Run the numbers on a casual neighbourhood restaurant and a Michelin tasting room and you will find the same kinds of findings in both. A dish that earns well and nobody orders. A crowd-pleaser that does not carry its margin. A table that is not turning, or a course that is not landing. The diagnostics are not tier-specific. The mathematics of contribution margin does not care what is on the awning.
What is tier-specific is what you do about it — and, just as much, how you say it. That second part is the one most analysis gets wrong.
The four tiers
The Couverte methodology runs on four restaurant tiers, and the tier is the first thing you set, before any analysis runs.
Tier 1 — Local and Casual. The neighbourhood room. The economics have to work first, and the room runs on table turns. A seat that does not turn enough at lunch is revenue left on the floor.
Tier 2 — Hotel F&B. The restaurant inside a property, where breakfast, lunch, dinner, and room service are all different businesses sharing a kitchen. Profit per seat-hour by daypart is where the money hides, and the F&B experience shapes how the whole hotel is remembered.
Tier 3 — Fine Dining. The destination room. Turnover is meant to be low — you do not want to flip a table here. The figure that matters is profit per cover: how much each guest is worth across a full, unhurried experience.
Tier 4 — Michelin. The tasting menu fixes both the duration and the price, so the time-and-space metrics nearly drop out. What is left is creative coherence — whether each course earns its place in the narrative, not in the spreadsheet.
That is the budget read of the tiers, and it is the obvious half. The half that does the real work is voice.
The same finding, four ways
Take one finding and carry it across all four tiers. A Puzzle: a dish with healthy contribution margin that almost nobody orders. The margin is already there. The problem is visibility. The action, in every tier, is some form of promotion.
Now watch what “promotion” has to become.
At Tier 1, it is a direct instruction. Move the dish to the top-right panel of a two-panel menu. Stretch the description to twelve to fifteen words and lead with value — house-made, locally sourced, served with. Charm pricing is fine here. The recommendation reads like a clear note to a working operator who has five other jobs today, because that is exactly who is reading it.
At Tier 2, promotion becomes a service play. Feature the dish as the chef’s recommendation at dinner. Brief the floor team on a two-sentence spoken description. Test it as a highlighted room-service item on the in-room card. The voice frames the change as something that serves the guest and the bottom line together — never as cold optimisation, because in a hotel the F&B team thinks of itself as hospitality first.
At Tier 3, promotion gets quieter. Position the dish next to the highest-priced plate so it works as an anchor. Lift a tableside element to raise perceived value without raising cost. Descriptions run short — four to ten words, technique and ingredient doing the work, no decorative adjectives. No currency symbols, no decimals, no price column. The voice is composed. It respects the room.
At Tier 4, you do not really “promote” at all. You ask a question. If the course serves the menu’s narrative arc, treat it as a signature moment. If its low draw is a matter of sequence — the wrong beat in the tasting progression — reposition it rather than reprice it. The voice never instructs the chef on creative matters. It raises coherence and lets the kitchen decide.
Same Puzzle. Same underlying mathematics. Four genuinely different recommendations, and four genuinely different registers.
Why the voice is not decoration
It would be easy to read the above as polish — the analysis is the substance, the wording is the garnish. That is backwards.
A recommendation that is technically correct and tonally wrong does not get a discount. It gets ignored. Hand a Michelin chef a Tier 1 instruction — “reprice for volume, stretch the description, test a price cut” — and you have not given them slightly-too-blunt advice. You have told them, in one sentence, that you do not understand their restaurant. Everything else you say after that is discounted to zero. Hand a casual operator who is mid-service a paragraph of fine-dining restraint about narrative arcs and perceived value, and they will nod, file it, and never act on it, because it does not sound like their world.
The display rules carry the same weight. Charm pricing reads as honest and friendly in a neighbourhood trattoria and reads as cheap in a fine-dining room. A price column makes a casual guest shop efficiently and makes a fine-dining guest feel they are being sold to. None of that is taste. It is how guests in each kind of room have been trained, over years, to read a menu — and a recommendation that ignores it is fighting the guest instead of the problem.
This is also why a generic analysis tool struggles here. It is straightforward to compute a contribution margin. It is much harder to know that the same number should produce a blunt instruction in one room and a careful question in another. That judgment — which finding, which tier, which voice — is the part of consulting that does not reduce to a formula, and it is the part the Verdict is built to carry.
The budget of a restaurant tells you what it can spend. The tier tells you how it thinks — and how it needs to be spoken to. Get the second one wrong and the first one never matters, because the advice was filed before it was read.