Check this page with an assistantOpens a chat asking it to summarise this article and name the evidence behind each claim.
Claude opens with the prompt on your clipboard: Anthropic does not support prefilled prompts on the web, and we would rather copy it than ship a button that drops it.
Your Profile Is the Product Page
Restaurant searches are overwhelmingly intent-driven rather than brand-driven: a type of food, near a place, sometimes with a constraint attached. That comparison happens on a map, using photos, reviews, hours, and menu information, and it frequently concludes before anyone visits a website.
The practical consequence is that effort spent on the website while the profile is incomplete is spent in the wrong place. A restaurant with a beautiful site, no menu on its profile, three photos from 2022, and unset holiday hours is losing the comparison before its website gets a chance to be good.
It also reframes what a restaurant website is for. Not persuasion, which already happened, but confirmation and action: is this still the place I saw, what exactly do they serve, and can I book or order without friction. Pages built to persuade get in the way of that.
What the Profile Has to Answer
Conversational and constrained searches are resolved from structured fields. Every blank field is an unanswered constraint, and unanswered constraints lose comparisons to restaurants that filled theirs in.
| What people want to know | Where it comes from | Common failure |
|---|---|---|
| Are you open right now? | Regular and special hours | Holiday hours never set, so the answer is wrong |
| Do you serve this dish? | Menu, as readable text | Menu is a PDF or missing entirely |
| Can I book? | Reservation link on the profile | Phone only, which loses evening bookings |
| What does it look like? | Current photos of food and room | Stock imagery, or nothing since opening |
| Is it right for my situation? | Attributes and review language | Attributes blank; reviews say only "great" |
| Can I order directly? | Ordering link you control | Only third-party links, so the margin goes elsewhere |
The last row is a commercial decision disguised as a settings screen. If the only ordering route on your profile is a delivery platform, you have handed them the transaction and the customer relationship for every order that starts in search. Adding your own ordering link alongside costs nothing and changes where repeat orders go.
Reviews Are Where Subjective Constraints Live
"Good for a first date", "quiet enough to talk", "handled our allergy well", "fast at lunch". None of these is a field on any profile. They exist in review text or they do not exist at all, which makes review language a discovery asset rather than a reputation metric.
That changes how you ask. Asking for a rating produces stars; asking what someone came for and whether you delivered produces sentences that answer future searches. A four-star review saying the kitchen handled a gluten allergy carefully at a busy service does more work than a five-star review saying lovely food.
Volume and recency also matter more than perfection, and both are things you can influence honestly by asking consistently rather than occasionally. Never incentivise positive reviews specifically, which platforms prohibit and detect. The mechanics of asking properly are in how to get more Google reviews, and the wider constraint-matching picture is in local businesses in AI search.
The Delivery App Trade
Third-party platforms buy you reach in exchange for margin and the customer relationship. That can be a rational purchase, particularly early. It becomes a problem when it is the only route, because then every customer you win belongs to the platform.
The workable position for most independents is both, arranged deliberately. Use the platforms for discovery, and make direct ordering visible and marginally better, whether that is a discount, a loyalty benefit, or simply being the obvious option on your own profile and site. The goal is that first orders can arrive anywhere and repeat orders come to you.
That only works if the direct path is genuinely easy. A direct ordering flow that takes six steps and a slow page loses to an app people already have, and the margin argument does not persuade a hungry person. This is one of the few cases where site performance is straightforwardly revenue, and the metrics worth targeting are in the Core Web Vitals guide.
Questions Restaurant Owners Ask
- Where do people actually choose a restaurant?
Overwhelmingly inside Maps, before your website is involved. Most restaurant discovery is intent-driven rather than brand-driven: people search for a type of food near a place, compare what appears on the map, and decide from photos, reviews, hours, and the menu. Your site is usually where a choice gets confirmed, not made.
- Does my restaurant website still matter?
Yes, for a narrower job than most restaurant sites are built for. It has to load fast on a phone, show the current menu as text, and let someone book or order in one step. Elaborate sites with the menu as a PDF and a slow image-heavy homepage lose people who had already decided to come.
- Should the menu be a PDF?
No. A PDF menu is slow on mobile, awkward to read, and difficult for machines to use, which means the dish someone searched for may be invisible even though you serve it. Publish the menu as real HTML text with prices, and keep the PDF only as an optional download if you must.
- Do reviews matter more than the rating?
Volume and recency generally do more work than a perfect average. A restaurant with a large number of recent reviews and a good-not-flawless rating tends to be treated as more established than one with a handful of perfect ones. What matters most is that reviews describe specifics, because that is where constraint-matching answers come from.
- Are delivery apps worth the visibility?
Delivery apps buy reach at a significant margin cost and accumulate the customer relationship rather than you. The pragmatic position for most independents is to use them for discovery while making direct ordering obviously available and slightly better, so repeat customers migrate to the channel you keep the margin on.
