How to evaluate a software QA company: the eight questions that matter

Eight questions to ask any QA company before you sign, with what a good answer sounds like and the red flag that should end the call.
By Tudor Brad, founder of BetterQA. Written from the questions buyers ask on sales calls. October 2026.
// short answer

To evaluate a QA service provider, ask eight things before you sign: how the price is built, whether you can try them first, who starts and when, whether they work inside your sprints, whether the same people stay, how they prove AI, security and compliance testing, what happens if you leave, and whether they also build the product. A good provider answers each one with specifics. A vague answer is the red flag.

Looking for a managed QA team? This is how QA outsourcing with BetterQA works, so you can put us through the same eight questions.

Most QA vendors look the same on a website: the same testing types, the same tools, the same certifications. They differ in eight places, and you only find them by asking. These are the questions that come up on almost every call we have with a buyer, with what a good answer sounds like and the answer that should end the conversation.

question 1

How is the price built, and what am I billed for each month?

why it matters

A low hourly rate with unclear hours costs more than a higher rate you can check. You want to know what you pay for, not just how much.

good answer

A clear model (time and material, or a fixed team), a rate range by seniority, monthly invoices after the work, and hours you can see broken down by task.

red flag

A per-seat tool licence on top of the people, or “we’ll send a quote” with no range before a long discovery phase.

question 2

Can we try you before we commit?

why it matters

A trial on your product tells you more than any case study. A vendor confident in its work makes trying cheap.

good answer

A short proof of concept on a slice of your product under NDA, or a first period you do not pay for if you are not happy, with agreed success criteria.

red flag

A minimum contract of several months before you see a single bug report.

question 3

How fast can you start, and with whom?

why it matters

Your release does not wait for their hiring. Ask who would start, not just when.

good answer

A start date in weeks, named engineers you can meet, and honesty that contracts and access often take longer than finding the people.

red flag

“We’ll start recruiting once you sign.” You are paying for their hiring time.

question 4

Will your people work inside our sprints and tools?

why it matters

Testing that happens after the sprint finds bugs after the sprint. The value is in being there when requirements are written.

good answer

They join your standups, your Jira or Linear, your Slack, and work on your project only. They read stories before code is written.

red flag

A separate portal you have to log into to see results, or engineers shared across several clients at once.

question 5

Who exactly will we get, and will they stay?

why it matters

What a tester learns about your product is most of their value. Rotation throws it away every few months.

good answer

A straight answer about seniority (a senior works without being managed, a junior needs direction and costs less), and the same people for the life of the project.

red flag

A senior on the sales call and someone else on the project, or “the team” with no names.

question 6

Can you test our AI features, security and compliance, and prove it?

why it matters

AI features fail in new ways, and regulated teams need evidence an auditor will accept, not just a pass rate.

good answer

Concrete examples (prompt injection, data leaking out, answers that sound right and are wrong), security testing as part of each release, and a per-release report of what was tested and what passed.

red flag

“Our AI tests everything.” Ask who checks what the AI produced. If the answer is nobody, keep looking.

question 7

What happens if we stop working with you?

why it matters

The test suite is an asset you paid for. Lock-in shows up on the way out, so ask on the way in.

good answer

Automation lives in your repository from day one, test cases export to a standard format, and the notice period is weeks, not quarters.

red flag

Tests stored in the vendor’s own tool with no export, or a long notice period you only learn about in the contract.

question 8

Do you also fix the bugs you find?

why it matters

A team that builds and tests the same code has a reason to look away. Independence is what makes a bug report trustworthy.

good answer

No: they test, and each report comes with steps, evidence and the likely cause so your developers fix it fast.

red flag

“We can do development too.” Then ask who reviews their code.

// the checklist to take into your calls
  1. Price model, rate range by seniority, monthly invoice after the work
  2. A trial or proof of concept with success criteria
  3. A start date in weeks and named engineers
  4. Inside your sprints and tools, on your project only
  5. Seniority stated plainly, same people for the whole project
  6. Real examples for AI, security and compliance, and a per-release report
  7. Tests in your repository, standard export, notice in weeks
  8. They test, they do not build your product

If you want to put us through these eight questions, book a 30-minute call.

Share the Post: