The menu in text, always
A PDF menu cannot be read comfortably on a phone, cannot be searched and cannot be found by anyone looking for a dish. Text menus are the single highest-return change most restaurant sites can make.
People arrive at a restaurant site for three things: the menu, the hours and how to book. Most SF restaurant sites make all three difficult.
Restaurant website design in San Francisco costs $2,500 to $12,000. The decisions that matter are putting the menu in readable text rather than a PDF, offering direct ordering so you keep the margin third party platforms take, and making hours and reservations obvious on a phone.
A PDF menu cannot be read comfortably on a phone, cannot be searched and cannot be found by anyone looking for a dish. Text menus are the single highest-return change most restaurant sites can make.
Third party delivery takes a substantial share of every order. Direct ordering from your own site will not replace them, but every order that moves across is margin you keep, and San Francisco diners will use it if it works.
Different hours by day, brunch, happy hour, holiday closures and the week the kitchen is closed. Wrong hours cost you a customer permanently and they are the most common error on restaurant sites.
Resy, OpenTable, Tock or your own system. The site's job is to make booking obvious in one tap rather than to replace a system your floor team already knows.
Not stock. Real photographs of dishes you actually serve, taken well enough to look like the room. This is one of the few places where image quality directly changes whether someone walks in.
Menus change with the season and sometimes with the day. If a menu change is an agency ticket, the site is wrong within a fortnight and your staff stop pointing customers at it.
| Project | Range (USD) | Timeline | What you get |
|---|---|---|---|
| Single restaurant | $2,500 to $6,000 | 2 to 4 weeks | Home, menus in text, hours, reservations, location, editable by your team |
| Restaurant with ordering | $6,000 to $12,000 | 4 to 7 weeks | Above plus direct ordering, payment, order management, promotions |
| Group or multiple locations | $12,000 to $20,000 | 6 to 12 weeks | Location pages, per-site menus and hours, shared brand, group reservations |
| Menu system only | $1,200 to $3,000 | 1 to 2 weeks | Text menus, editable structure, schema, mobile layout |
| Care plan | $120 to $600 per month | Ongoing | Updates, menu changes, monitoring, seasonal work |
If a quote for a single restaurant site is above these numbers, ask what specifically it includes that this does not.
| Ask them | A good answer sounds like | Walk away if |
|---|---|---|
| Who owns the code and design files? | You do, unconditionally, in the contract | Ownership ends when the retainer does |
| What happens when a connected system is down? | Retries, queues, alerts, defined behaviour | It will not be down |
| What does month two cost? | A number, before you sign | We can discuss that later |
| Can I see Core Web Vitals on three live sites? | URLs, and an honest word about the slowest | They show you portfolio shots |
Write the answers down. Six months in, nobody remembers what was said on a call.
| Problem | What it costs |
|---|---|
| Menu as a PDF | Unreadable on a phone, invisible to search |
| Hours wrong or incomplete | A customer who goes elsewhere and does not return |
| Reservation link buried | Bookings that go to a competitor with a visible button |
| No direct ordering | Every order paying third party commission |
| Stock food photography | Read as dishonest by diners who notice |
| Nobody on staff can change the menu | A site your own team stops trusting |
A single restaurant runs $2,500 to $6,000. Adding direct ordering runs $6,000 to $12,000. Groups with multiple locations run $12,000 to $20,000.
Because it cannot be read comfortably on a phone, cannot be searched, and cannot be found by anyone looking for a dish you serve. Converting menus to text is usually the single highest-return change on a restaurant site.
It will not replace the delivery platforms, but every order that moves across keeps a substantial share of the value that currently goes to a third party. San Francisco diners will use direct ordering when it works properly.
Whichever your floor team already knows. Resy, OpenTable and Tock all work; the site's job is to make booking obvious in one tap rather than to replace a system that is functioning.
One of the few places where image quality changes whether somebody walks in. Real photographs of dishes you actually serve, taken well. Stock imagery is noticed and read as dishonest.
They must be able to. Menus change seasonally and sometimes daily, and if a change requires a developer the site is out of date within a fortnight and your own team stops sending customers to it.
It does more discovery work than the site does for most restaurants. Hours, photographs and menu links should match the site exactly, because mismatched details weaken local ranking for no reason.
No. We work with SF restaurants remotely on Pacific hours and would rather say that than imply otherwise. Not carrying Bay Area overheads is part of why our numbers differ.
Point us at the page and we will put a number and a timeline against it. If a cheaper option does the job, we will point you at it.