As a "new" user on MemeStreams, a few comments.

It's hard to tell whether my blog is being read, but my memes are just not generating replies or re-recommendations, or if I'm just being ignored / not seen by the other users. There is a lack of feedback about what's happening. I can't 'see' site activity unless other users post entries of their own, and even then that doesn't necessarily correlate with a visit to the front page or the agent.

At the outset, I might be satisfied with just some click-throughs, even if I'm not garnering recommendations just yet. But I can't tell. After 39 straight posts with no response of any kind, a person can start to get discouraged. Whether 7 responses to 58 posts (12%) is a "good" feedback ratio is perhaps a matter for debate. I note that most "regular" users enjoy substantially higher feedback ratios. Although it only took a few days for me to get four of the top ten MemeStreams users into my "audience" graph, I can't really tell how meaningful this is. Below a certain threshold, being in a user's agent table is of marginal value, because there are so many total posts that mine never rise to the top, except on the slowest of days. (Back to "slow days" in a minute.)

If "community" is supposed to be an attribute of MemeStreams which separates it from a run-of-the-mill blog-hosting service, feedback for new entrants seems like a nice thing to have. (Should MemeStreams have a welcome wagon? Would it improve retention? Can it be institutionalized, like a jury duty kind of thing?)

I don't think regular users spend much time on the front page, which means that new users end up at the very bottom of a very long list in the agent. Especially as volume goes up on the system, it will be very hard for a new user to "break in."

One possible "break in" strategy is to go around the site, re-recommending memes previously posted by others -- particularly the ones that rank highest in each topic area. (As a new user, am I "supposed" to do this?)

If I'm new to the system, but already know a lot of people with blogs here, it's hard to get myself established. I have to spend a lot of time teaching the agent over a period of time, and quickly re-posting a bunch of old stuff in my friends' blogs to 'teach' the agent is unlikely to win me many new friends, because I'll appear to be completely "out of the loop" on things. If new users flood in, it may be particularly difficult to sort out the please-ignore-me-I'm-just-training-my-agent memes from the truly new stuff.

Basically, because all of the training is "on line", in the AI sense, there is no quick-start mechanism which does not have ripple effects. If my MemeStreams friends wanted to train their agents to rank me highly without waiting for this to happen 'naturally', they'd likely have to bias their response to my initial blog entries, which would create a ripple of these memes across my friends' social networks. This might annoy their other friends, which in turn might cause my friends to have to sacrifice some of their capital in order to rapidly train their agent to show them my blog entries at the proper place in their listing.

One partial alternative is for the already-established friend(s) to establish a Newcomers personal circle and add their incoming friends to this circle. Preferably, the system would let me auto-expire a 'membership' to a circle, otherwise the user would have to do this circle management manually, which might be kind of annoying. Then a friend could just run the agent with this circle to check in with newcomer friends.

An extension to this would be to auto-create a "newcomer" personal circle, or in the future, a public circle, for the same purpose. The idea is that newcomers are an unknown, awaiting evaluation, which is different from a long-time user with whom you've had many diverse interactions, and thus is (relatively) a known quantity. Right now the system treats new users in the same way as it treats long-time users that one has actively decided to ignore.

Back to "slow days." Some server log statistics would be helpful. Question: are low-post days also low-read days? Or do people visit the site just to read things, without posting new items of their own or re-recommending someone else's post/meme? It would be nice to have a catalog of "typical" scenarios for a user's interaction with the system. The statistics from 2003 suggest that weekends are slower, overall, than weekdays, but it does not distinguish between reading and posting. Perhaps a count of posts by day-of-week would be insightful; we could compare it to the overall activity figures.

If some days are low-post but not low-read, then it makes sense for a newcomer to try to "time" posts to take advantage of the lull. Is there any "harm" in doing so? Does it seem unfair? Should the system make it easier to do, perhaps by automating it through an "embargoed for release until" feature, as with some press releases?