Should I build a web app or a mobile app first?
By Bonaventure Ogeto, Founder and CEO, Mctaba Labs
Published
Build a web app first, designed for a phone screen. It opens from a link, needs no download, works on every device and can be changed the same day. Build a mobile app first only if the product must work offline, depends on device hardware, or will be opened many times a day. Consider WhatsApp or USSD before either.
"I am building an app" is how a first product usually gets described. Then the description turns out to be a booking system, a marketplace or a lending product, none of which needs to live in an app store.
"App" has become the word for "software product". That matters, because the two real options, a web app and a native mobile app, differ in cost, in speed and in how customers reach you. Picking the wrong one first makes it far more expensive to learn whether anyone wants the product.
Here is the verdict, and then the conditions under which it changes.
What is the short answer?
Build a web app first, and design it for a phone screen before you think about laptops.
A web app is a product that runs in the browser. The customer taps a link and it opens. There is nothing to download and nothing to update. Four things make it the right default for a first version:
- It opens from a link. Your first customers will come from WhatsApp messages, statuses and groups. A link takes them straight to the product. An app adds a store visit, a download and an install between the message and the first use, and each step is a chance to lose them.
- It works on every device. One build covers Android, iPhone and the laptop in the office.
- It changes the same day. A first version is wrong in ways you cannot predict. On the web, a fix is live when you publish it. A store app needs a new version, a review by the store, and customers who bother to update.
- It costs less to build and to change. One codebase, one team, no store accounts to set up.
"Mobile first" still applies. Expect your customers to open that link on a phone, often a modest Android phone on mobile data. Design for a small screen, keep pages light and test on a cheap handset.
What does a mobile app do that a web app cannot?
A native app can reach deeper into the phone, and it stays on the home screen. That is worth paying for when the product depends on it.
| A native app is better at | Why it matters |
|---|---|
| Working without a connection | The app and its data live on the phone and sync when the network returns |
| Device hardware | Reliable access to the camera for scanning, Bluetooth for printers and card readers, background location, NFC, biometrics |
| Being opened many times a day | An icon on the home screen and dependable notifications bring people back |
| Heavy or fast interfaces | Maps that move, video, drawing, anything that must feel instant |
| Store trust | Some customers trust an app from an official store more than a link, especially for money |
A web app can do part of this. Browsers can use the camera and location, and a web app can be installed to the home screen. But where the product depends on these things all day, native is the safer choice.
Which conditions flip the answer?
Five of them. If one clearly applies to your product, a web app first is probably wrong.
1. Offline use. Your users work where the network drops: field agents collecting data upcountry, riders moving between estates, a shop attendant whose data bundle has run out. If the job must continue with no signal, you need an app that stores work on the phone. Ask yourself what happens to your user when the bars disappear. If the answer is "they wait", the web is fine. If it is "they lose a sale", go native.
2. Device hardware. The product is built around the phone itself: scanning barcodes all day, printing receipts over Bluetooth, tracking a delivery in the background, reading a card. Test this early, because it is an easy condition to discover late.
3. Store trust. Your customers will hand over money or sensitive information and expect an official app. Lending, savings and insurance products often fall here. Being in a store signals that someone has checked you. Note the cost: each store requires a developer account, reviews what you submit and enforces its own policies. Google and Apple both publish their requirements, and financial products face extra rules.
4. Data cost. This one cuts both ways. A web page costs data every time it loads, so a heavy site punishes repeat users. A well-built app downloads once and then uses little. But the download itself is a barrier for someone who buys data in small bundles and has a full phone. For occasional use, a light web page wins. For daily use by the same people, an app can be cheaper for them over a month.
5. Frequency. How often will one person use the product? A few times a month, such as booking a service or paying fees, suits the web. Many times a day, such as a point-of-sale tool or a driver's job list, suits an app, because the user will want it one tap away.
Should I use WhatsApp or USSD instead of either?
Sometimes, yes. In Kenya the choice has four options, not two, and the cheapest of them is often the right first move.
WhatsApp. Your customers already have it open. A first version can be a WhatsApp Business number, a catalogue and a person replying. When volume grows, the WhatsApp Business Platform lets software send and receive messages, so orders, reminders and receipts can be automated inside a chat the customer already trusts. It suits conversational products: ordering, booking, support, simple questions and answers. It is weak for anything that needs a rich screen, such as comparing many items or filling a long form.
USSD. A short code such as *123# works on every phone, including feature phones, with no data and no download. It is the right choice when your customers are not on smartphones, or when the action is short and they should not need data: check a balance, pay a bill, confirm a delivery. Someone pays for each USSD session, either the customer or you, so ask the operator or aggregator how sessions are billed. The limits are real. Sessions are short and time out, menus are text only, and you need a short code from a mobile operator or an aggregator such as Africa's Talking, which carries setup and running costs.
A combination. A common pattern is a web app for the business owner with WhatsApp or USSD for the end customer. The shopkeeper manages stock in a browser while customers order in a chat.
There is a fuller comparison of these channels for payment products in USSD, app or WhatsApp.
How do I decide for my own product?
Answer these six questions about your first hundred customers. Do not answer for the customers you hope to have in three years.
| Question | If yes |
|---|---|
| Do most of them use feature phones, or avoid spending data? | Look at USSD first |
| Is the product mainly a conversation: ask, answer, confirm, pay? | Look at WhatsApp first |
| Must it work with no network? | Mobile app |
| Does it depend on scanning, printing, Bluetooth or background location? | Mobile app |
| Will one person open it several times a day? | Mobile app, or a web app now and an app soon |
| None of the above? | Web app |
Work from the top and stop at the first yes. If you reach the last row, build for the web and do not look back for six months.
A worked example with invented numbers. Say you are building a tool for small pharmacies to record sales. The attendant uses it fifty times a day, at a counter where the network is unreliable, and wants to scan barcodes. Three flip conditions apply, so the attendant's tool should be a mobile app. But the owner, who checks sales once each evening from home, needs only a web page. And the first test of the idea, before either is built, could be a shared spreadsheet and a daily WhatsApp summary.
What does this choice mean for the build?
Less than you might fear, if the product is built in the usual modern way.
Most of a product is not the screens. It is the server that holds the business rules, the database, the M-Pesa integration, the messages that go out. A sensible build puts all of that behind an interface that any front end can talk to. The web app is one client of it. A mobile app, a WhatsApp flow and a USSD menu can be added later as further clients of the same system.
So ask your developer one question before work begins: "If we add a mobile app next year, what do we keep?" The answer should be "almost everything except the screens". If it is "we would begin again", the design is wrong, whichever option you picked.
Choosing the web first lets you learn cheaply, in the place your customers already are, and build the app later, once you know which screens deserve it.
Key takeaways
- Default to a web app designed for a phone screen. It is the cheapest way to learn what customers want.
- Your first users will probably arrive from a link shared on WhatsApp, and a link opens a web page.
- Go mobile first when the product must work offline, needs device hardware, or is opened many times a day.
- If your customers use feature phones or will not spend data, look at USSD or WhatsApp before any app.
- A web app can be installed to the home screen and later wrapped or rebuilt as a store app, so the first choice is not final.
- Choose by where your customer already is, not by what looks most like a startup.
Frequently asked questions
- Do I need a mobile app to be taken seriously by customers or investors?
- No. Customers care whether the product solves their problem and whether paying feels safe. Investors care about usage and revenue. A web app with a hundred paying customers is more convincing than a store listing with a few hundred downloads and nobody active.
- What is a progressive web app?
- It is a web app built so that a phone can install it to the home screen, open it full screen and, with extra work, keep some features working without a connection. It is still one codebase served from your own website, with no store review. Support for features such as notifications differs between Android and iPhone, so ask your developer what will work on each.
- Can a web app take M-Pesa payments?
- Yes. An STK push sends the payment prompt to the customer's phone whatever they are using to browse, so the customer enters their PIN in the familiar M-Pesa screen. Payment providers such as Paystack and Flutterwave also offer hosted checkout pages that work in any browser.
- Should I build for Android or iPhone first if I do need an app?
- Build for the phone your paying customers carry. For a mass-market product in Kenya, expect that to be Android, but check with your own customers. If your customers are a narrow, higher-income group, check what they use before you decide. A cross-platform framework such as Flutter or React Native lets one team cover both, with some limits.
- How hard is it to move from a web app to a mobile app later?
- Easier than it sounds, if the first build was done sensibly. The server, the database, the payment integration and the business rules carry over. What gets rebuilt is the part the customer touches. By then you will know which screens matter, so the app is smaller.
Sources
- How to build a Progressive Web Application (PWA) with Next.js, Next.js documentation (Vercel). Checked on 19 November 2026.
- Get started with Play Console, Google, Play Console Help. Checked on 19 November 2026.
- Become a member, Apple Developer Program, Apple. Checked on 19 November 2026.
