Start by refusing to answer the question with headcount, test-case volume or automation percentage.
A large QA function may be excellent value if it protects critical revenue, catches expensive failures early, provides strong independent assurance and helps delivery improve. It may be terrible value if hundreds of people are executing low-value regression and feeding dashboards nobody trusts.
If the bill is obvious but the value is not, that is a perfectly reasonable thing to challenge. We can help trace what the function is actually buying you before anybody reaches for a blanket headcount cut.
Follow the outcomes
Look at serious production incidents and escaped defects. Look at release delays caused by quality. Look at how quickly teams get useful feedback. Look at how much test effort goes into critical risk versus repetitive low-value checking.
Then examine the operating model. Are testers embedded where they can influence design? Is there duplicated governance? Are multiple suppliers testing the same things? Is automation reducing effort or simply creating another estate to maintain? Are expensive specialists doing work that does not require specialists? If no one can explain the work behind the bill, start with what the test team actually does all day.
Do not ask a consultancy for a maturity score unless you need one
You can spend a remarkable amount of money being told you are ‘Level 2 moving towards Level 3’. The score is only useful if it leads to better decisions. Otherwise you have purchased a very expensive horoscope for your QA function.
Cost the friction as well as the people
A function can be expensive without having a huge headcount. Consider the delay caused by slow gates, duplicated reviews, unstable automation, hand-offs between suppliers and teams waiting for formal test cycles. Quality controls that routinely slow safe changes without catching meaningful problems have a cost beyond salary.
Equally, do not slash a team because the visible defect count is low. Low defect escape may be evidence that the controls are working. You need to understand contribution before removing capability.
The sensible review asks what risk each part of the function controls, how effectively it controls it and whether there is a cheaper or faster way to achieve the same outcome.
Sample what the people are actually doing as well. Senior testers spending their week copying results into presentations, automation engineers babysitting flaky suites and whole teams waiting for environments are costs that a headcount chart will not explain. Equally, a small group preventing a high-value incident every release may be worth far more than its size suggests.
If you want to know whether the cost is justified, we can review the function independently. We can look at spend, structure, delivery evidence, supplier performance, automation, incidents and reporting, then tell you where the value is, where the waste is and what we would change.