How to choose a software testing company
Thinking about outsourcing QA? Here are the questions to ask a software testing company, what a good bug report looks like, how pricing models compare, and the warning signs.

Hiring a testing company is harder to judge than hiring a designer or a developer. A designer shows you a design. A developer shows you a working feature. A tester's best work is a problem that never reached your customers, which means that when it goes well there is nothing to see.
So buyers tend to choose on price or on a confident sales call, and find out three months later whether it was a good choice.
This guide is a way to find out sooner. It is written by a testing company, so read it with that in mind. The questions apply to us as much as to anyone.
Start with what you need
"We need QA" covers several different jobs, and companies are rarely good at all of them.
- Manual testing: people using your product to find problems before release.
- Test automation: scripts that recheck the important paths on every change.
- Specialist testing: security, performance under load, or accessibility.
- Ongoing or one-off: a partner for every release, or a single check before a launch.
Knowing which one you want narrows the field quickly, and stops you paying for an automation team when you needed two days of careful manual testing.
Ask to see a real bug report
This is the most useful thing you can do, and few buyers do it.
Ask for an example report, with client details removed. A good one has:
- A title that says what is wrong, not where it was found.
- Numbered steps that let a developer reproduce it on the first attempt.
- What was expected and what happened.
- The device, browser and account used.
- A screenshot or recording.
- A severity, with a reason.
A weak report says "checkout not working on mobile" and leaves the developer to spend an hour finding out what that means. The quality of the report is the quality of the service, because the report is what you are buying. Our own standard is described in how to write a bug report.
Questions worth asking
Who will do the testing? The person on the sales call is often not the person who tests. Ask who is, and how long they have done this.
What will you test first, and why? A thoughtful answer starts with your riskiest features: payments, sign-in, personal data. A weak one promises to test everything.
Which devices and browsers do you test on? Real phones matter. Emulators miss problems that real hardware shows.
How will we hear about bugs? Into your own tracker, as they are found, is best. A PDF at the end of the month is the slowest way to learn about a serious problem.
What do you not do? A company that says yes to every request is not being straight with you. Good ones are clear about their limits.
What happens when you find nothing? It can mean the product is in good shape. It can also mean the testing was shallow. Ask how they would show you which it is.
Compare pricing models
How a company charges changes how it behaves.
| Model | How it works | Suits | Watch for |
|---|---|---|---|
| Hourly | You pay for time spent | Open-ended or changing work | Hours do not guarantee findings |
| Monthly retainer | A fixed team for a fixed fee | Products with regular releases | Paying in quiet months |
| Fixed price | One price for a defined scope | A single launch or audit | Scope disputes if the product changes |
| Pay per bug | You pay for confirmed bugs | Early products, tight budgets | Needs clear rules on what counts |
No model is best for everyone. The right one depends on how often you release and how predictable your budget needs to be. We explain the last model in pay-as-you-go QA pricing.
Warning signs
- A promise of zero bugs. Nobody can deliver this, and anyone offering it is selling.
- No questions about your product. A tester who does not ask what the product does or who uses it cannot know what matters.
- Automation as the answer to everything. Automation rechecks known paths. It does not explore, and a brand-new product usually needs exploring first.
- Reports full of trivial findings. Fifty spelling mistakes and no mention of the payment flow suggests effort went where it was easy.
- Reluctance to do a small trial.
Run a trial first
The safest way to choose is a short paid test. Give two companies the same part of your product for two or three days and compare what comes back.
Look at what they found, how clearly they wrote it up, and whether they noticed the problems you already knew about. One trial like this tells you more than any number of proposals.
Think about fit
A large testing firm with a formal process suits a bank. A startup releasing twice a week needs a few people who answer quickly and adapt. Neither is better. A mismatch in either direction is expensive.
Also consider time zones, how fast questions get answered, and whether the testers will talk directly to your developers. The work goes much better when they do.
About us
SQAfox is a small QA team. We do manual testing of websites and mobile apps, Playwright automation, and security testing. We are a good fit for startups and small product teams, and probably not for an organisation that needs a hundred testers on site.
If you want to judge us the way this guide suggests, start a project and ask for a short trial on one part of your product.