Thabo Sefolosha (a little bit of offense, like Adams) and Roberson (no offense) have also earned starting consideration over better players (Harden, Reggie Jackson) for "defensive purposes" like Adams, on the Thunder as well. I don't value that accolade as highly as it generally would be.
Where's the evidence that the front offices that rely on advanced stats are in the minority?
I disagree with your implication that Steven Adams has been included in the starting lineup simply for "defensive purposes."
The association by reference with Thabo and Roberson is misleading. Adams is much less of an offensive liability than those guys. Indeed, he provides a valuable complement to Serge Ibaka's skillset. That he isn't a total sieve like Enes Kanter doesn't mean he's in the starting lineup solely for his defensive ability.
Suggesting that GMs around the league place some emphasis on the fact that a player has carved out a significant, starting role on a good team, or the fact that a player has failed to do so on a bad-to-mediocre team lacking in talent, does not require the premise that GMs do not also take analytics into account. The point is simply that the existence of analytics, indicating that a player has value that may not be immediately apparent from the box score, is not by itself enough for talent evaluators around the league to value a player especially highly.
In short, I think Kelly Olynyk probably has more value around the league than most third string PF/Cs getting inconsistent minutes on a borderline playoff team, but that shouldn't be confused with him having significant trade value. He needs to actually put together at least a couple of months of productive basketball while the team actually wins games before we can expect him to hold a lot of value in trade discussions.
The only stretch I can think of where Kelly was regularly earning minutes and putting up points was in the last month or so of 2014, when the team was just playing out the string and getting beaten by double digits more often than not.