An app gets one minute. Somebody installed it standing at your counter or lying on a sofa, they open it, and within about sixty seconds they will either do the thing they came for or put the phone down — and the ones who put the phone down almost never come back to it. Everything in this topic is about that minute, because the retention numbers from topic 1 are decided here rather than later.
🚪 Why apps get deleted
The reasons are boringly consistent across every restaurant app ever built, and four of them cause almost all of it.
- A registration wall before any value. The app opens and demands an email, a password and a phone number before showing so much as a menu. This is the single biggest killer: the customer has not yet received anything and is being asked to pay in effort. Let people look first, and ask for an account at the moment it becomes necessary — when they order.
- Permissions asked up front. Location, notifications and contacts, all requested in the first ten seconds by an app that has not yet earned any of them. Ask for notifications after the first order, and explain in one line what you will send.
- It is slow. A phone app that takes six seconds to show a menu is competing against a browser that takes three, and losing.
- The menu is out of date. The customer orders something you stopped serving in March, or sees a price that is no longer true. This is the failure that turns an app from useless into actively harmful, and it is entirely a maintenance problem — the same one topic 3 warned you to name a person for.
Notice that not one of those is about design taste. The catalogue outline for this subject had a whole chapter on visual interface; in practice, what deletes an app is friction and neglect, not the colour palette.
⏱️ What must be possible in sixty seconds
Write these down as a test and try them yourself on the finished thing, with a stopwatch, on mobile data.
See the whole menu, with prices, without logging in. This is the most common reason anybody opens a restaurant app at all, and it must be the fastest path in the product.
Reorder the usual in two taps. For a returning customer this is the one thing an app does better than a browser, and it is the reason to have one at all — the previous topic suggested building version one around exactly this.
See the points balance and what it is worth, if you run a programme. Not the rules, not the terms: the number, and the sentence that says what it buys.
Find hours, address and a call button. The same four things the mobile site needed in the previous topic, for the same reason.
If any of those takes more than a few seconds or requires an account, fix that before adding a single new feature. And take out anything that is not on the list: a home screen with four clear things beats one with eleven, especially on a phone held in one hand on a street.
🖥️ The four screens that matter
Whatever the provider offers, the product a small restaurant needs is four screens, and everything else is decoration.
The menu, which is the reason most people open it, kept identical to the real menu and updated by the person named in topic 3. The order screen, with the previous order visible at the top so that repeating it is trivial. The account, which is really just the points balance, the saved address and the saved card. And the restaurant: hours, address with a navigate button, phone with a call button, and a message option that a human answers.
Two details that punch above their weight. Make the app usable while logged out for everything except paying — a menu that requires a login is a menu nobody reads. And make sure the app knows when you are closed and says so on the home screen, in words: nothing damages trust faster than an app that cheerfully takes an order at eleven at night for a kitchen that closed at ten.
🧪 The hallway test
Before launch, do the cheapest and most brutal piece of testing there is. Hand the phone to somebody who has never seen the app — a friend, a supplier, a new member of staff — and say one sentence: "order what you usually order." Then say nothing at all. Do not explain, do not point, and do not defend it.
Watch where they hesitate. Every hesitation is a real customer who will not ask you for help, they will just leave. Three people is enough to find almost everything, and the same test repeated after each change costs ten minutes.
Then do the version that matters most and that nobody does: try it yourself in the situation your customer is actually in. Standing outside, in the sun, on mobile data, with one hand, in a hurry, having not opened the app for six weeks and having forgotten the password. That is the real product. The one you have been looking at on a fast connection with your account already logged in is a demonstration.