Most restaurants in Mexico design their menu on intuition: what the chef likes, what customers order, what "looks good" on the card. Few apply a systematic framework to decide what to promote, what to reposition and what to cut.

Menu engineering is that framework. It's a methodology that classifies every dish on your menu along two variables: popularity (how often it gets ordered) and profitability (how much margin it leaves). The result is a four-quadrant matrix that tells you exactly where your money is.

The matrix: stars, plow horses, puzzles and dogs

Every dish on your menu falls into one of four quadrants:

Classification Popularity Margin Action
Stars High High Promote aggressively
Plow horses High Low Keep, optimize cost
Puzzles Low High Reposition, give visibility
Dogs Low Low Rework or remove

The logic is straightforward: you want more stars on your menu and fewer dogs. But the real value is in the puzzles and the plow horses, because that's where you can move margins without losing customers.

Real case: AMALAY, Monterrey

AMALAY is a brunch/café restaurant in Monterrey with more than 180,000 tickets processed over two years. Applying weekly menu engineering with real data, the classification looks like this:

Stars (high popularity + high margin)

  • Chilaquiles — the best-selling dish with a 42% margin. Consistently in the daily top 3. You don't touch it: you promote it.
  • Pancakes & Waffles — high weekend turnover, margin above average. Positioned as the "special of the day" on Saturdays.

Plow horses (high popularity + low margin)

  • Drip coffee — sells at almost every table but the margin is thin because of bean cost. Action: negotiate a better supplier price or raise it $5 MXN without denting demand.
  • Fresh juices — popular, but fresh fruit is expensive. Testing whether a 50 ml portion adjustment affects perception.

Puzzles (low popularity + high margin)

  • Acai bowl — 38% margin but low sales. Opportunity: move it to the first visual block of the menu, add a photo, have servers recommend it as "the most ordered this month" (framing).
  • Specialty croissants — high margin, but customers don't see them. Reposition on the menu.

Dogs (low popularity + low margin)

  • Truffle pasta — only 12% margin and sporadic sales. The ingredient cost doesn't justify keeping it. A candidate for removal or reformulation with more accessible ingredients.

Key data point: at AMALAY, simply repositioning 3 puzzles and cutting 2 dogs generated an estimated 12% increase in average margin per ticket over 8 weeks. Without raising a single price.

How to calculate each dish's classification

The calculation requires two data points per dish:

  1. Relative popularity — the share of total sales each dish represents. If your menu has 40 dishes and the average is 2.5%, any dish above that threshold is "popular".
  2. Contribution margin — selling price minus ingredient cost. Don't include fixed costs: only direct food cost.

With those two numbers, every dish lands somewhere on the matrix. The average margin across all dishes draws the line between "high" and "low".

Why do it weekly, not once a year

Most restaurant consultants treat menu engineering as an annual exercise. The problem is that demand shifts with the season, the weather, trends and even the day of the week. A dish that's a star in winter can be a dog in summer.

The advantage of automating it is that the classification refreshes every week with real sales data. No assumptions: just numbers.

Concrete actions per quadrant

With the stars

  • Put them in the most visible spot on the menu (first block, highlighted box)
  • Have servers mention them as "the house favorite"
  • Never lower their price. If anything, create a premium version

With the plow horses

  • Don't cut them: they bring people in
  • Look for ways to shave 1-2 percentage points off food cost without touching the recipe
  • Test whether a $5-10 MXN price increase affects volume

With the puzzles

  • Reposition them on the menu: move them to the "golden triangle" (center-top-right)
  • Add a photo or a more appetizing description
  • Feature them as the "chef's recommendation" of the week
  • Ask the service team to suggest them actively

With the dogs

  • If they can be reworked with cheaper ingredients, do it
  • If not, cut them. A shorter menu is a more efficient menu
  • Every dog you cut reduces kitchen complexity and waste

The impact in numbers

According to studies from the Cornell Hospitality Research Center, restaurants that apply systematic menu engineering report:

  • An 8-15% increase in average contribution margin
  • A 10-20% reduction in ingredient waste
  • A 5-12% increase in average ticket (without raising base prices)

The secret isn't charging more. It's selling more of what makes you more.

How Fullsite automates menu engineering

Fullsite runs this analysis automatically every Monday. The system takes the previous week's sales data, calculates popularity and margin per dish, classifies each one on the matrix, and sends the owner a report with suggested actions.

No Excel, no consultant, no spending a morning cross-referencing data. The menu engineering agent works on its own and delivers actionable conclusions.

"Menu engineering isn't academic theory. It's a practical tool that turns sales data into margin decisions. The difference is that now you can run it every week, not every year."