I think, something which is missed, is what I'd call the "mediametic horizon". In the "old days", people weren't just "doing their thing", there was more to it. Even, if you entertained just a small site, you were somewhat playing big media (compare the vintage radio allusions or the ubiquitous news signations that popped up about 1997.) Small websites were pulling for reach, dreaming of a mention on "What's cool" lists. (Besides other functions, HN may be the last true what's-cool list today.) Even those building professional sites were driven by this "mediametic horizon": you wouldn't build the same website twice, not even the same concept. Adding to and enriching what was the known standard, the known functionality, the known capabilities of the Web was the raison d'être of any new site. Becoming known and mentioned for such a feat was a major incentive, both for the designer-developers and for the clients (again, for reach and prestige). The truth being, the "old web" essentially required and depended on brokers of reach and prestige to develop this drive, and I do not see any candidates for fulfilling this function. (Apart from HN, that is.)
P.S./Edit: This was still true for the early 2000s: back then, I was involved in a major relaunch of the site of a telephone carrier and the clients were scanning relevant Usenet groups for critiques in order to evaluate their success. (Surely, this wasn't the norm, but it still did happen.)