TL;DR: Build the first version where your users already are when they need the product. If the product is used at a desk, inside a workflow, by a team or by a buyer who signs up through a sales conversation, start with a web app. If it is used in the moment, away from a desk, and depends on the camera, microphone, location, sensors, push notifications or offline use, start with a mobile app. A web app is usually cheaper and faster to change, a mobile app is usually stickier and closer to the hardware, and the real cost of getting it wrong is not the build but the months spent learning the wrong lessons from the wrong audience. Pick one surface, prove the loop, then earn the second.
"Should we build a web app or a mobile app?" is one of the first questions founders and product owners bring to us, and it is usually asked as a technology question. It is not one. It is a question about where your users are standing, what they are holding and what they are trying to finish when they reach for your product. Answer that and the technology mostly picks itself.
This guide is for anyone scoping a new product or a new line of business software and trying to decide which surface to build first. It covers the questions that settle it, what each choice really costs over time, the hybrid options in between and the mistakes we see most often.
What each option actually means
Three terms get blurred together, so it helps to pin them down.
- A web app runs in the browser on any device. It can be fully responsive, so it works on a phone screen, but it lives at a URL, updates the moment you deploy and needs no store approval.
- A mobile app is installed from the App Store or Google Play. It can be built natively or with a cross-platform framework, which is a separate decision we cover in native vs cross-platform app development.
- A progressive web app (PWA) is a web app that can be added to the home screen, cache content and, on most platforms, send notifications. It narrows the gap but does not close it, and browser support for its features still varies by platform.
The choice here is which of these your first version should be. Not which ones you will ever have.
Five questions that settle it
1. Where is the user when they need the product?
This is the question that decides most cases on its own. If the moment of need happens at a desk, in a browser tab next to email and spreadsheets, a web app meets people where they already are. Back-office tools, dashboards, admin portals, accounting and operations software almost always start here. A product like Fyuel, an accounting platform that brings customers, suppliers, tankers, ledgers, banks and reports into one real-time system, is used by people doing finance and operations work, and that work happens where the documents and the keyboard are.
If the moment of need happens in a lecture hall, on a roadside, at the gym or on a job site, a phone is what people have in their hand. LectureNotes AI records lectures and turns them into summaries and outlines, so it has to be on the device sitting on the desk during the lecture. Covoisino, a ride-sharing app built to make hitchhiking safer, uses QR verification between people standing next to each other. Neither product makes sense as something you open later on a laptop.
2. Does the product need the device itself?
Some capabilities are much stronger, or only available, in an installed app:
- Continuous or background use of the microphone, camera, GPS or motion sensors
- Reliable push notifications, especially on iOS
- Bluetooth, NFC and connected hardware
- Real offline use with local storage and later sync
- Health, payment and other platform integrations
The web has gained a lot of these over the years, but support is uneven and the edges are exactly where products break. If your core loop depends on one of them, that is a strong vote for mobile. If none of them appear in your core loop, the web is usually enough for version one. Offline needs deserve their own look, which we cover in offline-first app development.
3. Who buys, and how do they find you?
Products sold to businesses are usually discovered through a sales conversation, a referral or a search, and the first thing the buyer wants is to try it without installing anything. A login link they can forward to a colleague does more for a B2B pipeline than an app store listing. This is also where requirements like single sign-on, roles and admin controls show up, and those are easier to build and demo on the web. Enterprise-ready SaaS goes deeper on what those buyers ask for.
Consumer products often live or die on habit, and the home screen icon, the notification and the store listing are part of how habits form. If the plan depends on daily use by individuals, mobile tends to win.
4. How fast do you need to change it?
Early products change constantly. A web app ships a fix in minutes and every user has it immediately. A mobile release goes through store review, and users update on their own schedule, so you will support several versions at once and need a plan for forcing upgrades when an old version breaks. That is manageable, and we walk through it in the app store release process, but it slows the learn-and-change loop that matters most before product-market fit.
If you expect to rewrite onboarding five times in the first two months, that is a real argument for starting on the web, even for a product that eventually lives on phones.
5. How will you charge?
On iOS and Android, digital subscriptions and features used inside the app generally go through the stores' in-app purchase systems, which take a commission and add receipt validation and entitlement work. The rules on linking to outside payment have been changing by region, so check them for each market. On the web, you choose your own payment provider and pricing page. If the business model depends on annual contracts, invoices or usage billing, the web is the simpler starting point. If it is a consumer subscription, plan the billing layer carefully either way; our guide to subscription app development covers it.
What each choice costs over time
The first build is only part of the bill. A few differences show up in the running costs.
Web first
- One codebase, one deployment and one set of users on the latest version.
- Responsive design work so the product is usable on phones, even if it is not built for them.
- Browser testing across the main engines.
- No store review, no store commission on your own payments.
Mobile first
- Store accounts, review cycles, screenshots, listings and privacy disclosures for each platform.
- A device and OS matrix to test against, and yearly OS releases that can break things.
- Several app versions in the wild at once, which means your backend has to stay compatible with older clients.
- Usually a small web presence anyway, for the landing page, password resets, account deletion and legal pages.
Building both in version one is closer to two products sharing a backend than one product with two screens. For typical price bands and what drives them, see how much an MVP costs.
The options in between
The choice is rarely all or nothing, and a few middle paths are worth knowing.
- Responsive web app first, mobile later. The most common path for B2B and workflow products. You learn fast on the web, then build a focused mobile app for the jobs that genuinely happen away from the desk, such as approvals, field capture or notifications.
- Mobile app with a web companion. Common for consumer products. The app is the product; the web side handles signup, billing management, sharing links and anything that needs a big screen.
- PWA as a bridge. Useful when you want a home screen presence without store overhead, for internal tools or for testing whether people will actually come back daily before you invest in a full app.
- Web-based AR and 3D. For immersive experiences that need instant reach, such as marketing, retail or events, web technologies can skip the install entirely. We cover the trade-offs in WebAR without an app.
Whatever you choose, design the backend and API as if a second client is coming. Clean authentication, an API that is not shaped around one screen, and account-level data instead of device-level data make the second surface an addition rather than a rewrite.
Mistakes we see most often
- Building both at once to look complete. It doubles the surface you have to keep stable while you are still learning whether the core loop works.
- Choosing mobile because "everyone is on their phone." They are, but not for every job. Nobody reconciles a supplier ledger on a phone if they can avoid it.
- Choosing web to save money on a product that needs the device. A camera-first or location-first product squeezed into a browser often feels broken, and early users judge the idea by that experience.
- Treating responsive as an afterthought. Even a desk-first web app gets opened on a phone from an email link. That first impression matters.
- Skipping analytics on whichever surface you pick. You cannot tell whether the choice was right without measuring the loop. Product analytics for an MVP covers the minimum worth setting up.
How AI changes the decision
AI coding agents, used under senior review, have made the second surface cheaper to build than it used to be. Translating a working web flow into a mobile screen, writing the per-platform glue and generating tests are the kind of mechanical work they handle well. That is a big part of how we ship in days rather than months, described in our ship-in-days playbook.
What AI does not change is the cost of maintaining two surfaces and two release processes, or the cost of learning from the wrong audience. It makes "earn the second platform" faster to act on. It does not make "build everything at once" a better idea.
Where we fit
We build both, and our own products show why the answer differs by product. LectureNotes AI and Lifemaxxing AI live on mobile because that is where their users are when they need them. VoiceDash, a white-label client portal for voice AI agencies, is a branded web dashboard on the agency's own domain, because agencies and their clients log in from a browser to check results. Fyuel brings an entire fuel-trade back office into one real-time system. The choice in each case came from the user's moment of need, not from a house preference.
If you are writing up a new product, our guide to writing a software project brief includes the questions in this article, and answering them before you talk to any vendor will save you a round of rescoping.
Quick checklist
- Describe the exact moment a user reaches for the product: where they are and what they are holding.
- List the device capabilities your core loop needs, if any.
- Decide how buyers discover and try the product.
- Estimate how often you will change the core flows in the first three months.
- Choose your payment model and check what each surface requires.
- Pick one surface for version one and write down what would justify the second.
- Design the backend and API for more than one client from day one.
Weighing web against mobile for a new product? Book a demo and we will work through where your users are, what the product needs from the device and which surface gets you real learning fastest. See our work: LectureNotes AI, Fyuel, Covoisino, VoiceDash and more, shipped in days, not months.