Menu decisions

Menu engineering for a small restaurant

By RevenArc
November 5, 2026•4 min read

Menu engineering for a small restaurant is a way to review dishes using sales volume and cash contribution. It takes sales history from your POS and plate costs from your recipes, then groups each dish to help you decide what to change next.

The common four-way split

The common method divides a menu into four groups based on two questions: does the dish sell often, and how many dollars does it leave after food costs?

In 1982, hospitality researchers Michael Kasavana and Donald Smith introduced this matrix at Michigan State University, labelling the four groups Stars, Plowhorses, Puzzles, and Dogs. In daily restaurant operations, plain words describe what each group actually means:

  • Popular and profitable: dishes that sell frequently and leave higher-than-average dollars per plate.
  • Popular and not profitable: dishes that sell frequently but leave lower-than-average dollars per plate.
  • Profitable and not popular: dishes that leave strong dollars per plate but sell in lower volume.
  • Neither popular nor profitable: dishes that sell in low volume and leave lower-than-average dollars per plate.

To assign dishes to these groups, you calculate two cut-offs from your sales period: the popularity cut-off and the profitability cut-off.

Example: A five-dish entree category

Example: worked example with made-up numbers showing menu engineering cut-offs and grouping arithmetic.
Dish Price Plate cost Weekly orders Dollars per plate Landed group
Classic Burger $15.00 $5.00 110 $10.00 Popular and not profitable
Grilled Salmon $26.00 $11.00 110 $15.00 Popular and profitable
Ribeye Steak $35.00 $16.00 50 $19.00 Profitable and not popular
Veggie Wrap $12.00 $4.00 70 $8.00 Neither popular nor profitable
Fried Chicken Sandwich $18.00 $4.00 160 $14.00 Popular and profitable
Total / Average 500 $13.00

How the cut-offs were set

Here is the arithmetic used to set both cut-offs for this category:

1. Popularity cut-off:

  • Total weekly orders across all five dishes: 110 + 110 + 50 + 70 + 160 = 500 orders.
  • Average orders per dish: 500 divided by 5 dishes = 100 orders.
  • Dishes with more than 100 orders are high volume (popular). Dishes with 100 or fewer orders are low volume.

The original method by Kasavana and Smith puts the popularity line at 70% of the average, not the full average. Here that is 70 orders (100 times 0.7). The groups don't change in this example: the Veggie Wrap sits exactly on the line at 70, so it stays low volume, and every other dish is clearly above or below it. Use whichever line you prefer, but use the same one every time.

2. Profitability cut-off:

  • Dollars per plate (contribution margin) for each dish: selling price minus plate cost. For portion costing steps, see our guide on how to calculate plate cost.
  • Total dollar contribution generated by each dish:
  • Classic Burger: 110 orders times $10.00 = $1,100.00
  • Grilled Salmon: 110 orders times $15.00 = $1,650.00
  • Ribeye Steak: 50 orders times $19.00 = $950.00
  • Veggie Wrap: 70 orders times $8.00 = $560.00
  • Fried Chicken Sandwich: 160 orders times $14.00 = $2,240.00
  • Total dollar contribution for the category: $1,100.00 + $1,650.00 + $950.00 + $560.00 + $2,240.00 = $6,500.00.
  • Average dollars per plate: $6,500.00 total contribution divided by 500 total orders = $13.00.
  • Dishes leaving more than $13.00 per plate are profitable. Dishes leaving $13.00 or less are not profitable.

Evaluating items by dollars per plate avoids the distortion of percentage alone. As explained in our guide on restaurant food cost percentage, a low percentage doesn't mean a dish leaves you many dollars per plate.

What menu engineering cannot tell you on its own

A menu matrix gives you an organized snapshot of your sales history. It does not explain the full business context:

  • Season and weather: A hearty beef stew may land in the low-volume group in July and lead the category in January. The matrix reflects the weeks you measured, not year-round performance.
  • Why a dish sells: Sales figures cannot tell you why an item sold. A dish might sell because of server recommendations, prime menu placement, or because it satisfies a dietary requirement.
  • What happens if you change it: If you raise the price of a popular dish with low profit, sales volume might stay steady, or orders might drop and shift to another item. When considering price changes, review the arithmetic in how much to raise menu prices before adjusting the menu.

The four groups help you choose which dishes to inspect first. They do not replace operational judgment.

How to use menu engineering in your weekly review

Do not try to reprice or replace every dish at once. Pick one or two items from your review that show a clear reason for attention:

  • For high-volume dishes with low profit, check whether portion sizes drifted, negotiate vendor prices on key ingredients, or test a modest price adjustment.
  • For high-profit dishes with low volume, review item placement and description on the menu.
  • For dishes that are neither popular nor profitable, evaluate whether the dish serves a necessary dietary role before deciding to retire it.

If you use RevenArc, it connects to Square or takes your Toast Item Selection Details export and, once you enter plate costs, shows each dish's sales and dollars per plate in your weekly review, so you can see which group it falls into.

See what a menu review looks like.

The sample shows how a review can bring an item's sales, missing information and next question together in one place. Its data is illustrative.

Questions operators ask

What are the four categories in menu engineering?+

The common method divides dishes into four groups: popular and profitable, popular and not profitable, profitable and not popular, and neither popular nor profitable.

How do you set the popularity cut-off in menu engineering?+

You divide total category orders by the number of dishes to find average orders per dish, or multiply that average by 70% following the Kasavana and Smith benchmark.

What is the biggest limitation of menu engineering?+

It reflects historical sales but cannot tell you why an item sold, account for seasonality, or predict how orders will shift when you make a price or recipe change.

RA

RevenArc

Menu decisions and restaurant intelligence

We build RevenArc for restaurant operators who want a clearer view of their menu and a useful place to start the next review.

Feed Reader

Follow the next article.

Email updates aren't available yet. You can follow RevenArc in your feed reader to see new articles as they're published.

Open the RSS feed

Keep reading

Menu decisions

How to price add-ons and sides on a restaurant menu

How to price add-ons and sides on a restaurant menu comes down to margin dollars and attach rate rather than target food cost percentage. How portion costs, modifier pricing, and guest uptake shape your weekly menu contribution.

Menu decisions

What to do when beef goes up 12 percent

What to do when beef goes up 12 percent is a question most operators face when protein markets tighten. How wholesale protein hikes hit different plates, and your three operational options as an owner.

Share this article