Skip to content
Senior web & app development · 20 capabilities
oakstone.digital
Apps & software / capability

QA & testing
before your customers volunteer.

Manual and automated coverage that finds the failure before your customers volunteer to.

This is for you if
  • Releases regularly break something unrelated
  • Testing is one person clicking through from memory
  • You have automated tests nobody trusts
A sharper way forward

The work should
change the numbers.

Tools we reach for
PlaywrightVitestAxeLighthouse CIk6BrowserStackGitHub Actions

Every team already has a QA process. Usually it is called "the founder clicks around before release", and it works until the product has more paths than one person can remember.

Useful QA is risk-based. Instead of chasing a coverage percentage, we identify the flows where failure costs money or trust — checkout, authentication, billing, data export — and defend those hardest with automation that runs on every commit.

The goal is a suite your team trusts enough to release on a Friday, plus a clear picture of what is deliberately not covered.

What's included

What the engagement
actually contains.

01

Risk-based test strategy

Critical journeys identified with you, then coverage targeted where failure actually hurts.

02

Automated regression suite

End-to-end tests running in CI on every pull request, with flake treated as a defect.

03

Cross-browser and device coverage

Real-device and browser-matrix testing for the combinations your analytics say matter.

04

Accessibility testing

Automated axe checks in the pipeline plus manual keyboard and screen-reader passes.

05

Performance and load testing

Budgets enforced in CI and load profiles run against the traffic you actually expect.

06

Defect triage process

Severity definitions, reproduction standards and a queue your team can work from.

What you get
  • Test strategy with named coverage gaps
  • Automated end-to-end suite in your CI
  • Cross-browser and device test report
  • Accessibility audit against WCAG 2.2 AA
  • Performance budgets and load test results
  • Defect backlog with severities and reproductions
Signals it is time
  • Releases regularly break something unrelated
  • Testing is one person clicking through from memory
  • You have automated tests nobody trusts
  • Bugs are found by customers more often than by CI
Common mistakes we fix

What usually went wrong
before we were called.

  • 01Chasing a coverage percentage instead of defending the journeys that carry money
  • 02Tolerating flaky tests until nobody trusts the suite
  • 03Testing only the happy path on the browsers the team happens to own
  • 04Treating accessibility as an automated check and nothing more
How we measure it
  • Escaped defects per release
  • Critical journey coverage
  • Suite runtime and flake rate
  • Time from commit to green pipeline
Typical timeline

3–5 weeks to a trusted suite in CI

Practicalities

What week one
actually looks like.

First week

Critical journeys identified with you and ranked by what failure actually costs, then a harness running in CI with the first few tests green before anyone writes a hundred.

What we need from you
  • Repository and CI access
  • A staging environment with seed data
  • Test accounts and payment sandboxes
  • A defect triage owner
Not included
  • Manual regression testing every release, unless retained
  • Penetration testing
  • Fixing the defects found, unless scoped
  • Device farm subscription costs

Scope boundaries stated up front rather than discovered in a change request.

How the work runs

Four stages, visible throughout.

01

Model

The domain, data and permissions written down before any interface exists, because the schema outlives every screen above it.

02

Prototype

Flows and states tested as a clickable prototype, so the expensive disagreements happen in Figma rather than in code.

03

Ship in slices

One workflow at a time, in production, with tests and observability included in the definition of done.

04

Operate

Release tooling, monitoring and a documented handover so your team can run and extend what we built.

Questions buyers actually ask

QA & testing, answered.

Automate the journeys that carry money, data or reputation, and the regressions you have already been bitten by twice. Chasing a coverage number produces slow suites full of tests nobody reads. Twenty reliable end-to-end tests beat four hundred flaky ones.

That is the usual arrangement. We build the harness, write the first suite, wire it into CI and train your engineers to extend it. Some clients keep us on for release-cycle testing; most do not need to.

Automated tools catch roughly a third of real issues. We run those in CI and add manual keyboard, focus-order and screen-reader passes on key journeys, because that is where the failures that block actual users tend to live.

A better brief starts here

Make qa & testing useful.

Bring us the hard thing. We will bring a point of view, a senior team and a clear next step.

Start a conversation