Sooner or later someone says it in a meeting: "Our customers are asking for an app." It sounds like a spec. It rarely is one. Before you ask a custom software developer for a quote, or accept one, it pays to find out what the customer pictures when they say "app", because the answer changes the project completely.
There are three ways to respond: a native app published on the App Store and Google Play, a web app the customer installs from their browser, or a better website. What has changed is what a web app can do on an iPhone, and most comparisons online haven't caught up.
The short answer, for the busy
When a customer asks for an "app", they almost always want four things: an icon on their home screen, notifications, staying logged in, and fast access, sometimes offline. An installable web app now delivers all four on both platforms. Since iOS 26, any website added to an iPhone's home screen opens as an app, and web notifications have worked there since iOS 16.4, provided the user has added the site first. On Android, Chrome offers installation on its own. The app stores are still needed to be discovered by strangers, for device features Safari lacks, such as Bluetooth or NFC, and to sell digital content. As of October 2026, they cost US$99 a year at Apple and US$25 once at Google, a review on every update, commissions of 15 to 30% on digital sales, and two apps to bring up to the platforms' new requirements every year.
What "an app" means once you ask
Ask what the app would do that the website doesn't. The answers almost always fall into four needs.
The icon. Getting back to the business in one tap, without typing the address or going through Google.
Notifications. Being told the order is ready, the appointment is tomorrow, there are two spots left in Saturday's class.
Staying logged in. Not retyping a password every time they open their file.
Speed, and sometimes offline use. Seeing the information right away, even in a basement with no signal.
Notice what's missing: nobody asks to go through an app store. The store is the solution people picture, and none of the four needs requires it.
Two of these needs don't even call for an app. Staying logged in depends on how long a session the site grants. Speed is often a complaint about the current site, one that an app works around at great expense without fixing. Site speed can be measured and fixed. That's the third answer, and sometimes the right one.
What an installed web app does today
Adding a website to an iPhone's home screen has been possible since 2008. For 17 years, the icon only opened the site as an app if the developer had configured it to. Since iOS 26, released in September 2025, every site added to the home screen opens as a web app by default, separate from Safari. The WebKit team, which builds Safari's engine, put it plainly: there are now zero requirements for "installability" in Safari.
Notifications came earlier, with iOS 16.4 in 2023. They appear like any other: on the lock screen, in Notification Center and on a paired Apple Watch, through the same service native apps use. You don't need to be a member of Apple's developer program.
Two conditions matter. The permission request must follow a user action, such as a tap on a "Notify me" button. And on iPhone, notifications only work for a site added to the home screen. The site can't trigger that step itself: Safari gives it no install button, and the user has to go through the Share menu. That is the web app's real weak spot on iPhone. You have to teach the gesture, for example right after a first booking.
On Android, Chrome does that work. When the site is served over HTTPS, declares a name and icons, and the user has tapped the page and viewed it for at least 30 seconds, Chrome offers to install it. The site can also show its own "Install" button. Notifications work in Chrome even without installing.
The real limit is elsewhere. According to MDN's compatibility data, Safari supports neither Web Bluetooth, nor NFC reading, nor background sync.
What only the app stores give you
The stores give you three things: device features the web can't reach on iPhone, selling digital content through their payment system, and discovery. Someone who has never heard of you searches "yoga booking" in the App Store and finds you. For an app meant for your existing customers, that last advantage is worth close to nothing: they already know your name.
If none of the three applies, there's the temptation to wrap the existing site in an app. Apple saw that coming. Its guideline 4.2 on minimum functionality asks that an app include features, content and UI that "elevate it beyond a repackaged website", or it may not be accepted. Guideline 4.2.6 rejects apps created from a commercial template unless the content provider submits them directly. The "cheap app" that slips your website into a shell is exactly what Apple's review is built to filter out.
What the app stores cost
As of October 2026, Apple's developer program costs US$99 a year, billed in local currency where available. Google charges US$25, once, to open a Google Play developer account.
Apple's agreement sets a 30% commission on sales, and 15% on subscription renewals after a year. Its Small Business Program cuts the rate to 15% as long as net proceeds stay under US$1 million a year. Google, for Canada, charges 15% on the first US$1 million of annual earnings for developers enrolled in that tier, 30% above it, and 15% on subscriptions. A new fee grid has applied in the United States, the European Economic Area and the United Kingdom since June 30, 2026, and in Australia and Japan since September 30. For Canada, Google gives no date yet.
For most small businesses, these commissions don't apply. Apple requires that physical goods and services consumed outside the app be paid for by other means, such as Apple Pay or a credit card. Google exempts physical goods from its billing, along with services such as cleaning, gym memberships and live-event tickets. A garage, a clinic or a restaurant pays the store no commission on what it sells through its app.
The cost that remains is review. Every new version goes through Apple, which says it typically reviews at least 50% of submissions in under 24 hours and 90% in under 48. Google warns that for certain accounts, review can take up to seven days. On a website, a fix is live within minutes.
The bill that comes back every year
People think of an app as something you build once. The stores treat it as a subscription to their updates.
Every year, Apple and Google raise the minimum. Since April 28, 2026, every app uploaded to Apple must be built with Xcode 26 and the iOS 26 SDK. Since August 31, 2026, Google requires new apps and updates to target Android 16, and existing apps to target at least Android 15 to remain available to new users whose phones run a newer version.
So an app left untouched for two years doesn't stay frozen: to fix a bug, you first have to bring it up to standard.
A native app means two codebases (or a cross-platform framework that brings them closer), two accounts, two review queues and two calendars of requirements. The website still exists too. That's three surfaces to keep consistent, and the third is the one the budget forgets.
Three cases we see often
The client portal. This is the best candidate for an installable web app. Its users are already customers and log in to follow a file or see an invoice. A "your document is ready" notification and an icon cover the essentials. We've described what goes into a client portal, and how to connect it to your software through an API integration.
Booking. A customer books a few times a year. Making them download an app before booking adds a step your competitors don't impose. Booking should live on the website, where people already look for it by your name. Email or text reminders also reach people who never installed anything.
The loyalty card. This is where an app is most often a mistake. Apple Wallet and Google Wallet accept loyalty cards distributed through a link on your website or by email, with no app. The card lives in the wallet the customer already opens at the till, and it can show an up-to-date balance. The rules that govern points in Quebec are in our article on gift cards and loyalty programs.
A native app makes sense when the app is the product: you sell digital content to people who will discover you in the store, or your service has to talk to a Bluetooth device or read an NFC chip on iPhone. It then deserves a budget and an owner for years, and the tests in our article on custom software versus SaaS apply.
Where to start
Before any quote, ask your customers the four-needs question. If the list comes down to an icon, notifications, a login that sticks and speed, start with an installable web app and measure how many customers install it. If one need holds out (Bluetooth, NFC, digital sales, discovery by strangers), you'll have a specific reason to go into the stores, and a shorter spec.
At Peich, we build portals and web apps like these, and we'll tell you when a better website is enough.
→ Let's talk about the app your customers are asking for
Written byXavier Peich