There's an app for that

And there's a decent chance you haven't opened it since the day you downloaded it

That was Apple's slogan, and it was a very good one. It was also so successful that it quietly proved the opposite point.

Look at your own phone. A couple of hundred apps, probably. And a handful you would actually miss.

A few every day — messages, news, something social. A few now and then: banking, a booking system, the club or service you joined. And then the rest, which were all perfectly sensible decisions at the time.

That last part is the useful bit. Those apps mostly weren't bad. They worked. They just answered a question that didn't come up often enough to earn a place on the home screen.

One thing to be clear about before going further.

This is not an argument against building software. It is an argument about shape.

An app is one delivery mechanism among several. The same problem might be better served by a web app, an internal tool only your team touches, a piece of automation joining two systems you already pay for, an integration into software your customers sit in all day, or hardware with software running on it. I have worked on all of those. The question is rarely whether to build. It is which of those things the problem actually is.

An app happens to be the most expensive of them, the hardest to get installed, and the only one with a company standing between you and your customer taking a percentage. That doesn't make it wrong. It makes it a choice worth arguing for rather than assuming.

So when is the answer yes?

When you have found a real gap where software makes a genuine improvement or a genuine cost saving in a process that is already happening.

That last part is doing most of the work. Already happening. There are people, today, doing this thing manually or badly or expensively. You are not asking them to adopt a new behaviour. You are asking them to do something they already do, with less pain. That is a much shorter conversation than persuading someone to want something new.

If the process isn't already happening, you are not building a product. You are running a marketing campaign for a behaviour, and that is a far more expensive undertaking than anyone budgets for.

The one that comes up most: do I need an app to sell online?

No.

Before you build anything, and before you set up a Shopify store, try selling through Instagram or Facebook.

The commission feels like the objection. It's actually the price of the answer, and it's cheap. You find out very quickly what sells, at what price, to whom, and with what description. You learn which photograph works. You learn whether people ask a question before buying, and what the question is.

The moment you commit to a store, you have committed to a world of logistics, integrations, inventory, fulfilment and configuration in order to get exactly what you want. And you will still be paying commission somewhere in the stack.

Do the expensive thing after you know it works. Not before, on the assumption that it will.

Where an app genuinely beats a website

Some things really are better as apps. More responsive, more interactive, quicker to get the information across. Anything where the core action is short, repeated and frequent.

Booking is the obvious one. If you run a gym, a clinic, a studio, a consultancy where people book your time — a website makes them work for something they want to do quickly and often. An app doesn't.

Notice the pattern. Short, repeated, frequent. If your use case isn't all three, a good mobile-responsive site will probably serve you better and cost a fraction of the money.

The day-two test

Before anyone quotes you anything, write down two sentences.

What does someone do on day two?

Not day one. Day one is easy — day one is curiosity, and curiosity will get anything installed once.

And what does someone do on day thirty?

If you can answer both specifically, with an actual action rather than a benefit, you might have an app.

If day two is "explores the features" and day thirty is "continues to engage with the platform", an app is the wrong shape for it. Build the simpler thing — the site, the tool, the integration — and spend the difference on finding out what day two should have been.