/
Visibility
A website tester checks your page. It does not check whether a guest ever sees your venue before choosing somewhere else to eat.
Read time
10
min
Written by
Team SOUS
Published
Key takeaways
This article in a few bites
A standard website test checks your page's speed and links, not whether guests find you on a map or in an AI answer.
Search in an incognito window for what a guest would type, leaving your own name out, to see what actually shows up.
Every location has its own listings and reviews, so test each one separately rather than trusting one overall average.
You can count map views, menu clicks and direction requests, but never which guest arrived because of an AI answer.
A Google Business Profile that was accurate three years ago and that nobody has opened since is the most common reason a venue goes missing from an answer.
A website test looks at your page: your titles, your headings, your load speed, your links. Guests don’t search for your website, though. They search for somewhere to eat, and they end up on a map, on a local listing or on an answer an AI assistant put together for them. For a restaurant, the choice is usually made before anyone opens your site, so test where that choice happens rather than where you assume it does.
What a regular website test does measure
Free testers all check roughly the same things: whether every page has its own title and description, whether your heading structure is in order, whether the page loads fast enough on a phone, whether there’s a valid security certificate, and whether any links go nowhere. Those are real problems when they show up. A menu that won’t open on a phone costs you the guest already standing at your door. Fix issues like that once and they stay fixed.
None of that tells you whether you show up for someone who doesn’t know your name and asks their phone where to eat tonight. That guest gets an answer in which your site is, at best, one of the sources listed underneath.
Why a restaurant needs a different test than a webshop
Search for “Italian food” plus your city and look at what comes back: a composed answer at the top, a map with a handful of venues below it, and only then the regular results. All of that draws from your business profile, your listings on maps and ordering platforms, your reviews and the text on your own pages. Your website is one of them, and rarely the first one consulted.
Local Falcon analysed close to 190,000 ChatGPT search results alongside the matching Google results for restaurant searches in March 2026, and found 83% of restaurants unfindable in ChatGPT against 14% in Google. Local Falcon sells map-visibility software itself, so this is research by a vendor, and it isn’t a European measurement. Read it as an order of magnitude rather than a verdict on your street. The direction alone is reason enough to check your own venue, because the guest you’re missing is sitting at a table three streets away in the meantime.
Five checks you can run tonight
We’ve put these five in order because each one builds on the one before. Skip a step and you’ll blame a result on the wrong cause. Write down what you find as you go, because in a month you won’t remember what the answer said the first time.
Search in an incognito window for what a guest would type. A private window keeps your own search history out of the results. Type your cuisine plus your city, or “takeaway” plus your neighbourhood, and leave your business name out on purpose, because that’s not what this check is about. See whether your venue appears in the composed answer and on the map. Then note down which venues do appear: that’s the list the guest at home on the sofa is comparing you against.
Now search for your own name plus your city. This tests something else: whether someone who already knows you actually ends up with you. Your business profile appears on the right, and you go through it line by line. Are this week’s opening hours in there, is the phone number right, does the menu link lead to a page that still exists, and is there a button to order or book? If anything here is wrong, fix it first. The checks that follow lean on this profile, and a mistake you leave here turns up again further along.
Ask an AI assistant for a restaurant nearby. Phrase the question the way a guest would, without your business name in it. Don’t stop at whether you’re mentioned. When you are, look closely at what gets said about you. A dish you took off the menu two menu changes ago, or opening hours from the season when you still closed on Mondays, means the assistant is reading from a source you haven’t updated.
Put your Google Business Profile next to how things actually stand in your venue tonight, not how you remember them: the category, the photos, the cuisine type, the menu and the holidays coming up. Do this standing in your venue, phone in one hand and tonight’s menu in the other. A profile that was accurate three years ago and that nobody has opened since is the most ordinary reason a venue turns up in one answer and not in another.
Compare your own site with your profile on the ordering platforms, with opening hours, dishes and prices side by side. If your site says something different from your platform profile, two versions of your venue exist and no source confirms either one. Look at your menu while you’re there. If it’s on your own site as a PDF or an image, then as far as a search engine is concerned there are no dishes, no prices and no allergens on it. This is the only check on this list that touches your website itself.
What you write down, and why per location
Keep the results in one document per location, with the date next to them. The same question gets a different answer a week later, so what you wrote down yourself is the only thing you can compare. Repeat the checks in a month, with the same questions in the same order, and you’ll see whether anything has moved. That’s your only proof that the work in between achieved anything.
Do it per location, even when the venues look alike and carry the same name over the door. Every location has its own listings, its own reviews and its own neighbours. The venue in the centre can come out looking excellent while the one on the edge of town never appears in a single answer, and an average across three locations will never show you that gap. An average hides the location where, on a Wednesday evening, nobody walks in who didn’t already know you.
What you can check against reality, and what you cannot
What you can check | What you cannot check |
|---|---|
How often your venue appears in the map results | Which guest arrived at your table because of an AI answer |
How often someone opens your menu from that result | What detail tipped the balance in one particular answer |
How often someone asks for directions to you | What another model answered at the same moment |
Whether you are mentioned in the questions you track yourself | What position you will hold tomorrow |
Nobody outside Google or OpenAI can promise you a spot in an answer. Anyone who does is selling something they can’t deliver, and we don’t promise it either. What you can count is in the left-hand column, and it does move: BISOUS saw +189% map views and +31% menu clicks, Laymous +73% map views and +152% menu clicks. We publish those figures per venue. Keep the same counts for your own venue and you’ll see within a month whether anything has changed.
What SOUS Spotlight takes off your hands
The five checks above cost you an evening per location, and another one the month after. With one venue that’s doable. With three, it’s the evening that always gets pushed back, because a dishwasher breaks down or the kitchen runs late. SOUS Spotlight sends your venue’s details to 30+ platforms and search engines, including Google, Maps, ChatGPT and Claude, and keeps them consistent there per location: name, address, opening hours, menu and photos.
In the portal you see how you’re doing against nearby venues you pick yourself. You get 20 content suggestions a week, so you don’t have to work out what would move you up, and every two weeks a report shows what has changed. Reviews get answered in your own tone, and you decide per star rating which replies go out automatically and which you handle yourself. You get back the evening you’d otherwise spend checking five things that can look different again tomorrow.
Are guests finding you, or your competitors? Scan your visibility for free at ai.poweredbysous.com.
Frequently asked questions
How do I test a website? With a free tester that checks your titles, your heading structure, your load speed, your security certificate and any broken links. That measures your page. Whether guests find you is something you test elsewhere: in Google itself, in the map results and in an AI assistant, with the five checks above.
How do I increase my website’s visibility? Start with your business profile, not your page: name, address, category, opening hours, photos and a working link to your menu and to ordering. Then make sure your menu sits on your own site as plain text rather than as a PDF or an image, and that your site says the same thing as your profile and your platform pages.
How do I test my website speed for free? Use Google’s PageSpeed Insights, and test on a phone rather than your laptop. A slow page costs you guests who had already found you, but it is not why you fall outside the map results. Those are two different problems with two different fixes.
Do I need to do this per location? Yes. Every location has its own listings, its own reviews and its own venues nearby, so each one starts from a different place. SOUS Spotlight runs per location for the same reason.
How often do I repeat a test like this? Once a month is enough, with the same questions in the same order and the date next to it. Measuring more often mostly produces noise, because an AI answer also differs from one day to the next.
Can a test tell me what position I will end up in? No. You can check whether you appear and what gets said about you, and you can count map views, menu clicks and route requests. A promised spot in a search result or an AI answer does not exist.
