Menu Engineering for a Small Kitchen

Most menus are written by the chef and priced by feel. That works until margins tighten, and then a few dishes quietly carry the whole kitchen while others cost you money on every ticket. Menu engineering is the discipline of finding out which is which. It sorts every dish by two numbers you already generate every day, and it tells you plainly what to do with each one. You do not need a consultant or a spreadsheet degree. You need a couple of hours and a willingness to act on what you find.

What menu engineering actually is

Menu engineering is a simple sorting exercise. You take each item on a menu section, score it on two axes, and drop it into one of four buckets. The two axes are:

• How well it sells (popularity, measured in units sold over a period).
• How much money it makes you (contribution margin, measured in dollars per plate).

That is the whole idea. The power is not in the math; it is in refusing to treat all dishes the same. A best-seller with thin margins needs a different fix than a high-margin dish nobody orders. Menu engineering forces you to see those differences instead of averaging them away.

Do it one menu section at a time — appetizers against appetizers, entrees against entrees. Comparing a $6 side to a $32 steak tells you nothing useful.

Getting the two inputs

Contribution margin per dish. This is the menu price minus the plate cost (the cost of the ingredients that go into that one dish). If a sandwich sells for $14 and the ingredients cost $4.50, its contribution margin is $9.50. That $9.50 is what is left to cover labor, rent, and everything else — and eventually profit. You want this number in dollars, per plate, for every item.

To get plate cost, cost out the recipe: weigh or portion each ingredient, multiply by what you pay for it, and add it up. This is tedious the first time and fast every time after. Include the things people forget — oil, garnish, the bread that comes with it, the ramekin of sauce.

Units sold. This is how many of each item you sold over a defined window. A point-of-sale report gives it to you instantly. If you count by hand from tickets, that is fine too — just pick a clean window and count consistently.

Choose a window long enough to be representative. Four weeks is a reasonable default for most restaurants; it smooths out a slow Tuesday and a big Saturday. Avoid a holiday week or your one freak weekend, which will distort everything.

The four quadrants and what to do with each

Once you have margin and units for every item, find the midpoint of each axis. A common approach: an item is "high popularity" if it sells above the average units for its section, and "high margin" if its contribution margin is above the section average. That splits your menu into four groups.

Stars — high popularity, high margin. These sell well and make good money. They are your best assets. Do not touch the recipe, do not discount them, and do not bury them at the bottom of the menu. Give them the best real estate — top of the section, a box, a photo if you use photos. Protect their quality relentlessly, because a slip here costs the most.

Plowhorses — high popularity, low margin. Customers love them, but each one makes little money. These are dangerous precisely because they sell so much; a small margin problem, multiplied by high volume, is a big problem. Work the cost and the price, gently. Try a modest price increase (often a dollar or two goes unnoticed on a beloved item). Trim the plate cost with smaller portion tweaks or a cheaper garnish that nobody misses. Consider pairing it with a high-margin side or drink so the whole check improves even if the dish stays lean.

Puzzles — low popularity, high margin. These make great money but few people order them. The margin is already there; the problem is visibility or appeal. Move them to a stronger spot on the menu. Rewrite the description so it sounds worth ordering. Train servers to recommend them by name. Rename a dish that sounds unfamiliar. Sometimes the price itself is the barrier — a puzzle priced far above its neighbors may need a small trim to get tried. Give it a real chance before you give up on it.

Dogs — low popularity, low margin. Few orders, little money each. Your first instinct — cut it — is usually right, but check the role it plays first. A dog might be the only vegetarian option, a signature the regulars expect, or a cheap-to-hold item that anchors the menu. If it has no such role, remove it. Every dog you carry adds prep, inventory, and waste, and it distracts guests from items that pay you better. Cutting a dog is one of the fastest, cleanest wins in the whole exercise.

The mistakes that undo the work

Chasing food-cost percentage instead of margin dollars. This is the big one. A dish at 25 percent food cost looks better than one at 35 percent — but if the first makes you $4 a plate and the second makes you $11, the "worse" percentage is the better dish. You bank dollars, not percentages. Percentage is a useful sanity check, never the decision. Beware of cutting a high-percentage plowhorse that quietly earns more real money than your proudest low-percentage item.

Re-engineering too often. This is a seasonal exercise, not a weekly one. Recipes need time to settle, guests need time to respond to a new layout, and a four-week sales window can only tell you so much. Reworking the menu every few weeks means you are reacting to noise and never letting a change prove itself. Once a quarter, or when your costs or menu meaningfully change, is plenty.

If your data is thin

A small kitchen without detailed reporting can still do this well. Popularity is easy to eyeball — you already know your five best sellers and your dead weight. Write down a rough rank from memory, then sanity-check it against a week of tickets.

Margin takes more care but not more magic. Cost out your ten or fifteen highest-volume items first; those drive most of your revenue, and the long tail can wait. Round numbers are fine at this stage. The goal is not a perfect model — it is knowing which quadrant each dish sits in, and that survives rough estimates easily. A dish is a dog or it is not; you rarely need decimals to tell.

If you do run a point-of-sale system, the units-sold and item-level sales data it already captures turn the popularity half of this into a two-minute report — worth pulling before you start.

Start here this week

Pick one menu section — entrees is a good choice. List every item. Next to each, write two numbers: how many you sold in the last four weeks, and the contribution margin in dollars (price minus plate cost). Draw the average line on each column, and sort every dish into star, plowhorse, puzzle, or dog.

Then act on just three items: protect one star, fix one plowhorse's margin, and cut or promote one dog or puzzle. Three changes on one section is enough to move real money, and it is small enough to actually finish. Do the next section next month.