Why You Need a Mobile App — and When You Honestly Don't

I build mobile apps for a living, which makes me exactly the wrong person to ask whether you need one — the way a barber is the wrong person to ask whether you need a haircut. So I decided a while ago that the only way to be trustworthy on this question is to be genuinely willing to say no. And I do say it, more often than you would guess.

Here is the filter I actually use when someone brings me an idea — the same one, whether it kills the project or lands me the work.

The one question that decides most of it

Will the same person use this more than once a month? That is nearly the whole test. Apps live on home screens, and home screens are for habits. Food ordering, gym bookings, banking, learning, team tools — repeat loops, all of them. If your customer interacts with you once a year, they will not install your app, and honestly they shouldn't.

When I built TapMeal, the entire bet was on this: ordering food is something people do weekly, and the person reordering their usual biryani should never have to think. That is what an app is for. The same product as a website-only experience would be a worse product — not slightly worse, categorically worse.

Five signals you actually need an app

• Your product has a repeat loop. Order again, book again, check again. The reorder-my-usual button is the strongest argument in mobile.
• You need to reach users, not wait for them. Push notifications are the only re-engagement channel you fully own. If "we have no way to bring customers back" sounds familiar, this is your fix.
• You need the hardware. Camera, GPS, offline storage, biometrics, Bluetooth. My location app Nearby simply could not exist as a website — the phone's sensors are the product.
• Speed is the experience. A saved session, instant open, cached data. Every second of friction between intent and action costs conversions, and apps hold the speed advantage where it counts.
• Trust matters at a glance. Fair or not, an app signals permanence. For businesses handling people's money or data, being on the store is part of looking real.

When a website honestly wins

The mirror-image cases, so you can self-diagnose. If most of your traffic arrives from search — people comparing, researching, buying once — the browser is your home turf; an app would be a second product to maintain while your actual storefront starves. Content businesses, portfolio sites, one-off services, event pages: websites, all of them. A good responsive site with fast checkout beats a mediocre app every day of the week, and half the "we need an app" conversations I have are really "our website is slow" conversations in disguise.

The part people underestimate: an app is a product

The clients who regret their apps are rarely the ones who overpaid. They are the ones who treated launch as the finish line. An app needs updates when the OS changes, fixes when a library breaks, and attention when reviews come in. If nobody in your business will own it after launch, that — more than budget — is the sign to wait. (I keep a full breakdown of what apps really cost, maintenance included, if you want the numbers.)

And sometimes, the app is just... delightful

One more thing the frameworks-and-budgets talk misses: native apps can simply feel good in a way that builds affection for a brand. I post UI experiments constantly for exactly this reason — the details are the product. Case in point, a pizza-ordering flow I built just to chase that feeling:

No website version of that pizza builder would make anyone smile. Sometimes that is the honest answer to "why an app" — because done well, it is an experience people enjoy returning to, and returning is the whole game.

Where you probably land

If you read the five signals nodding, you are an app business — and 2026 has only sharpened that case. If you read them shrugging, spend the money making your website excellent instead, and revisit when the repeat loop appears. And if you are genuinely torn, describe your product to me — I will tell you which one I would build, even when the answer pays me nothing.