QA testing is the work of finding out whether software does what it's supposed to, before a customer does. It isn't clicking around hoping something breaks. It's forming a view of how a system is meant to behave, then deliberately looking for the places where that belief turns out to be wrong.
What the job actually involves
Most of it is thinking, not clicking. Before you test a feature you decide what "working" means, which cases matter, and where the assumptions are weakest — the empty field, the enormous number, the two people doing the same thing at once, the user who goes backwards through the flow.
Then you write it up. A bug report that a developer can reproduce in thirty seconds is worth more than ten vague ones, and writing clearly is a larger part of the job than most people expect.
The automation misunderstanding
Beginners are told constantly that manual testing is dead and they need Selenium or Playwright to get hired. This sends people into automation tutorials before they can test at all — and the result is someone who can write a script but can't say what the script should check.
Automation is valuable and worth learning. But it automates tests you already know how to design. Learning it first is like learning to type faster before deciding what to write.
Manual and exploratory testing are still employed widely, and the people who do them well find the things automated suites never catch, because an automated test only ever checks what someone already thought of.
What makes someone good at it
Scepticism. The willingness to believe something is broken when everyone else is sure it isn't.
Curiosity about systems. Wanting to know what happens underneath, not just whether the button works.
Clear writing. Half of testing is reporting. A finding nobody can reproduce doesn't exist.
Tolerance for being unpopular. You're the person who says the thing isn't ready. Done well, that's valued. It's still uncomfortable.
How people actually get in
Certification helps as a signal, not a skill. ISTQB Foundation is the common entry qualification and it gets your CV read. It doesn't teach you to test, and nobody good was ever hired because of it alone.
You can practise on real software today. Pick an app you use, test it properly, write the bugs up as though you were being paid. That portfolio is more convincing than a certificate.
The background doesn't have to be technical. Some of the best testers come from support, teaching, administration — jobs that train you to notice when something doesn't add up and to describe it precisely.
Learn one automation tool, after the basics. Playwright or Cypress. One is enough to show you can.
Where to start
If you can already describe exactly how something went wrong and why it matters, you have the hardest part. The rest — the vocabulary, the tooling, the certificate — is learnable, and faster with someone who has done the job telling you which parts are worth your time.