Menu design psychology: what the research shows

Two other price tips are popular but have less proof.

The first is to put the price at the end of the description, in the same size as the text, instead of in a column. WebstaurantStore says a column invites guests to compare prices and pick the cheapest dish. We found no test of this.

The second is price endings. In a small pizzeria with 1,271 customers, a pizza priced at 7.99 euros was chosen more often than the same pizza at 8.00 euros, but only when the other pizzas ended in zero. When every price ended in 9, or every price ended in 0, there was no difference. It is one small study, but it suggests a .99 ending works best on the one dish you want people to notice.

Do better dish names sell more?

In the tests we found, yes. In 2001, Brian Wansink and colleagues replaced plain names with descriptive ones for six dishes in a university cafeteria. Over six weeks and 140 customers, sales of those dishes were 27% higher. Diners also rated the food and the cafeteria more highly.

One caveat. Wansink later left Cornell after the university found he had committed academic misconduct in his research, and 18 of his papers were retracted. We found no retraction notice for this one, but we did not want to rely on it alone.

The second test comes from France. Nicolas Guéguen and Céline Jacob changed dish names in a small restaurant with 1,838 diners. A dish with a family tie in the name, such as "Grandma's home-made baked potatoes" instead of "baked potatoes", was ordered by 13.8% of diners, against 8.2% with the usual name. Names that hinted at tradition or patriotism also did better than the usual name.

Both studies are small, so treat the exact numbers with care. The pattern is clear enough to act on. Describe how a dish is cooked, where it comes from, or who it reminds people of. Keep it true, though. Wansink's own paper says descriptive labels should be used sparingly and appropriately.

WebstaurantStore suggests 12 to 18 words for each description. It also says words like "light" and "healthy" can make food sound less tasty, and that naming the ingredient, such as grass-fed beef, works better than making a health claim.

What about white space, colour and photos?

The evidence here is thinner, and most of it is industry advice. White space, a box or a border pulls the eye to a dish. WebstaurantStore says to highlight only one or two dishes per section. If you box everything, nothing stands out. It also suggests about one photo per page or section on a printed menu, and says red, yellow and orange stimulate appetite while blue works against it. We did not find sales tests for the colour claims in the sources we read, so treat them as design taste, not proof.

Menus on phones and delivery apps

Digital menus read differently. WebstaurantStore's guide says guests have a shorter attention span on a phone, so you need clear categories, fast loading and little scrolling. It suggests a best sellers group at the top and says delivery menus lean on photos. This is where good food photography does most of the selling, because guests cannot taste anything before they order.

How to find your best dishes

Design tips work best when you know which dishes deserve them. The standard method is the menu matrix, from Michael Kasavana and Donald Smith at Michigan State University (1982). You need 60 to 90 days of sales. For each dish, note the number sold, the price and the food cost. Price minus food cost is the profit on each plate.

Then work out the average number sold and the average profit per plate across your menu. A dish above average on both is a star. Popular with low profit is a plowhorse. Unpopular with high profit is a puzzle. Below average on both is a dog.

Puzzles are where design work pays off. A better name, a better position or a word from your staff can help them. Leave the recipe and portion of your stars alone. Plowhorses may need a small change to cost or portion size. Dogs can usually go, unless you need them for kids or dietary needs. Check the kitchen too. A high-profit dish that slows the line at peak service is not a star. WebstaurantStore suggests rebuilding the matrix every quarter, and again after any food cost change of 5% or more.

What holds up and what does not

A 2023 review by Robin Chark and Michael Ip looked at the menu design literature as a whole. It found that the reported effects look real and do not seem to come from p-hacking, which means choosing statistics until something looks significant. So menu design works. The trouble is that individual tips vary a lot in how well they have been tested.

Almost every study here is small, and many took place at one venue. Guests differ by country and by type of venue, so treat these findings as ideas to test, not guarantees.

How to test a menu change

Change one section at a time. Write down the date. Then compare sales for 30 to 90 days before and after, which is the window WebstaurantStore suggests. Look at units sold per dish and average spend per guest. Total sales can mislead you, because one busy week makes everything look better.

Menu design checklist

• Write the menu so it reads well from top to bottom.
• Put the dish you most want to sell first or last in its section.
• Cut the dishes that sell least and take the most work.
• Test prices as plain numbers, with no currency sign.
• Describe dishes by method, origin or memory, in about 12 to 18 words, and keep it true.
• Build a menu matrix from 60 to 90 days of sales before you redesign.
• Change one section at a time and compare sales.

Frequently asked questions

What is menu design psychology?