What it means
A restaurant may sell one dish 500 times a month at $20 contribution per serving and another 100 times at $35. The first brings $10,000 of total contribution, the second $3,500, before other costs, so a high per-serving margin alone does not identify the most important item.
Calculate each dish's selling price less recipe ingredient cost, then compare unit sales in a consistent category and period. Check tax and discount treatment before using point-of-sale data, and treat returns and complimentary meals consistently in the sales and cost records.
Compare mains with mains rather than with inexpensive drinks or sides whose natural price points differ. The classic matrix calls popular high-contribution items stars, popular lower-contribution items plowhorses, less popular high-contribution items puzzles and low-on-both items dogs, but those are prompts, not automatic commands to promote or remove.
A "dog" may be essential for a group with a dietary need, and a popular plowhorse may bring customers who buy high-margin drinks. Assess the whole basket and customer reason for visiting.
Recipe accuracy is central, because costs move with supplier prices, portions and waste. If the kitchen uses 220 grams of an ingredient while the recipe assumes 180, the displayed contribution is overstated, and a meal that needs unusually long preparation or ties up scarce equipment may need labour or capacity analysis when it changes the ranking.
Presentation can change demand, since clear descriptions, layout and staff recommendations can help customers discover a profitable but overlooked item, but changes should be tested in comparable periods rather than rearranging every dish and claiming the lift came from one label. A price increase can improve per-serving contribution but lower orders, so total category contribution and customer feedback should be tracked after any adjustment.
A test that increases one dish while cannibalising a higher-contribution dish may not help the business. Menu engineering also supports purchasing and waste control.
Removing a slow item can simplify ingredients, but a shared ingredient may still be needed for several popular dishes, so forecast ingredient quantities after menu changes, check shelf life and alternative uses before discarding stock, and use a limited seasonal special to test demand before committing to a permanent slot. For owners, the matrix is most useful as a conversation between kitchen, floor team and finance, not as an algorithm that dictates what guests should eat.
In practice
Real-world examples.
Example
A chef checks whether a popular low-contribution dish can be improved without hurting quality. A small change to portion size or a more cost-effective ingredient lifts the contribution. The chef tests the change over a comparable period.
Example
A restaurant promotes a high-contribution dish that few guests notice. It improves the menu description and asks staff to recommend it. Finance compares the dish's units and contribution before and after.
Example
Finance compares menu-item contribution with units sold and food waste. It finds that a slow dish shares an ingredient with several popular items, so removing it would not simplify purchasing. The team keeps the dish and adjusts its order quantities.
Formula
Calculation
Item contribution per serving = Selling price net of relevant sales taxes and discounts - Recipe food cost per serving
Worked example. A fictional dish sells for $80 net, costs $25 in ingredients and sells 200 servings. Contribution is $80 - $25 = $55 per serving and $55 x 200 = $11,000 across those servings before labour, rent and other costs. If a $5 discount raises sales to 220, contribution per serving becomes $50 and total contribution is $50 x 220 = $11,000 again on these simplified facts, so the discount added volume but no extra contribution. Measure actual behaviour rather than assuming a discount increases total contribution.
Worked classification using simple averages for illustration. A fictional menu has four dishes: A sells 200 units at $55 contribution, B sells 220 at $25, C sells 60 at $52 and D sells 50 at $22. The average is (200 + 220 + 60 + 50) / 4 = 132.5 units and (55 + 25 + 52 + 22) / 4 = $38.50 contribution. Dish A is above both averages, a star; B is popular but below the contribution average, a plowhorse; C is high contribution but unpopular, a puzzle; and D is below both, a dog. Each label is a prompt for investigation, not a verdict.Case study
Seen in the real world.
This illustrative and entirely fictional example follows Palmside Kitchen, an invented cafe. Its manager wanted to remove a slow-selling vegetarian dish because the matrix labelled it a dog. Staff explained that groups with one vegetarian guest often ordered several other meals when that option was available. The team measured group baskets, reduced ingredient waste and improved the dish's menu description.
It tested a revised portion before deciding whether to retain it. The invented outcome showed that item-level contribution was useful but did not capture every reason a customer chose the restaurant. The case shows why financial categories need operational and customer context.
Watch out
Common mistakes.
- Classifying items using outdated recipe costs.
- Removing a low-volume item without checking basket and customer effects.
- Treating high food-cost percentage as identical to low cash contribution.
Questions
People also ask.
What are the four classic categories?
Stars, plowhorses, puzzles and dogs, based on relative popularity and contribution.
Should every dog be removed?
No. Check strategic, dietary, basket and operational reasons first.
What data is needed?
Accurate recipe costs, net selling prices and unit sales for comparable periods.
From the founder's library

Take it further with the book.
Build your financial confidence beyond this definition. Shihan's full-length guide, Accounting Fundamentals, takes the same plain-English approach and turns it into a complete, practical playbook for non-finance managers, business owners and students - with chapter-end quiz answers and presentation slides included.
25% off with code MMHQ25, applied at checkout. Priced in USD - checkout may show the equivalent in your local currency.
View the book and save 25%