Many restaurants order a website that looks like everyone else’s, and then find out guests only open it for the menu and the address. Here is what a restaurant or cafe website needs first, what can wait, when your own app makes sense, and how to launch a first version without overspending.

A day in a cafe: a hypothetical example

Say you run a 40-seat cafe that also does takeout. Mornings are coffee and breakfast, lunch is the nearby offices, and evenings are groups booking tables. Some orders come through a delivery platform, some guests text you on WhatsApp or Instagram, and some call.

From the guest’s side, the questions are simple:

  • what do you serve and how much does it cost;
  • where are you and when are you open;
  • can I book a table for tonight;
  • can I order ahead and skip the line, or get it delivered.

From the host’s side, the problems are different: orders arrive in five places, bookings live in a notebook, and the Instagram menu has not been updated since last season. A good website or bot fixes both sides.

What a restaurant website needs first

Start with what guests actually come for. It is a boring list, but it is what brings people in.

  • A menu with prices. Sections, clear names, prices, photos of at least the main dishes. Not a PDF people have to pinch and zoom on a phone, but a normal page.
  • Address, map and hours. Mention parking and how to find the entrance if it is around the back.
  • One-tap contact. Phone, WhatsApp, Telegram. Many guests prefer texting to calling, so a visible messenger button matters more than a fancy animation.
  • Built for phones. People almost always open a cafe website on a phone, often on the go. Check every screen on a small one.

The key rule: the menu online must match the menu in the room. If prices are out of date, a guest arrives, sees different numbers and feels misled. Decide in advance who updates the menu, and ask for an admin panel where it takes a couple of minutes without a developer.

Table booking: start simple

Online booking is worth it if you have full evenings and guests ask about free tables in advance. There are two levels.

Booking request. The guest picks a date, time and party size and leaves a name and phone number. The host gets a notification, for example in Telegram, and confirms. It is quick to build, and for most places it is enough.

Floor plan. The guest sees the room and picks a specific table, and the system prevents double booking. It is nicer but costs more, and it needs discipline: if staff seat walk-ins without marking them, the plan soon stops matching reality.

Start with requests. Add the floor plan once there are more bookings than the host can confirm by hand.

Takeout and delivery: platform, your own website or a bot

This is where money is most often wasted. There are three paths, each with its own strengths.

A delivery platform. You build nothing, and new guests find you in a shared catalog. But the platform takes a commission on every order, decides how you are shown, and keeps the guest data. You cannot message a regular or reward them directly.

Ordering on your own website. The guest picks dishes, enters an address or chooses pickup, and can pay online. No platform commission, and the customer base is yours. But a website does not bring new guests on its own: it needs promotion, and you have to run delivery yourself.

A Telegram bot. The same ordering, inside a messenger: menu buttons, cart, address, confirmation. Nothing to install, and the kitchen gets the order straight into a work chat. It suits a short menu and regular guests. For a full comparison of formats, see website, app or Telegram bot.

Do not leave the platform overnight. Use it to reach new guests and move regulars to your own website or bot over time, for example with a card in the bag that has a promo code for direct orders.

Your own restaurant app: when it is worth it

An app feels like a serious step, but it only pays off under certain conditions:

  • many regulars who order every week;
  • a loyalty program you are ready to develop;
  • a need for push notifications about offers and order status;
  • several locations or your own delivery with steady volume.

If at least two of these are missing, few people will install the app. Nobody wants an app for a single lunch. Apps also have fixed costs: the Apple Developer Program at $99 a year, Google Play at $25 once, plus updates for new phone OS versions. What drives the price is covered in how much a mobile app costs, and if delivery is the core idea, see delivery app.

Loyalty programs: what can wait

Points, cashback, stamps and personal offers work well, but only once you have a guest base and know how people order. Building complex loyalty into the first version is risky: you guess the rules and then redo them.

Start small: collect guests’ phone numbers when they order or book (with their consent) and note who comes back. In a couple of months you will see who is worth rewarding and for what. Then the bonus rules are based on real data.

Other things that can usually wait:

  • online payment, if guests already pay on pickup or delivery;
  • a customer account with order history;
  • integration with your kitchen and inventory system;
  • multiple languages, if nearly all guests speak one.

Common mistakes

  • A beautiful site without a menu. Big interior photos, a video on the first screen, and the menu hidden or posted as an image. The guest goes back to the maps and picks another place.
  • A menu nobody updates. The site was handed over, but staff have no admin access or the panel is awkward.
  • Everything at once. Floor-plan booking, delivery, an app and loyalty in the first version. The project drags on, and half the features go unused.
  • Orders nobody sees. The site takes an order, but the notification goes to an inbox the kitchen never checks. Agree in advance where orders land and who handles them.

Where to start: a first-version plan

  1. Write down where guests and orders come from today: walk-ins, messengers, a delivery platform, calls.
  2. Pick one main goal: be found and show the menu, take bookings, or take direct orders.
  3. Build a simple first version: menu, contacts, map and one feature for that goal.
  4. Agree who updates the menu and who handles orders and bookings.
  5. After a month or two, look at the numbers, and only then add delivery, loyalty or an app.

This approach is cheaper and safer: you pay for what guests actually use. Why starting small works is covered in prototype and first version. For reference, at NOVO a first version of a website or web service starts at $1,200, and a prototype that shows the main screens before development starts at $300.

In short

  • Guests open a cafe website for the menu, address, hours and a contact button. That is the foundation.
  • The online menu must match the room, and your staff should update it without a developer.
  • Start booking with a simple request form and add a floor plan later.
  • A delivery platform brings new guests, while your own website or bot lets regulars order directly.
  • Your own app makes sense with many regulars and a loyalty program you will actually run.
  • Launch one main feature, look at the results, and only then add the next.