Skip to main content

02Practice

Web & Mobile App Development

Product surfaces where performance is a feature rather than a post-launch ticket. We build web applications and native-quality mobile apps from a shared design system, tested against real devices and throttled networks — because the median user is not on a new laptop wired to fibre.

Typical duration
2–3 weeks to start · 1–2 weeks
Engagement models
Product build · Front-end rescue · Ongoing product team
Practice lead
Named at proposal stage
A smartphone lying face down on dark wood beside the edge of an aluminium keyboard.

01The problem

Performance is a product decision, usually made by accident

Applications get slow one reasonable decision at a time. A date library here, an analytics tag there, a list that renders every row because the dataset was small in staging. Nobody approves a four-second load; it accumulates. By the time it shows up in a support queue, the fix is spread across forty files and three teams.

Mobile makes the same problem sharper. A screen that feels fine on a test device can drop frames on a mid-range phone with a full storage volume and two years of OS updates behind it. So we treat device class and network conditions as requirements written down at the start — a budget per route, a device matrix, and CI that fails the build when the bundle crosses the line.

02What GatiQ delivers

The work itself

Every item below is something you can inspect, run or read at the end of the engagement.

  1. 01

    A design system, not a folder of components

    Type scale, spacing, colour tokens and interaction states defined once and consumed by both web and mobile. Components ship with their loading, empty, error and disabled states, so those states exist in production rather than in a future ticket.

  2. 02

    Performance budgets enforced in CI

    A JavaScript budget per route, an image policy, and Lighthouse plus real-user metrics reported on every pull request. Crossing the budget fails the build the same way a failing test does, which means regressions get caught by the person who caused them.

  3. 03

    Offline and flaky-network behaviour by design

    For field, warehouse and clinical apps we specify per screen what is cached, what queues, and how conflicts resolve when two people edit the same record on opposite sides of a dead zone. Then we test it with the radio genuinely switched off.

  4. 04

    Accessibility treated as correctness

    Semantic markup, visible focus, sensible tab order and labelled controls from the first component onward. We run keyboard-only and VoiceOver or TalkBack passes before release, and fix contrast at the token level so it stays fixed.

  5. 05

    A release process your team can run alone

    Signed builds, staged rollouts, feature flags, crash reporting wired to a named owner, and a rollback that takes one command. Shipping on a Thursday afternoon should be uneventful.

03Engagement model

Three ways to buy it

Product build
End to end, from design system to store listing. Typically twelve to twenty weeks for a first release with a defined feature set, including two rounds of usability testing with your actual users.
Front-end rescue
You have an application that works but is slow, inaccessible or unshippable. Four weeks of measurement and targeted fixes, with a written list of what we changed and what we deliberately left alone.
Ongoing product team
A standing squad — product engineer, mobile engineer, designer — iterating on a live product. Monthly rate, quarterly roadmap, weekly release cadence.

04Typical timeline

What the calendar looks like

  1. 01

    Foundations

    2–3 weeks

    Design tokens, component library, navigation shell, CI and budgets.

  2. 02

    Core journeys

    6–10 weeks

    The three flows that carry most of the traffic, built and usability tested.

  3. 03

    Hardening

    2–3 weeks

    Device matrix, accessibility audit, offline cases, crash-free rate targets.

  4. 04

    Release

    1–2 weeks

    Store review, staged rollout, monitoring handover.

Ranges, not dates. We narrow them at the end of discovery and tell you the same week if they move.

05Technologies

What we reach for

Defaults, not rules. If your team already runs something else competently, that is usually the better choice.

Web
Next.js, React, TypeScript, Tailwind CSS, TanStack Query
Mobile
React Native, Expo, Swift, Kotlin, Flutter
State and data
Zustand, RTK Query, SQLite, WatermelonDB
Quality
Playwright, Detox, Storybook, axe-core, Lighthouse CI
Operations
Sentry, Fastlane, EAS, Firebase App Distribution