Menu psychology: what holds up in research, and what is myth

When you reprint a menu or set up a QR menu, you will find a lot of advice. Put your best dish in the top right corner. Remove the “kr.” from your prices. Give every dish a long, tasty name. Much of this advice is repeated from blog to blog, and very little of it comes with a study behind it.

This guide goes through the menu advice that researchers have actually tested. For each finding, it says how the study was done and how much weight you can put on it.

The “sweet spot” that eye tracking did not find

A popular idea says that guests look first at the upper right part of a menu, and that the dish you place there sells more. You will see it called the “sweet spot” or the “golden triangle”. We could not find a published study that shows it.

Sybil Yang tested the idea in 2012. Participants wore eye-tracking equipment while they read a two-page menu and chose a full meal. They read the menu in order, like a book, from left to right and down the pages.^{2 3} The study found no area where the eyes stayed longest. It did find a “sour spot” that got the least attention. On that menu, it held the information about the restaurant and a list of salads.³

This was one menu in a controlled setting, and the full paper is behind a paywall, so we rely on the press summary for the details. Still, it is the best test of the sweet spot that we know of, and it did not find one.

First and last in a list

The finding with the best support is simple. A dish sells more when it is first or last in its section than when it is in the middle.

Eran Dayan and Maya Bar-Hillel tested this in two ways.¹ In the first study, 240 students chose a meal from a menu with the dishes in different orders. In the second study, a small café in Tel Aviv used two versions of its real menu on alternating days for 30 days. The only difference was the order of 20 of its 60 items. The café recorded 951 orders from those sections.

In both studies, an item got more orders when it was at the top or the bottom of its section. For a few single items the difference was close to two to one. Across all items it was smaller. The authors summarise it as a gain from 45 % to 55 % of an item’s orders, or about 20 %.¹

Share of an item's orders that came when it was first or last in its section (2011)

Show the data

Study	Share of orders at the start or end
If position made no difference	50 %
Study 1: lab, 240 students	56 %
Study 2: café, 951 orders	55 %

The study has two limits. The café waiters knew about the test and wrote down the orders, so the data collection was not blind. The studies also used printed menus only.¹ Nobody has shown that the same effect holds on a kiosk screen or in a QR menu where guests scroll, although it is a reasonable guess.

In practice, put the dishes you most want to sell at the start or the end of their section. Put the dishes you keep for completeness in the middle.

Currency signs and price format

You may have read that removing the currency sign raises spending by about 8 %. That number comes from a single experiment.⁴

Sybil Yang, Sheryl Kimes and Mauro Sessarego tested three price formats at the lunch service of St. Andrews Cafe, a restaurant run by students at the Culinary Institute of America in New York State. The formats were “$20.00”, “20.” and “twenty dollars”. Each table got one format at random, and the study used the checks of 201 tables.

Overall, the price format did not have a significant effect on the check. Party size, the length of the meal and the table had much larger effects.⁴ Only a more detailed comparison found that tables with the plain number “20.” spent an estimated 5,55 dollars more than tables with the other two formats. That was about 8 % more per person. The authors themselves say that the result appeared only after they controlled for many other factors.⁴

In short: one restaurant, about 200 tables, a small effect, and dollars, not krónur. If you like how a clean number looks on your menu, use it. Do not expect it to change your month.

Descriptive names: a popular finding with a problem

Many menu guides say that a descriptive name, such as “Grandma’s slow-roasted lamb” in place of “Roast lamb”, sells more. The figure behind this advice usually comes from a 2001 cafeteria study by Brian Wansink and colleagues.⁵

Treat that study as unreliable. In 2018 Cornell University found that Wansink had committed academic misconduct in his research, including misreporting of research data and problematic statistical techniques. He resigned and left the university at the end of that academic year.⁶ A later Wansink paper on attractive names for vegetables in school canteens has been retracted.⁷ As far as we could check in the retraction records, the 2001 paper has not been retracted. It comes from the same lab, so we do not repeat its numbers here.

Better evidence exists, and it points in a similar direction. A preregistered study in five university dining halls in the United States compared labels for vegetable dishes. It covered 137.842 choices over 185 days.⁸ Labels that described the taste increased vegetable selection by 14 % compared with plain labels. They increased it by 29 % compared with labels that stressed health. The taste labels also did better than labels that only used positive words, fancy words or a list of ingredients.⁸

That study was about vegetables in student dining halls, not about a restaurant’s whole menu. The lesson that carries over is modest: describe what the guest will taste, and do not lead with health.

Calorie labels: a small effect, low certainty

Some guests want to see calories on the menu. What happens to orders when you show them?

A Cochrane review from 2018 pooled three randomised trials of calorie labels on restaurant menus, with 1.877 participants. With the labels, people bought on average 47 kcal less per meal. The 95 % confidence interval ran from 15 to 78 kcal less.⁹ For a 600 kcal meal, that is about 7,8 % less. The reviewers rated the quality of the evidence as low. They wrote that the estimate “may change with further studies”.⁹

Source: Crockett and colleagues, Cochrane.⁹

In Iceland, nutrition values on a restaurant menu are voluntary, as we explain in our guide to allergens on the menu. If you show them, expect a small shift in what people order, not a large one.

Anchors and decoys

Two more ideas are common in menu advice. An “anchor” is a very expensive dish that makes the other dishes look reasonable. A “decoy” is a dish that is priced to push guests towards the one next to it. Both come from well-known research on how people judge prices in general.

We did not find a published field test of anchors or decoys on real restaurant menus. The ideas are plausible, but nobody has shown them on menus. If you try one, measure it as described below.

Test your own menu with your sales data

Your own sales data can tell you if the findings above hold for your guests. The test that the Tel Aviv café ran is easy to copy.

1. Choose one change. For example, move a dish from the middle of the mains to the top. Change nothing else, not the price and not the name.
2. Alternate the versions. Use the old menu and the new menu on alternating days or alternating weeks. That way, a busy week or a cold week affects both versions in the same way.
3. Measure the share, not the count. Compare the dish’s share of all mains sold, not only the number of plates. A busy Saturday sells more of everything.
4. Give it enough orders. The café needed about 950 orders over 30 days to see an effect of this size.¹ With a small menu section and a quiet month, the difference you see can be chance.
5. Look at the margin too. A dish that sells more is good only if its margin is good. Our guide to food cost and menu pricing shows how to compare the margin of each dish.

On a QR menu or a kiosk, a test like this is easier because you can change the screen without a reprint. In Resto, the back office shows sales by dish and by category for any period, so you can compare the two versions side by side.