Live is not the same as found
Putting a page online and having a search engine reliably discover, read and understand it are two different jobs. A misconfigured robots file or a stray noindex, two addresses claiming to be the real one, a page nothing links to, a menu that only appears after scripts run, missing restaurant data, or search verification that was never finished: any of these can leave important pages undiscovered, excluded or misunderstood.
The owner rarely sees it happen. The site looks fine. What changes is where guests land when they search the restaurant’s name, its menu, its cuisine or its street: on a directory, on a marketplace that charges commission, or on the restaurant next door. The restaurant ends up paying for visibility it should own.
Fixing this on a general website builder usually means finding a technician, adding plug-ins and keeping them maintained. TableSpark treats search readiness as part of the restaurant website itself. No one can promise a ranking or that a page will be indexed; what can be done is to remove every avoidable reason for a page to be missed, and this page lists how.
- Dish names, descriptions and prices are written into that HTML, so a crawler reads the menu without running any script.
- Every page carries one canonical address on the site’s canonical host.
- Every live page embeds Restaurant structured data: name, address, telephone, opening hours, special opening dates, cuisine, price range, social profiles and a map link.
- After each publish, the site’s live addresses are sent to IndexNow and the sitemap is submitted to Google Search Console.
Pages a crawler can read
Each page of a restaurant site is rendered to complete HTML on the server and sent as the response. Dish names, descriptions and prices are written into that HTML, so a crawler reads the menu without running any script.
The live site shows the published version. Edits saved in the Builder stay private until you press Publish. Each page’s title uses the page’s own search title, then the site’s, then the restaurant’s name. The description follows the same order.
One address for every page
When the same page answers at two addresses, a search engine has to pick one, and it may pick the wrong one or split the credit between them.
- One canonical addressEvery page carries one canonical address on the site’s canonical host. The canonical host is the restaurant’s own domain once that domain is verified, and the tablespark.uk address until then.
- A permanent redirect to your domainOnce the domain is live, a request to the tablespark.uk address is redirected permanently to the same page on the restaurant’s domain.
- Only when the domain is readyThat redirect is switched on only after a DNS check of the domain passes. The check is repeated hourly and treats any doubt as not ready.
- A managed certificateA restaurant’s own domain is connected on Growth and Full with a managed certificate and a minimum of TLS 1.2.
Sitemap and robots
| Sitemap | Each site serves its own sitemap, listing the home page, every live page and each published post or event, dated with the last publish. |
|---|---|
| Robots | Each site serves its own robots file, which welcomes crawlers, keeps them out of the service’s internal paths and names the sitemap. |
| What stays out | Legal pages, Builder previews and embedded widgets are marked noindex so they stay out of search results. Every other live page is marked index, follow. |
| Wrong addresses | A wrong address on a live site returns the restaurant’s own page with a real 404 status and a noindex header. |
| Brief faults | If the database cannot be reached, the site answers with a temporary 503 and a retry time, so a brief fault is never read as a deleted site. |
Structured restaurant data
Structured data is how a page tells a machine, without guesswork, that this is a restaurant, where it is, when it is open and what it serves.
Every live page embeds Restaurant structured data: name, address, telephone, opening hours, special opening dates, cuisine, price range, social profiles and a map link.
The menu is included as Menu, MenuSection and MenuItem data, with prices in pounds, dietary suitability and allergens. A site that takes bookings declares it with a ReserveAction that points at the booking page, and a site that takes orders adds an OrderAction.
News posts and events carry Article or Event data with a breadcrumb trail. Each site also serves an llms.txt for AI assistants, with its name, description, menu link and live pages.
Old addresses keep working
A restaurant that replaces its website usually throws away years of links: the menu address printed on a flyer, the booking page a local blog linked to, the address Google has shown for a decade. Each one becomes a dead end unless something sends it on.
The redirect manager keeps up to 500 redirects per restaurant, permanent by default, each resolved in one hop: chains are flattened and loops are refused. A redirect can point only to a page of the same site or to one of the restaurant’s own hosts.
Common old addresses such as /menu or /book move by themselves to the one live page that does that job, when the match is unambiguous. The Builder suggests redirects from the page addresses of the previous website, and from pages renamed or removed since the last publish. Redirects are consulted only for an address that would otherwise be a 404, so normal pages never wait for the lookup.
Telling search engines
After each publish, the site’s live addresses are sent to IndexNow and the sitemap is submitted to Google Search Console.
A restaurant’s own domain gets its own Search Console property, verified by a file the service serves, and an hourly job completes this without a click from the owner.
The check before you publish
Before you publish, a checklist reads the pages as they will be served and lists what to put right: missing opening hours or address, empty pages, missing search descriptions, photos without alt text and internal links that lead nowhere.
It flags a search title longer than 60 characters and a description longer than 160.
Fast on a phone
Most restaurant searches happen on a phone, on the way somewhere.
- Photos are offered in four widths, so a phone downloads a phone-sized image.
- Images further down the page load lazily, and the main image is preloaded for each screen size.
- Fonts are served from the site’s own address and text stays readable while they load.
- A test renders the home page of every template and fails if the HTML is 88 KB or more.
- Site pages are sent over HTTPS with Strict Transport Security, a no-sniff header, frame protection and a referrer policy.
Questions owners ask
Can Google read the menu without running JavaScript?
Each page of a restaurant site is rendered to complete HTML on the server and sent as the response. Dish names, descriptions and prices are written into that HTML, so a crawler reads the menu without running any script.
What happens to the links to my old website?
The redirect manager keeps up to 500 redirects per restaurant, permanent by default, each resolved in one hop: chains are flattened and loops are refused. Common old addresses such as /menu or /book move by themselves to the one live page that does that job, when the match is unambiguous.
What structured data does a TableSpark site publish?
Every live page embeds Restaurant structured data: name, address, telephone, opening hours, special opening dates, cuisine, price range, social profiles and a map link. The menu is included as Menu, MenuSection and MenuItem data, with prices in pounds, dietary suitability and allergens.
How this page is kept true
This page makes 33 statements about how TableSpark behaves. Each one is recorded with the place in the code, the database or the test suite that enforces it, and an automated check fails when a statement loses its evidence. The suites named below exercise the behaviour described here.
Automated suites behind this page
- domain-search-sync
- seo-domain-gsc
- seo-tenant
- seo-tenant-agent-surface
- site-checklist
- site-redirects
- site-redirects-db
- tenant-hero-lcp
- tenant-not-found
- tenant-page-weight
- tenant-subdomain-lifecycle
- tenant-upstream-error-not-gone
Read next
See it on your own restaurant
Build your site free and publish when you are ready. Plans start at £19 a month, excluding VAT.