Web app vs mobile app for an MVP

Anchor the choice in where the user completes the job

Describe the moment of use: location, device, connectivity, session length, input method, privacy and interruption risk. A back-office workflow used at a desk points to a different surface from a field task involving a camera, sensor, intermittent network or one-handed action. Do not infer mobile need merely because target users own phones, or web suitability merely because a browser can display the screens.

Define the smallest complete outcome and which parts of the channel are hypotheses. If the question is whether users understand and value a workflow, a responsive web path may provide enough evidence. If the question is whether the product works under mobile permissions, background limits, offline states or store discovery, a web simulation may answer the wrong question. The MVP surface must match the uncertainty being tested.

Account for release systems and platform-specific capacity

Web teams still need compatibility checks, accessibility, secure deployment, monitoring and recovery. Mobile teams add signing identities, builds, device and operating-system coverage, permissions, store metadata, review rules and release tracks. These are not reasons to avoid mobile. They are work and ownership boundaries that belong in the estimate when the product actually depends on them.

Name who can test real devices, diagnose backend and client failures, prepare store submissions and respond when platform requirements change. A cross-platform framework can share some implementation, but it does not remove product decisions or platform obligations. Likewise, a progressive web app can use install-related standards yet still differ from native apps in capabilities and distribution. Verify the exact current browser and platform support.

Own distribution identities and shared product services

The company should control domain, DNS, hosting, analytics and backend accounts for a web app, plus developer-program and store identities for mobile where applicable. Keep shared data models and APIs documented independently of one client surface. This makes a later channel addition or replacement a planned product change instead of an emergency migration from an account owned by a supplier or individual.

Keep source control, domains, stores, analytics, cloud organisations and production credentials under explicit organisational ownership. Grant scoped, revocable access and rehearse handover before the final invoice. The supplier label matters less than whether another authorised person can inspect, deploy, recover and continue the product.

Test the channel-specific failure states

For web, test supported browsers, responsive layouts, keyboard use, network interruption, authentication and recovery on the intended devices. For mobile, add permissions denied, background and foreground transitions, offline behavior, upgrades, deep links and store-ready builds. Measure whether the intended user can complete the core outcome. A successful build or store approval is implementation evidence, not proof of demand or retention.

Define acceptance as observable evidence. Examples include a tested user journey, an approved build, reproducible checks, an incident runbook and a verified handover. None of those proves retention, revenue or product-market fit. Keep delivery proof, external provider approval and business results as separate claims.

Sequence channels only when each stage answers a question

Choose web first when it reaches the actual user context and provides a relevant learning loop without hiding a decisive native risk. Choose mobile first when phone-specific behavior is part of the value proposition and cannot be responsibly deferred. Choose both only when the first release genuinely needs both and the team can support two clients, not because broader surface area feels safer.

Write a review trigger for adding or changing the channel: observed desktop friction, required device capability, repeated mobile browser limitations, store distribution demand or support evidence from real use. Preserve shared contracts and acceptance journeys so another client can be introduced deliberately. Neither channel choice guarantees acquisition, approval, usage or commercial performance.

Decision criteria

Use the same questions for every option before choosing.