What your menu analytics tell you
Menu analytics explained: dish opens, searches, languages, sources and peak hours, what each number means and the decision it supports in your restaurant.
In this article
Menu analytics are the numbers a digital menu collects about how guests use it: how they arrive, which dishes they open, what they search for, in which language and at what time. A paper menu cannot tell you any of this. Your till tells you what sold; menu analytics tell you what guests considered, looked for and did not find. Read together, the two answer questions most owners settle by instinct.
Below, each metric, the decision it supports, and a 15-minute weekly routine. The examples use the insights in ManyMenu, but the reasoning applies to any digital menu.
What a digital menu measures that paper cannot
A QR menu, a menu link and a TV board all run in a browser, so visits can be counted without asking the guest for anything. What is worth tracking:
- Visits by source: QR scan, link or TV screen.
- Views, dish opens and dishes opened per visit.
- Average time spent on the menu.
- What guests search for, including searches that return nothing.
- Filters used, such as vegan, halal or gluten-free.
- Languages guests chose, compared with their phones' own languages.
- Countries and cities, devices, operating systems and browsers.
- Referrers: Instagram, Google and other sites that sent visitors.
- A weekday by hour heatmap, in the guest's local time.
- Available dishes nobody opened.
Each can be read over 7, 30 or 90 days and compared with the previous period.
A dish open is interest, not a sale
A dish open means a guest tapped the dish to see its photo, description and sizes. It measures curiosity, not orders. The value comes from comparing opens with your till.
- High opens, low sales. Guests are interested, then something stops them: usually the price, a vague description or a photo that does not match the plate.
- Low opens, high sales. Regulars who know the dish and order it without looking. Leave it alone.
- High opens, high sales. A star. Keep it where it is.
Say a café in Casablanca finds its avocado toast among the most opened dishes but sells only a handful a day. The photo works; the price at 65 DH probably does not. Try 55, or add an egg and keep 65, and compare the next 30 days. This is menu engineering with a second data source, covered in menu pricing psychology.
Dishes nobody opened deserve their own look. If an available dish was not opened once in 30 days, guests either do not see it or do not understand it. Rewrite the name, add a photo, move it up, or remove it.
Searches, filters and languages: free market research
Searches are guests telling you, in their own words, what they want. Searches with zero results are the most useful line in the whole report. "Decaf", "vegan", "kids" with no match means guests came looking for something you do not list. Either you do not offer it, which is a menu decision, or you offer it under another name ("déca", "plant-based", "petite portion"), which is a renaming job of two minutes.
Filters work the same way. If the gluten-free filter is used every week and only two dishes carry the label, check which other dishes qualify and label them.
Languages are where many venues leave guests behind. Insights compare the language guests chose on the menu with the language of their phone, and show whether you offer it. If a steady share of phones are set to German and your menu exists in French, Arabic and English, German is your next translation. With Morocco welcoming a record 19.8 million tourists in 2025 (+14%), this matters well beyond Marrakech. In ManyMenu, a new language is one tap: the AI translates the menu and keeps the names that should stay, so a tajine stays a tajine.
Where guests come from and when they look
Visits by source tell you which doors work. If QR scans are low compared with your covers, the codes on the tables are not visible: they are hidden behind napkin holders, too small, or missing on the terrace (see our French guide to creating a QR menu). If the link brings almost nothing and Instagram never appears among referrers, the link is not in your bio or nobody mentions it in stories.
Countries and cities separate tourists from locals. Devices matter too: if most guests open the menu on small phones, keep dish names short.
The heatmap shows when guests open the menu, by weekday and hour, in their local time. It informs three decisions:
- Staffing. A fast-food that sees views climb on Fridays from 23:00 knows when it needs a second person on the counter.
- Happy hours. If the bar is quiet from 17:00 to 19:00 on weekdays, that is your window.
- Breakfast cut-off. A brunch café in Rabat that sees views peak at 11:45 on Saturdays could extend breakfast to noon.
A 15-minute weekly routine
Pick a fixed slot, such as Monday at 10:00, and go through the same steps each week.
- Set the period to 7 days, compared with the previous 7 days (2 minutes).
- Check visits by source: a sudden drop in QR scans often means a table tent went missing (2 minutes).
- Read the zero-result searches and note any word that appears more than once (3 minutes).
- Compare the ten most opened dishes with the ten best sellers on your till (4 minutes).
- Pick one dish nobody opened and fix its name, photo or position (2 minutes).
- Glance at the heatmap and languages, then write down one action for the week (2 minutes).
Once a month, switch to 30 or 90 days for the bigger decisions: staffing, opening hours and new translations.
Caveats: small numbers, seasons and privacy
Menu analytics are a guide, not a verdict.
- Small numbers. At 60 visits a week, one group of tourists can double a language. Wait for a pattern over several weeks.
- Seasonality. Summer in Agadir is not winter in Agadir. Compare with the same season when you can.
- Ramadan and holidays. During Ramadan the heatmap shifts to the hours after iftar. Compare Ramadan with last Ramadan, not with the month before.
- Like with like. Compare a week with a week and a month with a month, and change one thing at a time.
On privacy, ManyMenu guest menus set no cookies. A visit is a random id kept per browser tab, and nothing personal is stored. The IP address is used only for rate limiting; country and city come from the network, and the device type from the browser.
Start with one question
Do not try to read everything at once. Choose the question that costs you most today, whether that is unsold dishes, empty hours or missing languages, and let the numbers answer it. You can see the guest side of a digital menu in the live demo, compare plans, and when you want your own numbers, request access to ManyMenu.
