Deciding before spending

Mobile app or website?

The test I run on the phone, in full — the one that can conclude you do not need an app.

The point is this: the question is not settled by three criteria of equal weight, but by one condition and then three criteria. The condition is frequency — an app lives on a home screen, and an icon opened once a year never pays for itself. If the frequency is there, a single criterion out of three is then enough: the phone itself, working offline, or coming back on its own. If the frequency is not there, no technical criterion saves the project and a website does the same job for less.

Answer four questions

No sign-up, nothing is sent anywhere. The answer appears at the bottom as you go.

Would your customers open it at least once a week?

Yes No, or I am not sure

The only question that can decide on its own. An icon opened once a year never pays itself back.
Does it need the phone itself?

Yes No

The camera used seriously, sensors, Face ID, Bluetooth, position tracked continuously.
Does it have to work with no network?

Yes No

The metro, a dead zone, a plane, a basement, a warehouse, a building site.
Do you have something useful to say between two visits?

Yes No

A notification that helps. A marketing nudge does not count — it is the first cause of uninstalls.

Answer the four questions: the result appears here. A website. The frequency is not there, and no technical criterion makes up for it. A website costs less, is fixed faster, and survives being left alone. It is the most common of the three outcomes, and I would tell you so on the phone in those words. Talk it through in 30 minutes The reasoning behind each question is just below. An installable web app. Your customers come back often, but nothing in your case needs the phone's own capabilities. A website that installs on the home screen does the job, with no store and no two apps to maintain. Its limits are real on iPhone: we look at whether they concern you. Talk it through in 30 minutes The reasoning behind each question is just below. A mobile app. The frequency is there and at least one criterion requires it: this is the case where an app is justified. The next step is not a quote, it is scoping — what it has to do, for whom, and what is deliberately left out of the first version. Talk it through in 30 minutes The reasoning behind each question is just below.

The condition that comes first: frequency

The criterion almost nobody puts first, and the only one that can kill the project on its own. An app takes physical space on a customer's screen: it asks to be installed, updated, and to survive the clear-out people do on their phones. The user pays that cost once at install and is repaid on every opening. If they do not open it often, they are never repaid.

It works when people come back

Every week, ideally several times. Ordering, checking, tracking, clocking in, booking, asking a question. Ed, the Totum group's app, holds on this point: you talk to your pharmacist when you have a question, and you have several a year, not one.

It fails when people come back rarely

Filing an insurance claim, booking a move, buying a mattress, making an administrative request. Nobody installs for that, and those who do uninstall straight after. On these uses a website is the right answer even when all three criteria below are true.

One case escapes the rule: an internal tool. If the users are your own teams rather than your customers, installing is no longer an obstacle — you deploy it — and frequency is decided differently. Say so early; it changes the whole reasoning.

The three criteria — one is enough

Once the frequency is there, any single one of these three justifies an app. They are capabilities a browser does not have, or only half has. All three at once is the clearest case; one is enough to decide.

It uses the phone, not just the screen The camera used seriously, sensors, Face ID, Bluetooth, position tracked continuously. A browser touches some of that, but with permission asked again on every visit, limits that differ from one browser to the next, and nothing that keeps running once the tab is closed. The other way round: taking a photo now and then justifies nothing. The browser does that perfectly well.
It works with no network The metro, a dead zone, a plane, a basement, a warehouse, a building site. An app keeps its data on the device, lets people work, and resyncs when the network comes back. A website stops on an error page. The other way round: if your users sit at a desk on wifi, this criterion does not count for you.
It comes back on its own A notification is the only way to re-open contact without buying attention again. A website waits to be visited, or pays to be visited. It is also the easiest criterion to waste: a useless notification is the first cause of uninstalls, and an uninstalled app does not come back. The other way round: if you have nothing useful to announce between two visits, this criterion is theoretical.

The three cases where a website wins

These are the three situations where I say no, and they come up often. A website is not the cheap version of an app: on these three uses, the app is the wrong tool.

People have to find youAn app is not found on Google. It is searched for by name, in a store, by someone who already knows it exists. If your problem is being found by people who do not know you yet, an app does not solve it — it moves it.
People buy rarelyA yearly purchase, a quote, a one-off booking. Installing costs your customer more than they will save by using it. A website opens in one tap from a link, with nothing to install.
It has to be shared and indexedA link is sent in a message, opens anywhere, gets referenced, gets quoted. Content locked inside an app exists for nobody but the people who already installed it — including for search engines and for the assistants now answering in your place.

What an app commits you to after launch

An app is never finished, and that is the part of the cost nobody shows you at the moment of deciding. iOS ships a major version every September, Android every year, and each one breaks something. There are two stores to keep, two reviews to pass, two sets of screenshots and descriptions to maintain. An app left alone for two years eventually stops launching at all.

A website, on the other hand, survives being left alone. It ages, it dates, it does not die. That is the difference to hold on to when you compare the two quotes: you are not comparing two purchases, you are comparing a purchase and a subscription.

What you do with your answer

Four outcomes, and only one of them leads to "we build an app". Here is what I say on the phone in each of the four.

The frequency is not there

A website, and I tell you on the call rather than in the third month. It is the most common of the four by some distance, and much the cheapest for you.

Frequency yes, no criterion

The middle case, and generally the worst advised. An installable web app — the user adds it to their home screen from the browser, with no store involved — often does the job. It has real limits, particularly on iOS, and we look together at whether they concern you.

Frequency yes, at least one criterion

An app is justified in your case. The next step is not a quote, it is scoping: what it has to do, for whom, and what is deliberately not in the first version.

You are still unsure

Starting with a website and building the app afterwards is a perfectly valid path, and often the right one. The website tells you who comes back and how often — which is to say it answers this page's question with your numbers instead of my criteria.

Three questions that always follow

It is a real third answer, and it is under-recommended because it earns less for whoever recommends it. A website can be installed on the home screen, work offline and launch full-screen with no store involved. What it does less well: notifications on iPhone remain limited, sensor access is partial, and you are not present in the stores — which matters if your customers look for you there. When your three criteria are all "no", that is almost always where to start.

Yes, and it is often the smarter order. The website answers this page's question with your own data: how many people come back, how often, to do what. You then reach the app decision with numbers instead of an intuition, and the scoping is half as long because you already know which features are actually used.

I build mobile apps — it has been my trade since 2013 and it is what I do well. Which means that when I tell you a website is enough, I lose the project, and that is exactly why the answer is worth something. If that is the conclusion, I tell you what to ask for and who to ask.

Your case fits none of these boxes?

That is the most common outcome. Thirty minutes is enough to settle it, and "no app" is an answer like any other.

Talk it through in 30 minutes The short test is on the home page

30 minutes, no slides, nothing to prepare.