I wish good luck to the new union. I really do. At the same time, I fear for the outcome of this action: having worked for years in the game industry, game QA is one the game industry's jobs with lowest entry barrier, employing the least paid and qualified professionals. I know a lot of great and hard-working people doing game QA, but they are among the easiest to replace and outsource. This actually explains why they would want to create a union, and we will all now see what happens.
Yeah if they push too hard, more developers will go the "early access IS QA" route. game devs have already started gamifying bug reporting for in game loot, skins, etc.
No one is more eager or qualified to give feedback about games but the gamers themselves. The downside in this industry is minimal. No one can sue a game dev if a bug slips out, assets can be rolled back or whole systems changed to negate their impact with no recourse available to consumers.
QA is not just about running the software, observing problems, and filing them. As you say, eager beta testers can do that!
QA is also about: 1. Exploring every corner of the software, comparing it to the written spec, and measuring any discrepancies, even those not visible to the user. 2. Exploring beyond the corners of the spec and using that exploration to improve the documents. 3. Writing software to automate this exploration and to automate defect-finding in general. 4. Fuzzing and other “red team” penetration testing. 5. Participating in design reviews and contributing towards the testability of the product. In other words working side by side with the developer to ensure the hooks are there to make the software testable and automatable. 6. Monitoring metrics and telemetry to find trends that are emerging post-deployment and getting them documented and reported before they are widely noticed.
And so on.
QA should be a highly skilled, highly technical job, a level equal to developers. If QA at your company is just waiting for a new build to get thrown over the wall so they can manually test it in a vacuum, your company might be doing it wrong.
General software QA does, or at least should, work like that but games QA isn't like general software QA. Since games software is basically a one-shot project (i.e. there's no expectation of long term support beyond a few patches) and not really amenable to automation, games QA are usually non-technical manual testers given fixed scripts to execute.
that's not a new concept; the first everquest is still being updated -- but parent is still right, the 'normal game' has an initial release, a few updates to fix those bugs, maybe a content patch, a few updates to fix those new bugs, and that's about it.
Agree with all of this. Our QA do all this and also provide dev support to designers, content creators and engineers, support build and editor stability in addition to the actual game quality, manage several labs with various hardware and set-up for reviews/playtests etc. and they do a lot of automation work and process improvements/documentation, including a lot of the onboarding documentation because as someone else pointed out, they have a good understanding of all aspects of the projects.
I really like this idea, but what you’re describing sounds more like “Software Engineer, Testing Department” than QA. In my experience, QA functions more like an extension of automated unit tests for the things an engineer hasn’t gotten around to automating yet.
SDET (Software Developer Engineer in Test) as a title/career path seems to have gone mostly out of vogue in the past decade or so unfortunately. Most companies I've worked for seem to hold the philosophy that the engineers developing the features themselves should do the testing. Which is a legitimate viewpoint, but there are of course pros and cons to everything and not every use case is really suited for that.
I think part of the problem for SDET is that pay bands are often different (read: lower) for SDET, and if you are doing SDET you already probably have the skills to be a developer and are going to opt for the higher pay band/career path. A lot of companies seem to hold a weird perspective that SDET is "lesser" or requires less skills than just a typical feature developer.
They're most definitely not professionals in the sense that their profession isn't playing the game competitively.
They might be better skilled than the average players, but their actual profession is QA, e.g. making sure the game runs, not that it's perfectly balanced.
As far as balance feedback, you can't do better than actual top players.
Actually pro players may have the deepest understanding of the balance but they only represent a tiny fraction of players and while useful for esports, implementing their expert feedback may alienate the broader audience.
QA may provide good balance feedback but are rarely valued for it/do it on their own time with career progression in mind. They often are a rare combination of players who also understand the dev side.
For actual tuning before release, playtesting and other user research involving players representative of the target audience is the gold standard.
Even early access fails for QA. I was randomly selected for a private beta of a game, played it, sent feedback, about 20+ bugs none of which were fully game breaking but I'd say some were pretty severe... Every bug was present on release. Every. Bug.
Corporate thinking being what it is, I don't hold out hope they will ever learn you can't lean and bean-count a creative industry, it isn't a widget factory.
Upside is when they lean so much their financials get shakey because gamers are staying away you can pick up their entire catalog for 75%+ off, even better if you pick a game with a solid mod community that patches and adds functionality.
All the widely anticipated games have more than enough folks nowadays willing to endure through janky private alphas and betas. So much so that it's common for a lot of bugs identified in later beta stages to not even be fixed.
So a large QA team is only needed nowadays for games that nobody is really that hyped about, or games that cannot risk any chance of plot-lines/art leaking before release.
With how common extremely buggy, near unplayable releases are these days, I'm surprised QA testing is still a thing at AAA studios, because it certainly doesn't show.
It doesn't show because the management has the same opinion as glimshe.
That's opinion is wrong though, because testing is more then just playing the game a little.
You get the same issues in companies that straight up don't value tests whatsoever. Unsurprisingly, bugs become increasingly frequent, especially if the software is developed by multiple people.
Methodically testing something in a reproducible manner, potentially through scripting to enable regression testing is challenging, in some cases more challenging then writing the code/behavior you're testing.
glimshe isn't wrong about how hard this is going to be for the QA people though. These companies don't value quality, which is precisely the reason why they're hiring the people that are easiest to replace. Most QA workers currently just go through the motions of a "manual acceptance test", with the result you just mentioned.
Good QA is not easily replaced but it's a job you under hire for a large part of the game lifecycle. It's cheaper to gut the team after a release and bite the hiring cost on the next game than to hold on to them. You don't need a legion of folks testing a prototype.
Pretty sad because good QA folks provide a lot of value despite what I'm reading in some of the other comments here.
Even if it just lead to OT pay and more visibility into planned layoffs, it would be a win for the QA folks. As it is, big studios plan crunch time leading up to a launch and then large layoffs. They never say who's getting laid off so you have to just pretend its not happening.
In the film industry, the fact the job ends is just part of the culture. At least owning up to that, paying OT, and giving folks an opportunity to line up another gig would be good. Even better if the union handles healthcare so there's no disruptions there.
I don’t think more testing will lead to better quality games necessarily. My experience is testing always comes up with more bugs than devs can fix already.
from what I heard from PirateSoftware about working QA at Blizzard making you a second-class citizen (and working at Blizzard in general sucks) and a friend who worked at a Call of Duty studio as QA who got treated bad enough to be a "stress casualty" (and never returned to the games industry) their standing can only improve.
I'd imagine the fear is they get crushed, replaced with the cheapest option available located in a place with even less rights. And everyone moves on with the narrative and another move towards labor rights dies with a whimper
> replaced with the cheapest option available located in a place with even less rights
Cheapest is not equal to less rights. Compared to US most developed countries are cheaper, but with more labor rights. So it would make sense to outsource it to, say, CEE EU countries.
If you outsource to anywhere, including within your own country, you've always got a game of telephone going on.
If you outsource to anywhere with a different language (which, from the US PoV kinda includes much of Europe, but less than you might expect from the national languages), what's lost in translation is comparable to the impact of machine natural language processing — so you get the same quality from ChatGPT etc., which is cool, but not quite good enough. (And when LLMs become that good, it will make a lot of people very sad).
If you outsource to poor countries, you also have to worry about the stability of the power grid and telecoms impacting your ability to communicate with the reliability you expect in corporate circles.
Fear of leaks, dealing with timezone differences that slow down the work, cultural differences leading to misunderstandings, or generally requiring more iterations to get the same result. Sure it might be cheaper even if it takes 10x as long, but games have tight roadmaps and hard deadlines so in some cases time may matter more than money.
At least in the software business, the cheaper options than eastern Europe usually have such broad cultural differences that the management overhead doesn't offset the cheaper cost.
A unionized work force might cost more and tip the balance the the other way, for better or for worse. But mostly the problem here is probably retaliation.
Either way, some work conditions are so bad that it is a risk work to taking over a status quo that harms workers in a way they can’t stay in the industry very long regardless. QA are certainly treated the worst in an industry that already has poor work conditions while representing more $$$ than the movie industry.
Yes, it's not like there aren't already third-party games testing companies in Eastern Europe and elsewhere. I'm not sure to what degree games companies use agile and devops practices though. Working with an outsourcer for QA is a sure-fire way to increase your lead times and slow down development.
Working at a very small place in house QA was absolutely impactful in being subject matter experts on what the software actually did and was supposed to do. I am not sure how you measure that.
A good QA might be even the folks that know most how system actually works and what are all the features. Your developer can end up easily work in single corner, with few happy paths. Where as QA should touch nearly all surfaces of product. And even have idea what are most of the specs.
Exactly, we had 1 to 2 QA people compared to 6 to 8 developers, roughly speaking the developers knew more about particular details of like a third of the system where QA knew all the interactions.
That's definitely a fair point. However, that knowledge should live with other members of the team and it can be helpful for QA to have less knowledge of the system. How can that be helpful ? -- so they can view the systems through the lense of a typical end user.
I worked in game testing at two major Japanese game companies in the Bay Area for a few years before switching to doing QA in the video industry. As game testers we’d write up a ton of bug reports but generally only the most egregious crashes and console manufacturer’s standards violations would get fixed (think using the wrong Sony or Microsoft terminology or symbols in the game UI).
By the time most games reached our offices for testing they’d already undergone most of the core testing in Japan or at their developer’s studios. We were essentially just there to make sure last minute bugs didn’t slip through the cracks. Management rarely prioritized gameplay or even functionality bugs to be fixed, unless it was easy to reproduce and clearly prevented overall progress in the game.
We shipped one big fighting game and wrote up a bunch of bugs that clearly showed that online play was nigh-unusable, and nothing got fixed. The game came out and was ripped apart in reviews for the low quality of its net code. But it had a bunch of single player modes that were rock solid so they let the online modes slide. This was a few years before fighting game developers started seriously considering the importance of online play, and taking major steps to improve its quality.