for eighteen months of bench discipline. Every operator I know who has done this math sticks with it.
A quick note on cross-unit bench in a multi-format region
One thing that made bench planning harder at Hana Group than at Zareen's was the format diversity. A Whole Foods embedded unit does not run like a Walmart embedded unit, which does not run like a Target embedded unit, which does not run like a Sam's Club. The peak hours are different. The guest expectation is different. The prep discipline is different. Even the vocabulary the team uses at the pass is different.
Which means cross-unit bench moves inside a multi-format region are not as clean as they look on a chart. A strong opening lead at a Whole Foods unit will need a real ramp before running an opening at a Sam's Club, even in the same city, because the volume curve is a different animal. I have made this mistake. I have moved a shift lead across formats too fast and watched them look overwhelmed on a Saturday morning that would have been routine at their home format.
The fix is to add a "format" column to the bench chart. Every second-column name gets tagged with the formats they have actually held the seat in. Cross-format cross-training becomes its own scheduled cycle, once per year per person, with the primary present. Small change. Real impact. In the last two years at Hana we cut cross-format transition friction roughly in half by naming this problem out loud.
The point
Bench strength is a boring discipline. It is not a strategy. It is a chart. It gets updated every ninety days. Every seat that can shut a shift down gets a first and a second. The second becomes ready by running the shift live, not by reading a job description. The chart lives where the general managers can see it, not where the board can see it.
The first time you save a Sunday morning because your closing lead's second successor stepped in without a phone tree, you will understand why this work is worth every hour it takes. The first time you lose a Sunday morning because the second column was empty, you will build the chart the next day. I would rather you built it before the phone rings.