Menu engineering is a way to look at your menu with two numbers instead of one. Most owners know which items sell most. Fewer know which items earn most per plate. Menu engineering puts both numbers on one chart and sorts every item into one of four groups. Each group has one clear action.
The trade names for the groups are stars, plowhorses, puzzles and dogs. KANJIN's Profit matrix uses plain names: Top earners, Popular, low profit, Profitable, rarely sold, and Low on both. This post uses both, so you can read either version.
The two numbers
For each item you need:
- Profit each: the price minus what the item costs you to make. Not a percentage. Cash.
- Weekly sales: how many you sold last week, or an average over four weeks.
Then find the middle value (the median) of each list. Items above the profit median are "high profit". Items above the sales median are "popular". That gives four groups.
| High profit each | Low profit each | |
|---|---|---|
| Sells often | Top earners (stars) | Popular, low profit (plowhorses) |
| Sells rarely | Profitable, rarely sold (puzzles) | Low on both (dogs) |
Why split at the median and not the average
One best seller can pull an average so far up that every other item looks weak. Say five items sell 40 to 60 a week and one sells 400. The average is over 100, and five items land in the "sells rarely" half. The median is about 50, and the items split evenly.
The median always puts half your items above the line and half below. That is the point. The chart compares your items with each other, not with a rule from a textbook.
A sample six-item menu
Here is a cafe menu with six items. Profit each is price minus cost. Weekly profit is profit each times weekly sales.
| Item | Price | Cost | Profit each | Weekly sales | Weekly profit | Group |
|---|---|---|---|---|---|---|
| Flat white | $4.50 | $1.10 | $3.40 | 320 | $1,088.00 | Popular, low profit (plowhorse) |
| Avocado toast | $12.00 | $4.20 | $7.80 | 150 | $1,170.00 | Top earner (star) |
| Chicken sandwich | $11.00 | $5.50 | $5.50 | 90 | $495.00 | Profitable, rarely sold (puzzle) |
| Salmon bowl | $16.00 | $6.00 | $10.00 | 40 | $400.00 | Profitable, rarely sold (puzzle) |
| Soup of the day | $7.00 | $3.50 | $3.50 | 55 | $192.50 | Low on both (dog) |
| Blueberry muffin | $3.50 | $1.60 | $1.90 | 120 | $228.00 | Popular, low profit (plowhorse) |
Total weekly profit: $3,573.50.
How the groups were worked out
Profit each, sorted: $1.90, $3.40, $3.50, $5.50, $7.80, $10.00. With six items the median is the average of the third and fourth values: ($3.50 + $5.50) ÷ 2 = $4.50. The sandwich, the toast and the salmon bowl are above $4.50. They are high profit. The muffin, the flat white and the soup are below.
Weekly sales, sorted: 40, 55, 90, 120, 150, 320. The median is (90 + 120) ÷ 2 = 105. The muffin, the toast and the flat white sell more than 105. They are popular. The salmon bowl, the soup and the sandwich sell less.
Now combine the two:
- Avocado toast is high profit and popular. Top earner.
- Flat white and blueberry muffin are popular but low profit. Popular, low profit.
- Chicken sandwich and salmon bowl are high profit but sell rarely. Profitable, rarely sold.
- Soup of the day is low on both.
If an item lands exactly on a median, KANJIN counts it in the higher group.
What to do with each group
Top earners (stars): keep and show
The avocado toast earns $7.80 a plate and sells 150 a week. Do not touch the recipe. Put it at the top of the menu. Ask staff to recommend it. Never put it on discount. A 20% discount on this item alone would cost $360 a week in profit.
Popular, low profit (plowhorses): small price rise or small cost cut
The flat white earns only $3.40 but sells 320 a week. A 25-cent price rise is $80 a week, or about $4,160 a year, with no change to the drink. Or find a cheaper milk supplier and keep the price. Do one, not both, and not by much. People buy these because of the price. A big rise can push them into the Low on both group.
The muffin is the same story at a smaller scale. At $1.90 profit and 120 a week, a 30-cent rise adds $36 a week.
Profitable, rarely sold (puzzles): make them visible
The salmon bowl earns $10 a plate, the most on the menu, but only 40 people order it. The chicken sandwich earns $5.50 and sells 90, just under the median. These items need to be seen. Move the bowl to the top right of the menu board. Give it a better name. Offer the sandwich with the soup as a set. If the bowl went from 40 to 80 a week, that is $400 more profit with no new item.
Low on both (dogs): cut, unless it has another job
The soup earns $3.50 and sells 55. It is last on weekly profit at $192.50. Cutting it frees a burner and a prep task. But ask first: does it do another job? Is it the only vegan option? Is it what one regular orders every day? If yes, keep it and stop worrying about it. If no, replace it with something that could be a Top earner.
Mistakes to avoid
Using profit percentage instead of cash. A 75% margin on a $4 item beats a 60% margin on a $30 item on a percentage chart. But the $30 item puts $18 in the register and the $4 item puts $3. Cash pays the rent. Use cash per item.
Judging on one week. A rainy week or a holiday changes the order. Use four weeks if you can.
Changing everything at once. Move one price, or one item, and watch for two weeks. Then do the next one.
Forgetting the items that are not on the chart. Kids meals, a gluten-free option, a plain black coffee. They can be Low on both and still matter, because they bring in the table that orders three Top earners.
Frequently asked questions
What is a menu engineering matrix? A chart with two axes. One is profit per item, the other is how often the item sells. Each axis is split at the median of your own items. That makes four groups: stars, plowhorses, puzzles and dogs. Each group has one action: keep, reprice, promote or cut.
How many items do I need? At least six. Below that the medians are arbitrary and the chart can mislead. A menu with 40 items works fine, but start with the 12 that sell most.
Should I use food cost percentage? Use it to spot a recipe that costs too much. Do not use it to rank the menu. Cash per item is what fills the register, so rank on cash.
Does this work for a salon or a shop? Yes. Use services or products instead of dishes, product cost instead of food cost, and weekly bookings or units instead of plates. The four groups and the actions are the same.
Try it with your own menu
Type your items into the free Profit matrix. It works out profit each, finds both medians, and shows which group every item lands in, with the action for each. If you want the same numbers kept on every tablet in the shop, along with the Slow Day email and the review requests, the Paid plan is $39 a month per location, cancel any time.