Pay-as-you-go QA: paying for bugs found, not hours spent
How pay-as-you-go software testing works, how findings are priced by severity, what counts as a finding, and when another pricing model suits you better.

Most testing is sold by the hour or by the month. You pay for time, and hope the time turns up something useful. Pay-as-you-go turns that round: you pay for what is found.
How it works
- We agree rates before any testing starts. There is a price for each kind of finding: a critical bug, a high, a medium, a low, and an improvement.
- We test your product and log each finding with its severity, steps to reproduce, and who found it and when.
- You receive a report listing every finding with its price, and a total.
- You pay for the findings in the report.
Every line on the bill is a specific problem in your product that you can open, read and check for yourself.
How severity sets the price
Not all bugs are worth the same. A typo and a checkout that takes payment twice should not cost the same to find. We use four levels for bugs:
- Critical: blocks a main journey, or loses money or data
- High: a main feature is broken, with no easy way round
- Medium: something is wrong, but people can still finish
- Low: small or cosmetic
Improvements are priced separately. An improvement is not a defect. It is a suggestion: a confusing label, a step that could be removed, a message that could be clearer.
What counts as a finding
This is where pay-as-you-go needs clear rules, so they are worth stating plainly.
- A finding must be reproducible. It comes with steps anyone can follow to see it.
- One problem is one finding. The same bug on five pages, from the same cause, is one finding, not five.
- Duplicates are not charged. If we report something twice, you pay once.
- Working as designed is not a bug. If the product does what you intended, it is not a defect. It may still be worth raising as an improvement, and you decide whether it is.
- You can disagree. If you think a finding is wrong or its severity is too high, say so. That conversation is part of the work.
Why teams choose it
The cost follows the value. If your product is in good shape, there is less to find and you pay less. If it has serious problems, you pay more, and you learn about them before your users do.
No commitment to start. You do not sign up for a month of testing to find out whether testing is worth it for you.
The bill is easy to check. Each charge is tied to a finding you can verify.
When it is not the best fit
Pay-as-you-go suits finding problems. It is the wrong shape for some other kinds of work:
- Test automation. Writing an automated test suite is building something, not finding defects. We price that as one agreed amount for a defined list of flows. See Playwright automation.
- Sign-off testing where the goal is "nothing found". If you need a full regression pass before every release and expect it to come back clean, you are paying for assurance, and a fixed price per pass is fairer to both sides.
- Long-term embedded testing. A tester working inside your team every day is better priced by time.
Questions to ask any testing provider
Whatever pricing model you choose, these questions tell you what you are buying:
- What exactly do I receive at the end?
- How is severity decided, and can I challenge it?
- What happens with duplicates and with behaviour I consider correct?
- Are fixes re-tested, and is that included?
- Who owns the report and any tests written?
How we do it
Manual testing with SQAfox is pay as you go: website testing, mobile app testing, API testing, performance testing and accessibility testing. Rates are agreed with you before we start. The report comes as a PDF or an Excel sheet, with an ID, severity, steps, reporter and time for every finding.
To see what a well-written finding looks like, read how to write a bug report developers can act on.