Menu engineering for Indian restaurants: the four-quadrant model, with examples

Most Indian restaurant menus are designed once, by the chef, in the first month of operation, and then never reviewed again. Five years later, the menu has 90 items, half of which lose money, and nobody can remember why the Veg Manchurian is on page 2 instead of the Paneer Lababdar.

Menu engineering is the discipline of treating your menu as a portfolio. Every dish gets two scores: how often it sells (popularity) and how much profit it makes per sale (margin). Plot them on a 2×2 grid and you get four categories. Each category has a different action.

The framework

• STARS — high popularity, high margin. Promote, protect, never discount.
• PLOWHORSES — high popularity, low margin. Re-engineer to lift margin.
• PUZZLES — low popularity, high margin. Reposition or repackage.
• DOGS — low popularity, low margin. Retire without sentiment.

Pulling the data

You need 90 days of POS data and your standardized recipe costs. Indostra exports both as CSV. If you are on paper, block out a Sunday and tally manually — it is worth it.

Stars

On the menu we analyzed (an 80-item North Indian thali venue in Pune), the clear stars were Paneer Tikka, Dal Makhani, and Garlic Naan. All three sold above category average and carried 68%+ margins.

The mistake most operators make with stars is discounting them in combos. Never combo a star with another star — combo a star with a plowhorse to drag the margin up.

Plowhorses

Butter Chicken was the obvious plowhorse — top-3 in popularity, but the margin was eroding because chicken thigh prices had climbed 28% over 18 months and the menu price had not moved.

Three options: raise the price (risky if your guests anchor on the old number), reduce portion size (acceptable if you redesign the plate to look generous), or reformulate (move from boneless thigh to a 60/40 mix with leg). We did option three. Margin recovered to 61%, guests noticed nothing.

Puzzles

Puzzles are the most interesting category because the diagnosis is usually a menu design problem, not a food problem. The Lamb Rogan Josh on this menu was high-margin and sold to almost nobody — because it was on the last page, with no photo, in 9pt type.

1. 1Move puzzles to the top-right corner of the page. Eyes land there first.
2. 2Add a photo or a 'Chef's Pick' badge.
3. 3Train waiters to mention puzzles when guests are deciding.
4. 4Re-measure after 30 days. If popularity does not double, the puzzle is actually a dog.

Dogs

The hardest part of menu engineering is killing dishes the chef loves. The Hyderabadi Biryani on this menu sold 4 units a week and lost ₹38 per unit because it had a 3-hour prep time and unique ingredients used nowhere else. It was a dog.

Retire it. Free up the kitchen labor, the cold storage shelf, and the dwindling slot of guest attention. Replace it with nothing — a smaller menu sells more per item, every time.

Your next 30 days

1. 1Week 1: Pull 90 days of POS sales by item.
2. 2Week 1: Cost every recipe to ±5% accuracy.
3. 3Week 2: Build the 2×2. Print it. Stick it in the office.
4. 4Week 3: Identify 3 stars to promote, 3 plowhorses to re-engineer, 5 dogs to kill.
5. 5Week 4: Reprint the menu. Train staff. Re-measure in 60 days.