This is a follow-up to
a diary I wrote a couple of weeks ago asking people when they'd registered. I got a very enthusiastic response, and thought people might like to see the final results. There are some interesting (to me, anyway) patterns in the results. In addition, it looks like the effects of YKos may be having an effect in bringing in new users, even after the publicity has died down.
The gory details are below the fold.
First, the overall roadmap:
I'd like to thank all of the people (roughly 200) who dug up their registration info. A special thanks to jotter, for the daily high-impact-diary series. The user-ID information in those diaries is but a tiny fraction of the goodness contained therein.
So, what does this graph tell us? dKos has had three (so far) overall growth periods. In the initial period, starting back in October 2003, people registered at the rate of approximately 50 per day. Growth stayed pretty consistent at this rate until the fall of 2004, and the run-up to the Presidential election.
During the height of the election season, with people's minds focused on politics, dkos grew fast. 260 people/day fast. Or, put another way, it took almost a year to get the first 15,000 registered users. The second 15,000 took roughly two months.
The third phase has lasted from November 2004 to the first YearlyKos, just a few weeks ago. During phase three, growth has been pretty constant at 100 per day. Note that even after the burst of people who came in with the election, the long-term trend is still double the rate pre-election.
This brings up an obvious question: are we on the cusp of another surge in growth? It's likely that there will be another spike in new arrivals once election season really begins to heat up in a few months, but has a new surge already arrive? The next graph says "maybe yes":
This is the last six weeks worth of registrants, along with the daily registration rate. First off, it's pretty clear that the publicity surrounding YKos caused a massive spike in registrations during the week of the convention. What's less apparent at first glimpse is that, now that the publicity has died down, registration rates are still higher than they were pre-convention. The two blue lines show this; the lower of the two is the average rate of registration for the month prior to YK (102 people/day). The upper is the post-YK rate (140/day). Is this significant? I don't know; ask me in a month. The early signs are, however, that it is. The actual spike from the conference publicity has clearly died down, and it looks like we've reached a new equilibrium, a fourth growth phase.
By the way, if this rate holds, the 100,000 mark will be reached 43 days from now. Just 6 weeks.
I'm very curious to see whether the recent publicity will cause yet another spike. Time will tell.
One last thing: When I said there were three phases, I skipped the "zeroth phase". That's the early days of the Scoop version of this site; during the changeover, all of the users of the old version registered all at once. As you can see from this last graph, the initial registration rate was quite impressive:
So, if you see somebody with a UID under about 2500 or so, that means they registered in the first two or three days. As best we could determine, on the first day, registrations were coming in at the rate of several dozen per hour.
-dms