I wish I could find it but there was a article written by someone who I think had some connection to sales about how OS/2 was effectively "sold wrong" by IBM who just didn't seem to know how to sell software like OS/2.
I remember posting that article on a company intranet once and a bunch of former IBM sales folks (who now worked at the same company I did) chimed in to echo the idea that IBM had a neat product with OS/2 but as an organization had no idea what to do with it.
Not to say it would have overtaken Windows, but it also struggled because it was sold by a company who didn't know what to do with it.
I remember in the warp days they were throwing free cdroms at anyone who asked, full licensed. I got several at trade shows. They were trying to sell it for sure.
No marketing campaign or OEM sales strategy would have saved OS/2.
All the things that made it better than Windows 3.0/3.1 also meant it was totally unsuited to fill the disruptive low-end of the PC market that ate Windows up.
The resulting Microsoft platform dominance led to ecosystem dominance led to software library dominance and that was the ballgame.
"IBM who just didn't seem to know how to sell software like OS/2"
Clearly, IBM was never that great at marketing much of anything IMO. However, many at the time also believed that in addition to not being that good at marketing, IBM (collectively) wasn't really overly interested in marketing OS/2.
My understanding is that by Windows 95, IBM's PC division preferred to sell computers with Windows to OS/2.