I'm surprised that 29% get some attention. The nature of the messages, the quality of attention, and the point of it all, of course, are additional considerations.
Also, in many cases where Twitter can be very useful... as in sending out and receiving announcements... utility has nothing to do with actions of response.
I'd say that the twitter spectrum is so broad and fragmented that these stats by themselves don't mean much. The (overall) Retweet Time Span Histogram looks like retweets are almost reflexive (automated?) in a great many areas, with an exponential decay. But, in more focused areas, that may not be the case at all.
Only 1.53% of Twitter conversations are three levels deep - after the original tweet, there is a reply, reply to the reply, and reply to the reply of reply.
...only a small number of users actually have the ability to engage on Twitter in a significant way.
From my personal perspective... and I can see a great many uses for the platform... one of the great problems I have, at least in my limited efforts to play with it deeply, is it implicitly assumes a continuing presence.
"Conversations" extended over time become hard to follow. If there's a threading capability for general back and forth messages, I've missed it. So, the platform itself demands a kind of daily addiction, if not more frequent than that. While the platform may generate momentum, it also may require it. This, from a semi-outside perspective, feels like a limitation.