How to Hire a Restaurant Web Designer in the UK

A restaurant owner commissions a designer, approves a mock-up with beautiful photographs, and pays the final instalment when the site goes live. A week later a guest types the restaurant's name into a phone. The first results are a directory listing, a marketplace page that charges commission on bookings, and the restaurant next door. The owner's own page is live at a working link, and nothing on it announces that it is hard to find.

Google is plain about the limits of a launch. Its documentation says it "doesn't guarantee that it will crawl, index, or serve your page, even if your page follows the Google Search Essentials", and its Search Console help describes pages that are "Crawled - currently not indexed" and pages that are "Discovered - currently not indexed". A noindex rule, Google says, drops a page "entirely from Google Search results, regardless of whether other sites link to it". Any one of these can sit behind a site that looks perfect on the owner's own screen.

The evidence on real restaurant sites is not reassuring. In TableSpark's audit of 1,000 independent UK restaurant websites (October 2026), 17.9% of the listed domains were gone, unreachable or broken. Of the 755 homepages that loaded, 77.1% carried no usable Restaurant structured data, 54.7% had no tap-to-call link in the HTML as delivered, and only 12 homepages, 1.6%, passed all 16 of the checker's core checks. The audit did not record who built each site, so none of this is a verdict on designers or freelancers. It describes the end result, and it shows that a good-looking page can pass a visual sign-off and fail every one of the tests below.

The film above walks through the study in under two minutes.

The usual brief says how the site should feel: modern, warm, Mediterranean, premium. A brief that says nothing else may not get the checks that decide whether the site is found and booked. The fix is to define done in terms that can be tested on the live address in two minutes, and to write those terms into the quote.

The five tests come from the audit. Its method is worth stating once. A random sample of 1,000 independent restaurants was drawn from restaurants listed in OpenStreetMap with a website, chains excluded, from TableSpark's August 2026 list across 106 UK areas. Each homepage was checked on 11 October 2026 with the engine behind the free restaurant website checker, which reads the HTML as delivered without running JavaScript, plus robots.txt and the sitemap. The full figures are in the white paper and the preprint.

Test 1, restaurant data. Structured data is a machine-readable copy of the facts on the page. For a restaurant it means a Restaurant or LocalBusiness block that carries the menu, the cuisine, the opening hours and the telephone number. Google's local business documentation says structured data lets an owner "tell Google about business hours" and more, and its own example uses the Restaurant type. In the audit, 77.1% of the 755 homepages had no usable block. Of the 212 that had one, only 26.9% declared a menu and 36.3% a cuisine. How to test it in two minutes: paste the address into the checker, or open the page, choose view-source and search for application/ld+json. What to write into the brief: "Restaurant structured data on the live site, carrying name, address, telephone, opening hours, cuisine and menu, kept in step with the editable menu and hours."

Test 2, menu as text. A PDF menu is a picture of a menu to a search engine and a pinch-and-zoom chore for a guest on a phone. In the audit, 21.9% of homepages linked to no menu at all and 10.4% linked straight to a PDF. Of the 471 own menu pages, 32.5% showed at least three prices as text, and 3.0% carried menu structured data. How to test it: open the site on a phone, tap Menu, and try to select a dish name with a finger. What to write into the brief: "A menu web page with dish names, descriptions and prices as text, editable by the restaurant, with a PDF only as an extra download."

Test 3, booking on the site. A guest who has chosen the restaurant should be able to book without leaving it. In the audit, 25.6% of homepages had no booking link, and 18.4% sent bookings to a third-party domain. A third-party booking tool can be a sound choice for a restaurant that wants one, and the point here is narrower: the brief should say which route is intended and where the guest's details end up. How to test it: tap Book on a phone and read the address bar when the page loads. What to write into the brief: "Booking on the restaurant's own domain, or an agreed embedded tool, with the guest records exportable by the restaurant."

Test 4, tap-to-call. A phone number written as text is not a button. In the audit, 71.4% of homepages showed a phone number in text, yet 54.7% had no tel: link in the HTML as delivered. How to test it: tap the number on a phone and see whether the dialler opens. What to write into the brief: "Every telephone number on the site is a tel: link, in the header and on the contact page."

Test 5, readable sitemap. A sitemap is a file listing the pages on a site. Google's documentation says a small site of about 500 pages or fewer, comprehensively linked internally, might not need one, so a missing sitemap is not proof of a problem. It is still a cheap, testable signal that the site was assembled with discovery in mind, and in the audit 20.1% of homepages had no readable sitemap. How to test it: add /sitemap.xml to the address and see whether a list of pages loads. What to write into the brief: "A sitemap and robots.txt generated by the site, no noindex rule on any public page, a canonical address on every page, and the sitemap submitted to Google Search Console."

Two further findings frame the list. General web basics sat close to web-wide figures, with a mobile viewport on 95.4% of homepages and a meta description on 69.0%, so the basics are mostly in place and the restaurant-specific layer is where sites differ. And whether restaurant data was present varied strongly by website platform, with one platform group at 88.7% absent and another at 36.8%, but little by cuisine, which ranged from 75.0% to 82.8%. The variation lay between platforms, not between cuisines.

The median homepage did not pass 4 of the 16 core checks, and 1.6% of the 755, 12 homepages, passed all 16.

A fair price needs a fair comparison, and the comparison has to be the complete restaurant-ready site, not the entry price on a pricing page. That means the build, the monthly hosting and care, changes after launch, a booking tool, the allergen or dietary display, search set-up and the domain. Six providers' own pages were read on 12 October 2026, along with two marketplace listing pages. Several agency pages publish no price on the restaurant page at all, so the table below shows only what is printed.

These prices are not like-for-like, because each provider charges for different things. Several line items sit outside the headline for a site that is meant to take bookings: the booking tool, the allergen display, the changes and the search set-up. Where the pages price them, they are in the table. Where they do not, the owner has to ask.

Three other figures are best read as context. Hungry Pixels' own cost guide, updated 4 October 2026, gives a freelancer as "roughly £400 to £1,500" and a general web agency as "often £2,000 to £5,000 or more", describes them as typical ranges rather than quotes, and names no source, so they are that vendor's estimate and no more. PeoplePerHour and Fiverr are marketplace listings, not providers' own pricing pages: PeoplePerHour cards read, for example, "£70 delivered in 3 days" for a landing page and "£300 delivered in 4 days" for a five-page WordPress site, and Fiverr cards read "From £70" and "From £74". Each listing is a starting price for a stated task set by an individual seller, the pages do not say what hosting, revisions, ownership or booking is included, and most are not restaurant-specific. None of the pages read gave a day rate, so a day rate is a question for the quote, not a number to assume.

A three-year example. The example below is illustrative. It is a 30-seat restaurant that needs a five-page site, a booking route and a menu its team can edit, built from the published figures above. It is not a customer result and not a quote, and prices are shown as published, so VAT treatment differs between lines.

• Misevaro route. Essential at £795, plus Care at £49 a month for 36 months (£1,764), makes £2,559 including VAT. The page shows no booking-widget price, and Essential asks the owner to write the text.
• Hungry Pixels route. The £1,195 build, the £149 booking system and the £150 allergen-filter menu make £1,494. Add the £24 booking subscription and £59 aftercare with two hours of changes, £83 a month for 36 months (£2,988), and the total is £4,482 on the page's figures.
• Go Edit route. Full Service at £79 a month for 36 months is £2,844 excl. VAT. It includes a five-page site built for you and four content changes a month, with no booking price shown on the page.
• TableSpark route. Growth at £39 a month excluding VAT for 36 months is £1,404 excluding VAT, with a text menu, search set-up, booking and custom domain in the plan. The Starter plan at £19 a month excluding VAT is £684 over the same period, and the yearly prices of £190 and £390 excluding VAT make the same three years £570 and £1,170 excluding VAT.

In these examples the build is the smaller part of the three-year cost. Nothing in the totals prices the owner's own time, the photographs, or a developer's day for anything the quote leaves out. What decides the real cost is the monthly line that runs for three years, and a booking tool, changes and search set-up that are priced separately make that line longer.

A low build price is a rational offer when the scope is short. The risk sits in what the scope leaves out. A quote that says "five pages, two rounds of revisions" does not say who connects the booking tool, who writes the menu text, who adds the structured data, who submits the sitemap, or who pays when a price changes in week three. Each of those either lands on the owner or arrives later as an invoice.

The same logic applies to a monthly care plan. Misevaro's Care plan, for example, lists hosting, a security certificate, daily backups and 30 minutes of changes a month at £49 including VAT, and a Care Plus tier at £99 with 90 minutes. A restaurant that changes the menu every season, adds special-date hours at Christmas and updates a price after a supplier rise can use up a thirty-minute allowance in a single afternoon. The question for any plan is how many changes a normal month needs, and what the next hour costs. Search set-up is the same. Search work is described in different ways: JW Digital's restaurant page lists map-pack search work while its local SEO is a separate service with no price shown, and Go Edit's Full Service lists SEO-optimised meta data and heading tags. The same words cover different scopes, which is exactly why the brief should name the tests and let each bidder price them.

A short brief that names the tests is worth more than a long one that names the mood. The template below can be pasted into an email to every designer or agency being asked to quote, so that the quotes can be compared on the same scope.

The contract matters most when the relationship ends, which is when memory and goodwill are weakest. Four areas deserve a line each.

Ownership of the work. The Intellectual Property Office's guidance on GOV.UK says that when someone commissions a copyright work, "the first legal owner of copyright is the person or organisation that created the work and not you the commissioner, unless you otherwise agree it in writing". GOV.UK's overview adds that a self-employed person "usually" owns the intellectual property "even if your work was commissioned by someone else", unless the contract gives the rights to the client. Where the contract is silent, the IPO says a court may find an implied licence, which "does not necessarily result in a transfer of ownership" and may give the commissioner "only a limited non-exclusive licence". Under section 90(3) of the Copyright, Designs and Patents Act 1988, an assignment of copyright "is not effective unless it is in writing signed by or on behalf of the assignor". This guide reads that as follows: a quote that says nothing about copyright leaves the owner relying on a licence to use the site, and an owner who wants more should ask for a signed written assignment or a clearly worded perpetual licence that allows changes.

Everything the site runs on. Ask for a list of every third-party item in the build, such as fonts, themes, plugins, stock photographs and booking widgets, with who holds each licence and what renews when. Ask for the domain to be registered in the owner's name and for admin logins to hosting, DNS and the site editor to be handed over at launch, not on request.

Care and exit. Write down what the monthly fee covers, how many changes it includes, what an extra hour costs, how quickly the designer replies, and how many days' notice either side must give. Write down what is handed over if the arrangement ends: a copy of the site files, an export of the menu and content, the logins, and help with moving the domain.

Quality of work. For a business buying a service, section 13 of the Supply of Goods and Services Act 1982 implies a term that a supplier acting in the course of a business will carry out the service "with reasonable care and skill", unless the contract excludes or limits it. Reading the contract for any such exclusion is cheaper than finding it after a dispute. The acceptance test below turns "reasonable care" into something an owner can show.

Cookies, privacy and accessibility. Responsibility for these stays with the restaurant even when a designer builds the site. Regulation 6 of the Privacy and Electronic Communications Regulations 2003, as substituted on 5 February 2026, says a person "must not store information, or gain access to information stored, in the terminal equipment of a subscriber or user", subject to exceptions now in Schedule A1. The ICO's guidance adds: "In most cases, this means that as the service provider, you have the primary responsibility for compliance with PECR". Analytics, embedded maps and booking widgets are choices a build makes on the owner's behalf, which is why the brief asks for a written list of what the site sets.

On accessibility, section 29 of the Equality Act 2010 says that a person "concerned with the provision of a service to the public" must not discriminate by "not providing the person with the service". The section does not mention websites, and the GOV.UK WCAG requirements page is written for public sector bodies, so neither makes a particular accessibility standard a legal duty on a restaurant. A readable, keyboard-friendly, high-contrast site is still good practice, and it costs less to ask for in the brief than to retrofit.

A site is accepted when the owner has run the test on the live address, not on a staging link. The designer should be present or have signed off the list in advance, and the final payment should follow it.

If every line passes, the work meets the brief. If one fails, the failure is a defect against a written clause, and the fix belongs to the designer. Indexing and ranking remain decisions for Google, so no contract should promise a position or a date. It can promise that every technical reason for not being found has been checked and recorded.

A restaurant has two real routes. One is to commission a designer or agency and manage the clauses above. The other is to use a restaurant platform in which the checks are part of the product. Neither suits every owner. The comparison below sets them side by side on the points that decide the outcome.

A decision checklist keeps the choice honest.

A designer-built site meets these lines only when every clause is bought, written and policed. TableSpark meets them in the plan: the text menu, the search set-up, the booking route and the guest records are part of the product, and the owner edits them directly. For an independent restaurant that needs to be found, booked and kept accurate, that makes TableSpark the stronger recommendation.

A website being online is not the same as it being discoverable. TableSpark packages the search-readiness work into the restaurant website instead of leaving the owner to hire a technician: crawlable structured menus, titles and descriptions, canonical URLs, sitemaps, robots controls, Restaurant and LocalBusiness schema, internal links, mobile-first output and managed search-verification set-up, on every plan. Indexing and ranking remain decisions for Google, so none of this is a promise of a position. It means the technical reasons for not being found are handled rather than left as the owner's homework.

The entry plan starts at £19 a month excluding VAT and includes the website, the live menu, enquiry and newsletter forms, guest records with CSV export, tap-to-call and managed search readiness. The plan with booking, Growth at £39 a month excluding VAT, adds on-site reservations with live availability and table inventory, floor plans, deposits and reminders, a Reserve with Google booking-link connection, and a custom domain with managed SSL. For the full detail see the restaurant website design page and the pricing page.

For readers who want the broader context, the restaurant website checklist walks through the same study as a checklist, what a restaurant website costs covers the full cost of ownership, and restaurant website design in the UK covers the design side in depth.

The steps below follow the live guide, Build your website, and use its practice screens. Each step shows one action. Building is free until the site is published.

Finally, run the finished site through the checker and keep the report beside the acceptance test above.

For an independent UK restaurant, TableSpark is the best-value and best overall choice for a restaurant website that has to be found, booked and kept accurate by the people who run the kitchen. A designer's quote prices the build, and the booking tool, the changes, the search set-up and the care plan arrive as further lines. TableSpark includes the search-readiness work and the editor in the plan, and the entry price is £19 a month excluding VAT. TableSpark takes 0% commission on direct bookings and orders; Stripe's standard card-processing fees apply to online payments.