On the radar
Small plates Natural & low-intervention wine Omakase counters Zero-proof cocktails Shorter tasting menus Hyper-local sourcing Dynamic menu pricing QR menus fading Chef's-counter seating Live-fire cooking Regional specialisation Snack-bar formats Small plates Natural & low-intervention wine Omakase counters Zero-proof cocktails Shorter tasting menus Hyper-local sourcing Dynamic menu pricing QR menus fading Chef's-counter seating Live-fire cooking Regional specialisation Snack-bar formats
Skip to content
All journal entries Menu engineering

Your best-selling dish might be hurting your margin

By Sagar Sharma 3 min read

Every menu has a dish the room loves — the one the regulars order, the one staff recommend without thinking, the one you’d never dare remove. It is also the dish nobody ever audits, precisely because it is loved. And if the numbers underneath it are thin, that affection is costing you money on every single cover.

Popularity hides the leak

In menu engineering, a dish that sells well but earns little is a Plowhorse — high popularity, low contribution margin. It is the most dangerous quadrant on the menu for one reason: popularity acts as camouflage. Nobody questions the dish everyone orders. The sales report looks healthy. The dish feels like a winner. Meanwhile each order leaves less on the table than it should, and because the volume is high, the leak is large and steady.

A Star and a Plowhorse can look identical from the floor — both busy, both beloved. The only thing that tells them apart is contribution margin, and that is exactly the number popularity makes you stop checking.

How to find yours

You are looking for the dish where volume and contribution disagree. The check is simple:

  1. Pull contribution margin — menu price minus plated food cost — for every item in a category.
  2. Pull popularity — each item’s share of orders in that category.
  3. Find the high-volume, low-contribution item. That’s your Plowhorse. The most-ordered dish sitting below the category’s average contribution margin is the one quietly working against you.

Then do the multiplication that makes it real: a small shortfall in contribution, times a large number of covers, is a meaningful sum every week. The danger of a Plowhorse isn’t the per-plate gap — it’s that the gap is multiplied by your highest volume.

What to do about it

The instinct is to leave it alone — it’s why people come. That instinct is right that you rarely cut a beloved dish, and wrong that you should do nothing. The moves, roughly in order of least disruption (the full set is in the matrix breakdown):

  • Nudge the price. A loyal, high-volume dish absorbs a modest increase without losing its crowd — and on that volume, a small rise is the fastest margin you’ll find.
  • Re-spec the plate. Trim cost the guest doesn’t perceive: a right-sized portion, an equal-quality but cheaper component, a garnish calibrated to what actually gets eaten.
  • Redirect some demand. Use position and menu design to put a higher-contribution neighbour in the favourite’s light, so some of that traffic shifts one dish over.

Done well, the dish stays loved and starts to pay. A worked Plowhorse can cross into Star territory without losing an order.

Why it stays hidden

The reason this dish escapes scrutiny is structural, not personal. Menus reviewed once or twice a year don’t catch it, because by the time anyone looks, the affection has hardened into “that’s just our signature dish.” Looking often — on the cadence we argue for in why menu engineering is a weekly discipline — is what keeps the question alive: not is it popular? but is it popular and paying? You can see the whole picture for a real restaurant in the sample Verdict.

The dish everyone loves deserves the most scrutiny, not the least. Popularity is a reason to look harder, not a reason to look away.

Start with the Verdict

Reading is the easy part. The Verdict is the decision.

A Verdict applies the same thinking these notes describe to your own menu and market — five deliverables in five days — free.