Development
Software testing checklist: how we test websites and apps before launch
The checklist we run before any website, shop or app goes live, what each check catches, how long it takes and what testing adds…
· 8 min read
DevOps is a set of habits and tools that make changing software safe and routine. Every change is tracked, checked automatically, released the same way every time, and watched once it’s live. The opposite, which plenty of UK businesses have without realising it, is someone editing files over FTP on the live server on a Friday afternoon and hoping.
You don’t need a platform team or a Kubernetes cluster to get the benefit. A business with one WordPress site and a part-time developer gets most of the value from five or six basics done properly. We set those up for companies across the UK, sized to the risk rather than to whatever is fashionable on conference stages this year.
Signs you need some DevOps
- Changes are made directly on the live site and nobody can say what changed last week, or who changed it.
- There’s no staging copy, so updates are tested on customers.
- A plug-in update has taken the site down before, and getting it back took a day.
- Only one developer knows how to deploy, and they’re freelance, or leaving.
- You have backups, in theory, but nobody has ever restored one.
- You find out the site is down when a customer phones to tell you.
If two or more of those sound familiar, the fix is usually a few days of work, not a transformation programme.
What we set up
- Git version control for your theme, plug-ins and custom code, so every change is recorded, reviewable and reversible.
- A staging environment that mirrors live, with a sanitised copy of the data, for testing updates and new features.
- Automated deployment (CI/CD) through GitHub Actions or GitLab CI, so a release is a merge, not a manual upload.
- Automated checks before anything goes live: code linting, tests where they exist, and Lighthouse checks for performance and accessibility so a heavy image or a missing label doesn’t slip through.
- Scheduled off-site backups of database and files, with a documented restore we’ve tested end to end.
- Monitoring for uptime, errors and slow pages, with alerts to the person who can act on them.
- Docker containers for custom applications, so development, staging and live behave identically.
- A short runbook: how to deploy, roll back, restore, and who to call.
Sized for small teams, not enterprises
Here is roughly what we’d put in place at three common sizes. The point is to stop at the level the risk justifies.
| Your setup | What we’d recommend | Typical effort |
|---|---|---|
| Brochure WordPress site, occasional edits | Git for the theme, staging, nightly off-site backups, uptime monitoring | 6 to 10 hours |
| WooCommerce shop taking daily orders | All of the above, plus automated deployment, hourly database backups, error alerts and a tested rollback | 12 to 20 hours |
| Custom Laravel or Node app, two or more developers | CI pipeline with tests, Docker environments, tagged releases, staged database migrations, log aggregation | 20 to 40 hours |
We’d only suggest Kubernetes or a multi-region set-up with a clear reason, such as genuine traffic spikes or a contractual uptime commitment. For most UK SMEs it adds cost and complexity they’ll never use.
How we work
Audit
We look at how your site or app is hosted, changed and backed up today, and write down the risks in plain English. This usually takes two to four hours.
Quick wins first
Backups and monitoring come first, because they protect you from the worst outcomes immediately. Version control and staging follow.
Automate the release
We build the pipeline that runs the checks and ships the change. Rolling back becomes one action rather than a frantic hour on the phone.
Hand over
You get documentation and a recorded walkthrough, so whoever works on the site next, us or someone else, follows the same process. We don’t lock you into our accounts; the repository, pipeline and backup storage sit in accounts your company owns.
Testing in the pipeline
Automated checks catch the obvious mistakes, but they work best alongside a short human checklist before each significant release. Our software testing checklist is the one we use on our own projects, and we adapt it for clients’ teams.
Technology
Git on GitHub, GitLab or Bitbucket; GitHub Actions or GitLab CI; Docker for custom apps; WP-CLI and Composer for WordPress; Cloudflare for DNS, caching and the firewall; and hosting on managed WordPress platforms, Linux virtual servers or AWS, Google Cloud and Azure. Our cloud solutions page covers hosting choices in more detail.
What DevOps costs
DevOps work is booked as custom web development hours at £60 an hour excluding VAT, with volume discounts. Worked examples:
- WordPress site with Git, staging, automated deployment, backups and monitoring, about 15 hours: £900 less 10 % = £810 plus VAT (£972 inc. VAT).
- Custom application pipeline with tests and containers, about 20 hours: £1,200 less 15 % = £1,020 plus VAT.
- Larger clean-up of an inherited app with no tests and manual deployments, 50 hours: £3,000 less 20 % = £2,400 plus VAT.
Running costs are often close to zero for small sites, because GitHub Actions and most uptime monitors have free tiers that cover them. Off-site backup storage for a typical business site is a few pounds a month.
Inherited a site from another agency?
That’s most of our DevOps work. We start with an audit of the code and hosting, bring it under version control and add staging, backups and deployment without changing how the site looks or behaves. It works the same in Birmingham, Cardiff or Belfast, on a cloud server or on a £5-a-month shared plan.
Make your next release a non-event
Tell us how your site or app is updated today and what has gone wrong before. We’ll suggest the smallest set-up that removes the risk. Contact us or call 020 3918 3814.
