Skip to content
All guides

Platform decisions

Should you start with a website or a mobile app?

Choose the approach after understanding the job and the device constraints. Compare access, offline behavior, device integration, and the work required to maintain each option.

Jump to the checklist

Look at the task, not only the screen size

A task performed on a phone does not automatically require an installed app. If a journey begins with a shared link, happens occasionally, or involves public content people need to discover, a responsive website may be a suitable starting point.

If people repeat a task every day, use specialist equipment, or work where connectivity is unreliable, describe those conditions first. Base the decision on the real journey and the devices the organization can support, rather than on the impression that an app is inherently more complete.

  • How people arrive

    A search result, a message containing a link, a managed device installation, or an app people find and install themselves.

  • How often they return

    A one-off transaction, regular repeated work, or a tool that stays open throughout a working session.

  • How the task finishes

    The result must be sent immediately, saved locally for later, or coordinated with another device.

Compare one tradeoff at a time

Use this comparison to ask better questions, not to declare one platform universally better. Actual effort depends on workflows, integrations, testing, and support. The label website or app is not enough to determine a budget.

Questions to compare before choosing a first release
AreaWebsite / web appInstalled app
First useOpen a URL; useful when sharing a linkPlan distribution and onboarding
Device featuresVerify support in target browsers and devicesVerify operating-system APIs and permissions
Offline workDesign storage and verify platform supportAlso design storage, retries, and reconciliation
UpdatesPlan web deployment and cache behaviorPlan releases, distribution, and older app versions
MaintenanceTest agreed browser and workflow coverageTest operating systems, devices, and update paths

Prototype the capabilities that determine the choice

For camera capture, barcode scanning, location, Bluetooth, or background work, do not infer support from a demo on a different device. Build a small proof of the critical task on the actual device, browser, and permission policy. Include what happens when permission is denied.

A progressive web app can add capabilities such as installation and offline operation. Support and behavior vary by platform, so distinguish essential requirements from optional improvements. Provide a workable alternative when an optional capability is unavailable.

Separate the interface decision from the data system

Web and mobile interfaces can share a backend, but the team still needs clear API contracts, authentication, permissions, and a plan for changing data formats. Having an API does not make every channel behave consistently by itself.

With several devices or offline work, identify the source of truth, rules for simultaneous edits, and how pending submissions are shown. Give records identifiers that support tracking and the chosen duplicate-prevention strategy.

  • Accounts and permissions

    Define sign-in, expiry, and revoked access consistently across every channel.

  • Data and status

    Distinguish locally saved work, work being sent, and records confirmed by the server.

  • Releases

    Plan for web and mobile clients using different API versions. Do not change data contracts without a migration path.

Choose a release that tests the core workflow

Some projects start with a web journey for customers and an internal screen for staff. An app follows when recurring work or a device capability justifies another channel. That sequence lets the team test the process and data before increasing the number of interfaces.

For work that depends on offline operation or specialist hardware, postponing the mobile experiment can uncover a fundamental constraint too late. Start by testing the requirement that carries the most uncertainty and has the greatest effect on the platform choice.

Record the decision and when to revisit it

Write a short decision record covering the core task, target devices, capabilities already tested, alternatives considered, and limitations the team accepts. Where evidence is missing, define an experiment instead of making a promise.

State what would trigger a review: supporting new equipment, introducing offline work, or changing distribution. Future maintainers can then understand why the first release had its particular boundaries.

A checklist for the next team conversation

  • The primary user task is clear
  • Entry points and frequency of use are understood
  • Target devices, browsers, and operating systems are listed
  • Critical capabilities have been tested on real devices
  • Offline and duplicate-submission behavior is defined
  • Interface work is separated from data and API work
  • The first-release choice has reasons and review conditions

Choose the channel that lets the core task succeed under real constraints. Add another channel when there is evidence that it improves the work.

Talk through your project