Mobile app or PWA for a D2C brand? Start with repeat-purchase frequency, weigh push notifications against the cost of two platforms, score seven questions, and run a pilot before committing.
At some point every D2C founder is asked, usually by an investor or a well-meaning friend, 'why don't you have an app?' The honest answer is often 'because we sell something people buy twice a year and nobody would keep it installed'. But the question deserves a real framework, because for some brands an app is the single best retention tool they will ever build, and for others a progressive web app gets most of the value for a fraction of the cost.
This post is that framework. It is written for a founder or e-commerce manager deciding whether to build, what to build, and how to find out cheaply before committing.
Start with the question behind the question
'Should we build an app' is never really the question. The underlying goals are usually one or more of: increase repeat purchase, reach customers directly without paying a platform each time, feel like a serious brand, or satisfy someone who thinks an app is a milestone. Only the first two are worth engineering for. Write down which of them you actually mean, because a PWA serves the first two well and a native app serves all four, at very different costs.
Repeat-purchase frequency is the first filter
An app lives on a phone that already has too many apps on it. It survives only if the customer has a reason to open it more often than the phone has a reason to suggest deleting it. So the first number to look at is how often your typical customer buys.
- Weekly to monthly (groceries, supplements, coffee, pet food, personal care refills): an app is plausible. The habit already exists; the app makes it shorter.
- Every few months (fashion, cosmetics, home goods): borderline. An app works if it offers something beyond the shop, such as content, a loyalty balance, or early access that the customer values between purchases.
- Once or twice a year (jewellery, furniture, electronics, luxury): a native app is almost never justified by purchase frequency alone. The retention job is better done by email, WhatsApp and a fast mobile website.
Pull the real figure from Shopify: orders per customer over the last twelve months, split by cohort. If the median customer has ordered once, you are building an app for a minority of your base, and the maths has to work on that minority alone.
What a PWA can do now
A progressive web app is your website, built to standards that let a phone treat it like an app: it can be added to the home screen with its own icon, launch full-screen without browser chrome, cache a shell so it opens instantly and works partly offline, and, on Android and on recent versions of iOS for web apps added to the home screen, receive push notifications.
- One codebase, which is the website you already maintain. Every improvement to the site is an improvement to the app.
- No store listing, no review process, no waiting for approval before a festive campaign goes live.
- No forced update. Everyone is on the latest version the moment you deploy.
- Push notifications with permission, on the platforms above, which for a D2C brand is most of what an app was wanted for.
- It is discoverable through search, which a native app is not.
Its limits are real. The install is less familiar to customers than an app store download, push on iOS depends on the customer having added the site to the home screen, and it cannot use the deepest device features or appear in the app stores, which some customers treat as a signal of legitimacy.
What a native app genuinely buys you
- Presence in the app stores, and the credibility that comes with a listing and reviews.
- Reliable push notifications on every platform, with rich content and no dependence on the customer having added anything to a home screen.
- Device features: biometric login, camera-based try-on for jewellery or eyewear, contacts, NFC, background sync, native sharing.
- A smoother feel for heavy interactions, such as a large image-led catalogue or a configurator, that a browser can approximate but not quite match.
- A stronger signal of intent. Customers who install a native app have, in our experience, told you something about themselves that customers who bookmark a site have not, and they behave accordingly.
The question is whether any of those is worth the cost, and the cost is not the build.
The real cost is two platforms, forever
A native app is at least two products, iOS and Android, plus the website you still have to run. Cross-platform frameworks such as React Native with Expo let one team write most of the code once, and we use them, but they do not remove the second platform; they hide it until something platform-specific breaks.
- Every feature ships three times: web, iOS, Android. Every bug is potentially three bugs.
- Two store review processes, two sets of policies that change, two sets of certificates and signing keys that expire at inconvenient moments.
- Operating system updates that arrive on a schedule you do not control and occasionally break something that was working.
- Crash monitoring, analytics and push infrastructure that need to be wired for both platforms and reconciled with the website's.
- Support: 'the app is not working' becomes a meaningful share of tickets, and every one needs device, OS version and app version to diagnose.
- In India specifically, most customers are on Android, with wide variation in device quality and storage. App size and performance on cheaper phones matter more than they do in markets where the iPhone dominates, and an app that is uninstalled to free up space is worse than a site that was never installed.
Budget the maintenance as a permanent line, not a launch cost. If that line is uncomfortable, the decision is made.
A decision framework you can run in an afternoon
Score each statement one point if true. It is deliberately blunt.
- Our median customer buys more than four times a year.
- We have a specific reason a customer would open the app between purchases, and we can name it.
- We need a device feature (camera, biometrics, NFC, background sync) that a browser cannot provide.
- Push notifications are central to our retention plan, and we have content worth sending at least weekly.
- We have, or will fund, a developer responsible for the app for at least three years.
- Our customers, in our own data, ask for an app or install competitor apps.
- Our mobile website is already fast and converts well, so an app is an addition rather than a rescue.
Six or seven: build the native app, probably cross-platform. Three to five: build the PWA, run the pilot below, and revisit in a year. Two or fewer: fix the mobile site and invest in WhatsApp and email; an app would be an expensive way to avoid that work.
Run a small pilot before deciding
The cheapest way to learn whether your customers want an app is to give them a lighter one and watch.
- Make the existing store installable as a PWA: a manifest, a service worker for the shell, and a proper icon. On a Shopify theme this is a modest piece of work; on a headless Next.js storefront it is largely built in.
- Add an install prompt at a sensible moment, such as after a second visit or after an order, not on first landing.
- Add push with a clear opt-in and a promise you will keep, such as restock alerts or a weekly drop, and send only that.
- Run it for one full purchase cycle for your category, which for most brands is eight to twelve weeks, and through at least one campaign, ideally a festive one, when intent is highest.
- Measure four things: what share of visitors install, what share of installers allow push, the repeat-order rate of installers against a matched group of non-installers, and how many uninstall or revoke push.
If installers repeat meaningfully more and stay installed, you have found the minority the native app would be built for, and you know its size. If they do not, you have saved yourself a year of maintenance, and the PWA is still a better mobile site than you had before.
Where we stand
We build both. LeelaRam Technologies has a mobile application for iOS and Android in development, built cross-platform with the same verification discipline as our Shopify work, so we are not arguing against apps. We are arguing for deciding with your own purchase data and a pilot rather than with a feeling. In our experience most D2C brands that run the pilot end up with a very good PWA and a WhatsApp programme, a few build the native app with confidence, and none of them regret having checked first.
