Where service breaks at peak

Every kitchen has a volume at which it stops being able to hold its standard. Below it, the system works. Above it, ticket times stretch, plates leave slightly wrong, and the floor starts apologising. Most operators know that this point exists. Far fewer know where it is.

## Finding the constraint

A service does not have many bottlenecks. It usually has one, and it moves depending on the mix.

The way to find it is to watch, not to ask. During a peak service, record three things: when each ticket is fired, when each course leaves, and where tickets are waiting when they wait.

A queue in front of one station at peak is the constraint. It is often not the station people assume. A grill that everyone believes is the pressure point can be waiting on a garnish from cold prep; a fryer can be holding the whole pass because of one item everybody orders.

## The mix is the hidden variable

The constraint moves with what people order. A menu with three items that all finish on the same station has a different breaking point at a table of six ordering all three than at a table ordering across the menu.

This is why menu engineering and operational capacity are the same conversation. An item promoted heavily in marketing, sold well at peak and finished on the busiest station will find the constraint faster than any volume increase.

## Staffing against the demand curve, not the average

Most rotas are built on shift blocks and headcount targets. Demand does not arrive in shift blocks. It arrives in a curve with a sharp peak, and a rota that covers the average is short at the peak and long either side of it.

Plot arrivals or ticket volume in fifteen-minute intervals for a representative week. The shape is usually sharper than anyone expects. Then compare it to the rota. Most venues find they are paying for coverage in the hours before and after the peak and short-staffing the twenty minutes that decide the evening.

## The recovery point

There is a second thing worth timing: how long the operation takes to recover after the peak passes. A kitchen that returns to normal ticket times within fifteen minutes of the peak ending is running with genuine capacity. One that stays behind for an hour was over its limit long before the visible failure.

The recovery time is often a better early indicator than the peak time itself, because it shows whether the system had any slack at all.

## What to do with the finding

Once the constraint is located, the options are narrow and specific, which is the point.

**Move work off the station.** Preparation that can be done earlier, a component that can be finished elsewhere, an item that can be assembled at a different point.

**Change the menu at that hour.** Some venues run a reduced peak menu for exactly this reason, and it is a legitimate operational decision rather than a compromise.

**Add capacity at the constraint only.** An extra person somewhere else does not help and can make the pass more crowded.

**Change the sequence.** Firing rules, hold times and how the pass is expedited often produce more than an additional pair of hands.

## Practical recommendations

1. Time tickets by station for at least three services, including the hardest one you have.
2. Plot demand in fifteen-minute intervals and overlay the rota on it.
3. Identify where tickets wait, not where people look busy — these are frequently different.
4. Measure recovery time after peak as a capacity indicator.
5. Check whether the item you are promoting finishes on the constrained station before the campaign runs, not after.
6. Re-check after any significant menu change. The constraint moves.

## In short

Service failure at peak is a specific, locatable event, not a general shortage of effort. Finding the station, the hour and the mix that produce it turns an expensive guess about staffing into a narrow and usually cheap intervention.