Same complaint, completely different cause. This operator ran the party-size check described in the cold food case study and it came back flat: temperature complaints were spread evenly across twos, fours and sixes. So the handoff was not the problem, and the three pre-shift changes that fixed the other venue would have achieved nothing here.
What sorted it was tagging the complaints by dish instead of by table.
Nine of the eleven temperature mentions in a month named the same plate.
Composite again, and the numbers are representative. The method is the part worth copying.
One plate, thirty-one per cent of the complaints
The dish was a pan-roasted sea bream with a butter sauce, plated on a cold garnish of shaved fennel. It was the second most popular main on the menu and, by the kitchen's own assessment, the best thing they made.
It also lost heat faster than anything else on the pass, for three reasons that are obvious once written down and invisible while you are cooking it:
- A thin fillet on a wide plate. Maximum surface area, minimum mass. It had perhaps ninety seconds of acceptable temperature from the moment it was plated.
- A cold component in direct contact with the hot one. The fennel was pulling heat out of the fish from the second it landed.
- A sauce that broke visibly when it cooled, which meant guests could see that the plate had been waiting even when it still tasted fine. Several of the reviews described the sauce, not the temperature.
None of this is a service failure. Runner discipline was good, the pass was clearing in under a minute, ticket times were healthy. The dish simply required a standard of logistics that a full Friday dining room cannot supply, and it had been on the menu for eleven months.
The three candidate fixes, and the one they chose
There is a temptation to solve this by trying harder — priority runners, a warmer plate, a shout across the pass. That works for a fortnight and then service gets busy again.
The real options were narrower than that:
| Option | What it costs | What it risks |
|---|---|---|
| Re-engineer the plate | A week of kitchen development | The dish stops being the dish |
| Move it to a smaller, deeper bowl | An hour and a crockery order | Looks less impressive |
| Take it off the menu | A popular seller, gone | Regulars notice |
They took the middle one, plus one recipe change: the fennel became a small side rather than a bed, so the fish sat on hot plate rather than on cold vegetable. Deeper bowl, less exposed surface, no cold contact. Development time, about two hours. Crockery cost, under £200.
Temperature mentions in the following month: one, and it named a different dish.
The general point, which is the reason to read this
Every menu has one or two plates that are quietly generating a disproportionate share of your complaints, and the reason you have not found them is that complaints get filed by symptom rather than by dish. "Cold food" goes into a temperature bucket, gets discussed as a kitchen consistency issue, and the kitchen — correctly — reports that its consistency is fine.
The diagnostic question is not are we sending food out cold. It is which plates cannot tolerate a normal Friday. Those are two different investigations and only the second one has an answer.
A plate is a candidate if it has any two of:
- A thin or small hot component on a large flat surface
- A cold element touching a hot element
- An emulsified or gelling sauce that changes appearance as it cools
- A garnish or assembly step that happens after the protein is cooked
- Anything the kitchen describes as needing to "go straight out"
That last one is the tell. If a dish needs to go straight out, it will eventually not go straight out, because that is what a Friday is.
How to run the check on your own menu
Take three months of reviews. Pull every comment mentioning temperature, texture, a broken or split sauce, or food that had "been sitting". Tag each one with the dish if it is named, and discard the ones that are not — vague complaints are not useless, but they are not what this exercise is for.
Then count by dish and divide by that dish's share of covers. A plate that is eight per cent of your mains and thirty per cent of your temperature complaints is not a service problem. It is a design problem, and it will keep producing those reviews for exactly as long as it stays on the menu in its current form.
Most operators who do this find one dish. A few find that their complaints are genuinely spread, which is the outcome that sends you back to the handoff investigation instead — and it is worth knowing which of those two worlds you live in before you spend a week fixing the wrong one.
The work is the tagging, and the tagging is the bit that stops happening in February. OMMU reads the reviews as they arrive, keeps the count by dish and by theme without anyone maintaining a spreadsheet, and says in plain English when one plate starts pulling more than its share.