Testing software is layered work, and the layer a check happens at decides what kind of bug it can catch. The International Software Testing Qualifications Board, a standards body whose glossary defines the vocabulary the discipline runs on, defines unit testing as “a test level that focuses on evaluating the smallest part of code testable in isolation,” and integration testing as “a test level that focuses on interactions between components or systems.” A unit test checking a single damage-calculation function in isolation can confirm the arithmetic is right without touching a renderer, a network layer, or a save system at all; it will not catch a bug that only appears when that function’s result gets passed to another system that misreads it. That’s integration testing’s job instead, run after individual units are already trusted, specifically to catch failures that only show up at the seams between systems.
Neither layer is specific to games, but a shipped game adds testing categories most software never needs. Certification is the platform holder’s own review, the process a console manufacturer runs before allowing a build onto its storefront, checking technical requirements like load time budgets, save data handling, controller support and stability under conditions the platform defines rather than the studio. A build that passes every internal test a studio has can still fail certification, because certification is checking against a separate, external standard the studio doesn’t write itself. QA, quality assurance, is the discipline that runs continuously underneath both of those, and it isn’t only automated: the UK’s official careers guidance for the role describes a games tester’s day-to-day work as playing a game specifically to break it, writing bug reports precise enough for an engineer to reproduce the problem, verifying fixes once they land, and checking for issues like copyright problems or text errors that an automated test would never catch, since they require a human judgment call about what a player will actually notice.
The International Game Developers Association’s own description of who makes a game names QA and localization testing alongside programming, design and art as distinct disciplines in their own right, not as an afterthought bolted onto engineering. That standing matters for how testing actually gets organized at a studio: a dedicated QA team, running structured test passes against a game’s current build, exists alongside the engineering tests that get run automatically on every code change, and the two catch different things. An automated unit or integration test suite runs the same checks every time, fast enough to run on every change a programmer makes, and it’s good at catching a regression the moment it’s introduced. A human QA pass is slower and less repeatable, but it’s the layer that catches the failures no automated test was written to look for in the first place, because nobody anticipated that exact interaction until a tester found it by playing the game the way a real player eventually will.
Certification sits above both as the final gate a shipped build has to clear, and it’s also, per the UK careers guidance already cited, one of the more common entry points into the games industry from outside it: a tester’s day-to-day work of finding and precisely describing failures is a skill that transfers directly into QA management, and from there into design, engineering or production roles, for someone who started without a shipped game credit to their name at all.