Indeed: for example, Crawford dropped on Twitter a while back that most people playing do not use Feats, and a lot of people on here were pretty resistant to the knthat Feats are a variant rule that isn't normally in play when 5E is played.
People are "resistant" to that idea because it's an extrapolation that is not the most rational explanation of the data he's present, and he doesn't claim what you're claiming.
We know from other data that the vast majority of people play 5E in the 1-10 level range. I can't remember the exact figures, but I think it actually centered on like 3-8. That alone would explain most games not using Feats, because it doesn't make any sense to take Feats over ASI for like 90% of classes in that level range, and the "casual" option is to simply take an ASI, so ironically, less optimization-prone players are likely to take the optimal option.
Thus the most logical explanation for his data is that due to the low levels 5E is played at, plus people who genuinely, intentionally exclude Feats (even if that latter is only 5, 10, or 30 percent of the total players, I believe we have no credible figures on that), the majority of games would not include any Feats. This is certainly reflected in the campaigns I've played in, in 5E. None of them have banned Feats. But, let's see, yeah in the 6 of them I'm confident about reporting on, 4 don't actually have any PCs with Feats.
There's also the question of how you count Ravenloft. I've excluded it, but we have a small Ravenloft campaign (3 PCs), and none of them have Feats, but all of them have Supernatural Gifts or whatever they're called, which the DM was willing to allow to be swapped for Feats, but no-one made that choice.
It the 4/6 no Feats because it's a "variant rule" that's intentionally "not in play"? No. It's because we don't usually play long enough to reach the levels where they make sense given their direct competition with powerful ASIs. I don't think it's remotely a coincidence that the two games where PCs do have Feats are the two highest-level ones.
He did not dive into specifics, but he was countering the narrative that people playing the game make choices based on power optimization:
He is absolutely saying that but there's a huge problem with his claim - the most common D&D races are some of the most overpowered, and indeed, the PHB has probably the most unbalanced races in all of 5E in it, indeed it definitely does if you take the recent Multiverse update and apply it.
Elves, Half-Elves and Dwarves, for example, are very common, and pretty powerful. So how would we even tell if say, 55% of players were picking races primarily for power? I think the racial line-up would nearly identical to what it looks like today. Yeah, if 100% were, or 80%, it might look a bit different.
Also the most extreme optimizer in any group I play in picked non-variant human in the last two campaigns, and not for aesthetics, he genuinely thought it was the best optimization. That might mean he's an incompetent optimizer, but he's trying lol.