Menu Sales Mix Analysis: How to Read POS Data Before You Change a Menu

What a sales mix review answers, and what it cannot

A sales mix review answers questions you can settle with evidence: Which items do guests order most? Which items bring in the most revenue? Which items sell in volume but contribute little? Which items almost nobody orders, yet still consume prep time, shelf space and menu attention?

It cannot tell you why. A dish can sell poorly because it is priced wrong, described badly, placed at the bottom of a long section, or simply unknown to a new server. It cannot tell you what guests would have ordered if the item had been different. And it cannot tell you whether a low-volume item is a hidden margin star that earns its place through high contribution per plate.

That boundary matters because it changes how you use the numbers. Treat the mix as a starting point for a conversation with the kitchen and the floor, not as a verdict you act on alone.

Pulling the right numbers from the POS

You need four fields per item for a defined period: units sold, net revenue, the menu price at the time of sale, and — where you have it — a current cost or contribution figure. Most POS systems can export this by item for a week, a month or a quarter.

Choose the period deliberately. One weekend is too short and too dependent on weather and events. A quarter smooths out most noise but hides a dish that was added three weeks ago. For most independent restaurants, a full month that excludes major holidays is a reasonable first pass, compared with the same month last year if you have it.

Before you read anything, clean the export:

• Remove or separate items that are not menu decisions, such as gift cards, deposits, delivery fees and staff meals.
• Group modifier sales under their parent item, or a dish with five paid add-ons will look weaker than it is.
• Check that combo and bundle sales are recorded the way you think they are. If a combo posts as its own item, its components disappear from the mix.
• Note any menu change inside the period — a price change, a new photo, a section move — because it splits the data.

Reading popularity and contribution together

Sort the export by units first, then by revenue. You are looking for three shapes.

The first shape is the high-volume, high-revenue item. These are the dishes guests come for. They deserve the best placement, the clearest description and a price you can defend.

The second shape is the high-volume, low-revenue item. It may be a genuine value leader, or it may be underpriced for the labour it takes. This is where reprice and rework decisions usually live.

The third shape is the low-volume item. Some are new dishes still finding an audience. Some are seasonal. Some are simply dead weight. You cannot tell which from the mix alone, so the review flags them for a decision rather than deleting them automatically.

Worked example: a ten-item menu

The numbers below are illustrative, not benchmarks from any real restaurant.

A café sells 4,200 items in a month across ten menu items. Two items account for 1,900 of those units: a breakfast sandwich at $9 and a cold brew at $5. Together they are 45% of units but only 31% of revenue, because both sit at the low end of the price range.

At the other end, a grain bowl at $14 sells 260 units and a steak salad at $18 sells 190 units. Together they are 11% of units and 22% of revenue.

Three items sell fewer than 60 units each in the month. One is a seasonal soup, one is a new dish added three weeks ago, and one is a sandwich that has been on the menu for two years with no changes.

The decisions that follow are modest and specific: keep the two volume leaders where they are and protect their placement; check the grain bowl's portion cost before touching its price; leave the new dish alone for another month; ask the kitchen whether the two-year-old sandwich is worth its prep time; and reprice or rework the breakfast sandwich only if its contribution is genuinely below the menu's average.

Turning the review into decisions

The review is only useful if it ends in a written list with owners and dates.

Record the reason for each decision. Next month, when the same item appears again, you will want to know whether you already tested this idea.

• **Keep:** items with healthy volume and contribution. Protect their placement and do not change them casually.
• **Reprice:** items where the cost or labour has moved and the price has not.
• **Rework:** items with good demand but poor contribution — smaller portion, simpler garnish, cheaper accompaniment, or a higher-margin side.
• **Remove:** items with low volume, low contribution and no strategic role. Check the guest conversation first.
• **Review again:** new, seasonal or data-poor items. Give them a date and a threshold, then decide.

Using the review without overreacting

Two habits keep a sales mix review from becoming a monthly purge.

First, change one thing at a time where you can. If you reprice, rework and move an item in the same month, you will not know which change did the work.

Second, watch the mix after the change. A reprice that lifts contribution but halves volume may be a good trade, a bad one, or neutral depending on the kitchen's capacity. You cannot know until the next period closes.

When the decisions are ready to publish, the MenuCrafters AI menu builder keeps the menu you edit as the source for both the hosted QR page and the print-ready PDF, so a repriced item changes in one place instead of three.

Limitations and assumptions

This method assumes your POS item names match your menu items, your modifiers are attached to parents, and your export excludes non-food lines. It assumes the period is long enough to be representative and that no major menu change happened inside it. It cannot measure why guests choose one dish over another, whether a low-volume item is strategically important, or how demand would respond to a price change. Contribution figures depend on a current, verified recipe cost; if your costing is out of date, the mix will point you at the wrong dishes. Small menus, new restaurants and seasonal businesses should expect more noise and lean more on the "review again" decision than on removal.