Menu Engineering for Small Restaurants: What to Push or Cut
A practical guide to menu engineering for small restaurants: sort items into stars, plowhorses, puzzles and dogs, then decide what to feature, reprice or cut.

Your menu has 35 items, the kitchen is slammed every Friday, and somehow the bank account doesn't show it. Your best seller turns out to barely make money, and that specialty dish you added last spring still ties up cash in the walk-in. Menu engineering for small restaurants is the fix: a simple way to decide, with real numbers, which dishes to push, which to reprice and which to cut.
This isn't just for big chains anymore. According to the National Restaurant Association's 2026 State of the Restaurant Industry report, 82% of operators said their food costs went up in 2025, and average food costs are now more than 35% above pre-pandemic levels (citing U.S. Bureau of Labor Statistics data). The same report found 42% of operators said they weren't profitable in 2025. When every ingredient costs more, every spot on your menu has to earn its keep.
What menu engineering actually means
You look at every item and ask two questions:
- How many do you sell? (popularity)
- How many dollars does each sale leave you after food cost? (contribution margin)
Those two answers put every item into one of four boxes, and each box comes with a clear move.
| Category | Sells | Margin | What to do |
|---|---|---|---|
| Star | A lot | High | Feature it and protect it |
| Plowhorse | A lot | Low | Fix price, portion or cost |
| Puzzle | Little | High | Give it visibility or rename it |
| Dog | Little | Low | Cut it or rebuild it |
One key detail: measure margin in dollars, not percentage. A soda can run an 85% margin and still leave you fewer dollars per sale than a burger at 60%.
Step 1: Pull one month of numbers
For each item you need:
- Units sold last month. Pull it from your POS, your order tickets or your online orders. If orders come in by text or WhatsApp, count them; it's much easier when they arrive as clean, itemized orders.
- Plate cost. What one serving costs you: ingredients, plus packaging for takeout, plus sides and sauces. If you don't have this yet, start here: how to calculate food cost per dish.
Then calculate contribution margin per item: menu price minus plate cost.
Step 2: Sort your menu (worked example)
Here's an example with made-up numbers for a small taquería selling six items in a month:
| Item | Sold | Price | Plate cost | Margin per item | Total margin |
|---|---|---|---|---|---|
| Carne asada tacos (3) | 340 | $12.00 | $4.50 | $7.50 | $2,550 |
| Birria tacos (3) | 260 | $15.00 | $6.00 | $9.00 | $2,340 |
| Shrimp tacos (3) | 80 | $16.00 | $5.50 | $10.50 | $840 |
| Chicken burrito | 220 | $11.00 | $5.75 | $5.25 | $1,155 |
| Veggie bowl | 50 | $10.00 | $4.75 | $5.25 | $263 |
| Torta cubana | 50 | $13.00 | $7.00 | $6.00 | $300 |
Now set two cutoff lines:
- Popularity: total sold (1,000) divided by number of items (6) = about 167. A common rule of thumb uses 70% of that as the cutoff: about 117 units. Above that, the item is popular.
- Margin: total menu margin ($7,448) divided by units sold (1,000) = $7.45 average margin per item.
Result:
- Stars: carne asada and birria tacos (popular, margin above $7.45).
- Plowhorse: chicken burrito (popular, but only $5.25 per sale).
- Puzzle: shrimp tacos ($10.50 per sale, but few people order them).
- Dogs: veggie bowl and torta cubana.
A spreadsheet and one afternoon is all it takes.
Step 3: What to do with each category
Stars: protect them
They pay the rent. Don't mess with the recipe and don't bury them.
- Put them first in their section with your best photo.
- If you raise prices, go easy here. This guide walks through it: how to raise menu prices without losing customers.
- Use them as the anchor of your combos.
Plowhorses: make them work harder
Guests love them, but they leave little on the table. Options:
- Nudge the price. Moving the burrito from $11.00 to $12.00 lifts the margin to $6.25.
- Check the portion, or swap one pricey ingredient for an equivalent.
- Pair it with a high-margin side or drink in a combo. See how to build combo meals that raise your average ticket.
Puzzles: get them seen
They make good money, but guests skip them or don't get them.
- Rename them so people know what they're getting: "Shrimp tacos" says little; "Chipotle shrimp tacos with mango slaw and lime crema" sells.
- Add a photo. An item without a photo on a menu full of photos is invisible.
- Move it up in its section and have staff recommend it.
Dogs: decide without guilt
- If nobody's loyal to it, cut it. Fewer items means less inventory, less waste and faster tickets.
- If a group of regulars always orders it, rebuild it: new recipe, portion or price, and give it one month.
How to feature items on a digital menu
On a printed menu, the classic trick is where an item sits on the page. On a phone, the game is different: guests scroll with a thumb and decide in seconds. What matters most:
- Order within each section. Whatever shows first gets seen most. Put stars and puzzles on top.
- Photos. On a small screen, the photo outsells the description.
- Short, specific descriptions. Two lines: main ingredient, how it's cooked and one craveable detail.
- Fewer choices per section. Fifteen tacos overwhelm people. Seven or eight strong options convert better.
- Combos up front. If you have a combo, make it visible on the first screen.
The upside of a digital menu is that you can make these changes today without reprinting anything. With Menymat, for example, you just text the bot on WhatsApp something like "move shrimp tacos to the top" or "raise the chicken burrito to 12", and the menu on your link and your existing QR code updates right away.
How often to run the numbers
- Monthly: review sales and margins. About 30 minutes once your spreadsheet is set up.
- When a key ingredient jumps (beef, cooking oil, cheese, avocado): recalculate every item that uses it.
- Seasonally: decide what comes on and what comes off.
Change one thing at a time and give it a month. If you change an item's name, photo, price and position all at once, you'll never know what worked.
Quick menu engineering checklist
- [ ] I have units sold per item for last month.
- [ ] I have an up-to-date plate cost for every item.
- [ ] I calculated dollar margin for each one.
- [ ] Every item is tagged star, plowhorse, puzzle or dog.
- [ ] Stars are at the top with great photos.
- [ ] Plowhorses got a price, portion or combo fix.
- [ ] Puzzles got a new name, a photo or a better spot.
- [ ] Dogs are cut or being rebuilt.
- [ ] I set a reminder to review in 30 days.
If you want to test these changes on a digital menu you run from WhatsApp, Menymat has a 7-day free trial with no contract.
FAQ
Does menu engineering work for a small restaurant?
Yes, maybe even more than for a big one. With a short menu, cutting one dog or repricing one plowhorse shows up in your numbers fast.
How many items should my menu have?
There's no magic number, but fewer usually wins. As a practical guide, aim for seven or eight options per section and drop items that neither sell nor make money.
Why measure margin in dollars instead of percentage?
Because you pay bills in dollars. A low-priced item with a high percentage can leave you less money per sale than a pricier item with a lower percentage.
How often should I do menu engineering?
Once a month as a routine, plus any time a key ingredient spikes. Change one thing at a time so you know what worked.
Do I need special software?
No. A spreadsheet with units sold, price and plate cost per item is enough. Just make sure all the numbers come from the same period.