Menu Engineering: Which Dishes Actually Make You Money?

Your best-selling dish might be your worst margin performer#

Every restaurateur has a dish they are proud of. The one customers order constantly, the one that gets mentioned in reviews, the one the chef considers their signature. And sometimes, that dish is a margin disaster. A classic example: a £14 fish and chips that costs £5.80 to produce (41.4% food cost) sells 180 portions a week. It contributes £1,476 gross profit. Meanwhile, a £16.50 chicken dish that costs £3.90 (23.6% food cost) sells 60 portions — contributing £759 gross profit but with a far superior margin. You think fish and chips is your best dish because it sells the most. It is — in volume. But the chicken dish makes more money per portion. Menu engineering surfaces these insights and lets you act on them.

The four quadrants of menu engineering#

The classic menu engineering framework — developed by Kasavana and Smith in the 1980s — plots every dish on two axes: gross profit contribution (how much money it makes per portion) and popularity (how often it is ordered). Stars: high profit, high popularity. These are your best dishes. Protect them, keep them on the menu, and make them prominent. Plowhorses: low profit margin but high popularity. Guests love them but they are not making you much money — reduce their cost or raise price slightly. Puzzles: high profit margin but low popularity. Potentially great dishes that need better menu placement, description, or promotion. Dogs: low profit and low popularity. Candidates for removal. AskBiz generates this quadrant automatically from your POS sales data and your recipe costs.

Why menu mix matters as much as individual dish margins#

Menu mix is the proportion of your total sales that comes from each dish. If 40% of your covers order your lowest-margin dish (the Plowhorse), your overall food cost is dragged up even if your Stars are perfectly priced. A 5% shift in menu mix from low-margin to high-margin dishes can improve your overall food cost percentage by 1.5-2 percentage points without changing a single ingredient or raising prices. That shift happens through menu psychology: placement, naming, photography, and description. Stars go top-right or top of category — the natural eye landing zone. Plowhorses go at the bottom. Dogs disappear. AskBiz tracks menu mix weekly so you see the impact of any layout or description changes on actual ordering behaviour.

How to calculate gross profit per dish#

Gross profit per dish = Selling price − Food cost (the direct ingredient cost at recipe quantities). This is different from food cost percentage, which is food cost ÷ selling price. A £20 dish with a £5 food cost has a 25% food cost percentage and £15 gross profit. A £10 dish with a £2 food cost has a 20% food cost percentage and £8 gross profit. The second dish has a better food cost percentage, but the first contributes almost double the gross profit per portion. Menu engineering uses gross profit per dish — not food cost percentage — as the primary performance metric, because you bank gross profit, not percentages. AskBiz calculates this automatically from your recipe library and updates it whenever ingredient costs change.

Redesigning your menu based on data#

Armed with the quadrant analysis, you make three types of changes. First, reposition: move Stars to premium menu real estate. In printed menus, that is top-right of each section and top-left in two-column layouts. On digital menus, it is the first item in each category and any featured or "chef recommends" tag. Second, reprice: Plowhorses that guests order regardless of price have inelastic demand — a £1-1.50 price increase will not significantly reduce orders but will improve gross profit per portion. Third, remove or rework: Dogs that are neither popular nor profitable consume kitchen prep time, ingredient purchasing, and menu space. Remove them or radically simplify them.

Seasonal menu changes and profitability tracking#

Every time you change the menu — seasonal updates, specials, new concepts — the profitability picture changes. A new special added in October might sell well but have an unforeseen 38% food cost because the seasonal ingredient is expensive. AskBiz tracks new dishes from day one. After 14 days of sales data, new dishes are automatically plotted in the quadrant analysis alongside your existing menu. You know within two weeks whether a new dish is a Star or a Dog — not at the end of the quarter when the accountant notices the food cost spike.

Real impact: a UK Indian restaurant#

An Indian restaurant in Manchester with 45 covers ran menu engineering analysis through AskBiz. Findings: three of their eight main courses were Dogs (ordered by fewer than 4% of covers, with food costs above 35%). Their two highest-margin dishes were buried in the middle of the menu. Their most popular starter was a Plowhorse at 38% food cost. Actions taken: removed two Dogs, moved the high-margin mains to top positions, raised the Plowhorse starter price by £1.50, rewrote descriptions for two Puzzle dishes. Eight weeks later: food cost dropped from 34.1% to 30.6%, average spend per cover increased from £23.40 to £25.10. Monthly profit increase: £3,800 on unchanged revenue.

• Menu engineering is the practice of analysing every dish by profitability and popularity — then designing your menu to maximise both.
• Most restaurant owners think they know which dishes perform best.
• The data usually disagrees.