One of the most puzzling data points of the past few years has been AMD’s complete failure to gain market share according to the Steam Hardware Survey. We’ve always warned readers that the SHS might not be accurate, based on problems we’ve observed in the data set in the past, but it’s never been clear what the problems were. Other analyst firms have reported share gains for AMD, but the Steam Hardware Survey — which represents the best public-facing data we have access to — has never shown a real gain for AMD.
A recent interview of Scott Herkelman by Hot Hardware clears up this point. I wrote about this interview in a different story yesterday, but this point is specific enough to deserve its own breakout.
According to Herkelman, the problem with SHS goes back to the August 2017 discontinuity. The problem was introduced because of a bug in the Steam Hardware Survey, but there’s allegedly a systemic undercount in AMD systems that persists to the present day.
Steam’s survey, according to Herkelman, isn’t meant to measure hardware market share for companies. It’s supposed to tell developers what kind of products are in-market. The reason the Steam hardware survey went nuts back in 2017 is that Steam was over-counting every single individual login at an iCafe as another instance of that computer’s system configuration. Imagine if you had 10 of your best friends over to play Steam games on your PC. Every single one of them logged in to their own accounts — which led to 10 copies of your system config being uploaded to the platform and counted as separate submissions.
That’s basically what happened with Steam. And according to AMD, while the company made some corrections to its data, Valve has never been particularly concerned with making sure its numbers track real-life market share. AMD, meanwhile, is drastically underrepresented in iCafe gaming.
“They did change their algorithm a little bit, but they really aren’t motivated to go in and change this,” Herkelman said, “because the purpose of their data is not for market share. The purpose of their data is to show general trends to game developers… it definitely doesn’t track our real share…. you can see the same thing actually happen in our CPU share. It’s still under-represented, it’s the same exact curve, and it’s all related to iCafes.”
AMD CPU market share on Steam. This doesn’t seem to be quite the latest data, because it shows AMD gaining market share against Intel. According to the SHS, AMD has been losing market share to Intel for several months, despite reports indicating the opposite is happening.
To bolster his point, Herkelman released images like the above, showing how AMD CPU adoption rates changed dramatically when Steam added PUBG in China, then changed again afterward when the company re-updated its algorithms. In both cases, AMD’s market share was lower as a result of the update.
The problem, of course, is that Herkelman’s comments don’t change the fact that the Steam Hardware Survey is still our best source for specific data on what kind of hardware the gamer community is using. Even when AMD and Nvidia release market data, they rarely release information on specific products or price points.
The one thing that makes no sense in all this is why Valve doesn’t care about inaccuracies in its own data set. The purpose of the SHS may not be to present accurate market share data, but presenting developers with inaccurate data is scarcely better. If developers think that more gamers own GTX 1060 and 1050 Ti cards than actually do (those being the top two GPUs on the survey, with 16.01 and 10.63 percent market share respectively), then they’ll draw the wrong conclusions about which cards they should target for future development.
The only conclusion we can draw is that Valve doesn’t feel whatever inaccuracy remains is enough to impact what developers do. AMD obviously felt strongly enough about the topic to publicly state the problems with using the SHS for market share estimates. It’s not clear what impact this might be having on Nvidia cards, either — adoption rates on some of those GPUs may also be skewed by errors.
We’ll continue to refer to the SHS at times because there’s little practical choice. It’s the only data set of its type available publicly. But this could explain why AMD’s overall CPU market share numbers have been ticking up in other reports but have remained fairly static on Steam. If Chinese iCafe installations grow more quickly than other types of gaming and AMD isn’t represented in that market, it’s not going to appear to gain much market share in either CPU or GPUs. We’ve mostly used the SHS to compare generational adoption for Turing versus Pascal, which should be less-impacted. But if the figures for AMD adoption are incorrect, the figures for at least some Nvidia SKUs will be as well.