I still kinda wonder if they saw the success of the invite system for gmail (I remember a lot of late nights begging for an invite on various forums) and thought that it would work again.
The critical difference is gmail still worked just fine with hotmail, yahoo mail, aol, etc. Wave was useless if both sides didn't have it.
As I recall, at one point Wave sort of had enough of an XMPP bridge that you could terribly IM a Wave without having a Wave invite if you were one of the 20 people still using XMPP that month and your friends with Wave knew a "secret" @ mention and you felt like learning an XML mini-DSL of pseudo-commands and kinda-unidiffs to read the changes from the people actually in Wave.
There was also plenty of talk about the "eventual" email bridge and real multi-server Wave federation, neither of which properly happened. (At least not in the invite months).
Though, yeah, Wave really could have used the network effects of non-scarce invites, because it wasn't as interoperable or as much of an "open standard" as it wanted to be. Or it should have had all that interoperability and open standards properly ready at launch and the Google server could have just been sold as the "best" of several options (and people waiting for invites could self-host; that might have done enough for viral class projects in college environments).
They'd already experienced the downsides of an invite-based rollout for a closed network, thanks to Orkut in the mid-2000s.
It flopped in English-speaking countries because invites were so limited when people first started talking about it, but became a success in Brazil and India as the buzz built a little later there, by which time it had become easier to get and share invites.
They then compounded the error by force-partitioning their users between the existing service and an invite-only New Orkut, with no easy way to communicate between the two.
That disaster was still playing out when Wave launched, so at least some part of Google ought to have been aware of the importance of network effects for a product of this type.
not to mention that i think there was some google+ initiative back then (i might've gotten the timing wrong tho). There's some office/department political machinations in the background, and the fallout of that ruined wave.
At that time email was validated, there was no doubt people wanted it, gmail was just better email. Contrast that to something like Wave which requires people to try something really new.
What could Wave have done better? explain why they need invites? Even better, expose their reasoning, eg they don't need to ease server pressure but they need quality signups? Anything fun I'm missing? Like skin-in-the-game moves from the private side, for macroscopic values of skin?
For Wave, I'd imagine they needed to publish data on which fun parts keep the new users returning ---there were MANY!
(So, we're both clearly not wishing to see their bugs from swiftly tilting these parts :)
The critical difference is gmail still worked just fine with hotmail, yahoo mail, aol, etc. Wave was useless if both sides didn't have it.