Revealed: The Declining LGF Registration Rate
In the past, The Boiler Room Crew has provided accurate “signs of the LGF decline” statistics, including commenting rates, unique nic participation, individual nic stats, and karma levels. These things are relatively easy to glean, since the data is sitting right there in plain view on the site. All one has to do is figure out a way to efficiently grab it and stick it into a spreadsheet, then activate the various filters for analysis (for the most part).
Now, I know we’ve noted it before, but the engineers of The Boiler Room thought we’d do a little more work to see just how poorly Johnson’s doing in another department: recruitment of new accounts (dubbed “hatchlings”). After all, this is a great gauge of the general popularity and community aspect of the blog (the site that was, not so long ago, a strong contender for “best online community”). In light of Johnson’s political 180° turn a few years ago, has he been able to attract a new base of lizards? Just how many folks are joining LGF in recent months/years, and how does that number compare to the old days?
For our purposes, this project was much more of a challenge. Unlike the commenting stats, the info we needed isn’t in plain sight. Sure, we could use an alternative method (see graph, left) to get a general idea, but that wasn’t giving us the true number. We thought that the info we needed was buried somewhere on the dashboard of the LGF blog engine, and virtually inaccessible to us, whether you’re logged in or not. What we discovered though, is that Johnson’s blog coding actually left the clues right under our noses, like unknowingly sailing over a shipwreck every day. (Some of the pieces we already had, and didn’t realize it!) Once the proverbial light bulb appeared, we grabbed as many strawberries as we could (over 2,500, as it turned out), and began to plug the figures in. Even if not unexpected, our findings are a bit shocking…and not in a good way for Mr. @Lizardoid.
First though, I’d like to quickly revisit a time back when Johnson was a wingnut, and registration windows were short. When the announcement came and they were opened, doors to the kingdom generally got busted. In those days, Johnson would usually follow up and tell everyone how successful it was (the following is a typical example, from July ’07):
Other info about registration was volunteered in the comment section by CJ in more popular days, as summarized by this gif montage:
But eventually, we assumed that the rate began to slow considerably, and as that happened, he just stopped talking about it. In fact, the last time CJ even mentioned the word “hatchling” was in November of 2010 (article #37494), when we saw this:
(A “sock”, for those who aren’t familiar, is an account which has characteristics of belonging to someone who has previously registered under another nickname. These days, a large percentage of new accounts are socks).
That last hatchling count indicates quite a drop off from those previous examples, but to some may seem still somewhat respectable. Going forward, however, we didn’t have Johnson’s report to work with, so we were kinda out of luck. He’d post up an open registration thread, but wouldn’t leave any clue as to how many were joining.
To give a full visual of Johnson’s announcements over the years, everything was plotted, and Engineer No. 5 presents and explains the timeline of those open registration windows and hatchling counts (or lack thereof):
Chart #1 – Plot (black lines) of the final registration counts for announced registration openings. These windows were typically 30-60 minutes long. The short red lines just mark when CJ announced an open registration, but failed to publish the count afterward – his method of hiding the decline.
Indeed, over the last handful of windows of the timeline, Johnson was actively “selling” the benefits of an LGF account, in an attempt to attract newbies. But how can we calculate a recent hatchling rate, if CJ’s too embarrassed to tell us how many actually signed up? Hence, the need for that hidden data.
Results of the 2nd Great Strawberry Heist
Well, luckily we have some smart and resourceful people here in The Boiler Room, and those aforementioned strawberries can tell us everything. For the time being, we’re going to refrain from explaining exactly how we got all this information*. But rest assured, because we have it, we can provide the following charts and analysis with confidence. It was only a matter of putting it all together. For that, we turn back to Engineer No. 5:
Chart #2 – A December calendar showing the open LGF registration windows… Green = Open, Gray = Closed (or unknown – I had a couple of glitches). For the 40 days shown, registration was open for a total of at least 282 hours (11 days, 18 hours).
I highlighted the 3 week period that we have registration information for. For that period, registration was open for at least 108 hours (4.5 days) during which CJ got at most 7 legit (non-sock) registrations.
#449 Charles 4/3/2008 18:10:48When I don’t announce open registration, it’s amazing how many fewer people notice it. It’s been open for almost 4 hours today, and less than 25 signups.
Let’s extrapolate that one. Back then, an unannounced open registration snagged about 6 regs/hour, or 1 every 10 minutes.
Flash forward to 2011. A 3 week series of unannounced open registrations snagged 7 legit regs in 108 hours, or 1 every 926 minutes.
Here’s where it gets really interesting. The data we grabbed allows us to plot the registration activity at LGF over the years:
This is the new accounts/month timeline. Notice that there hasn’t been significant activity in quite some time.
This is the timeline for total registration counts. Notice the flattening line over the past few years.
Bottom line, if you take this data and average it out by year, this is what it looks like.
Last we checked, there were 35,578 total user accounts registered since it was implemented in 2004 (we won’t mention how many are banned/blocked, as that is for another thread, and of course the number who are actively commenting these days is less than 1% of that). But check the “Grim Milestone” portion of the second graph: starting in mid-2007, the time to add the next 3,000 accounts keeps taking longer and longer, to the point where we really don’t know when (or if) they’ll get to the 36,000 mark.
Here’s a summary of the fun facts:
- LGF had about 1250 registrations during 18 hours on the first day it was open (6/15/2004).
- LGF has had about 1250 registrations during the 16 months from Sept 2010 and to the present.
- The number of new LGF registrations has decreased by at least 30% every year since 2007.
- The 4th year of decline, 2011, saw a whopping 67% fewer signups than 2010.
- During the period 2007-2011, the LGF registration rate declined by 92%.
- In 2007, open registrations typically netted 150-200 signups per hour. (The “announced counts” chart confirms that.)
- In 2011, perpetually** open registration netted about 50 signups per month.
It’s no wonder why Charles now keeps unannounced registration open so often, tries to boost signups with occasional announced “sales pitch”, and doesn’t mention how many hatchlings he gets anymore. The takers are scarce these days.
What’s more, the accounts that do sign up are socks, with increasing likelihood. Take the aforementioned 3 week span in December that we tracked as an example. There were a grand total of 11 accounts created, and we know that at least 4 of them were socks (which means that the real number could be much higher).
What’s not forgotten is the fact that the site currently has more bells and whistles than it ever has in the past: LGF spy, LGF Pages, auto-twitter updates, avatars, etc., plus at least one video posted per day… yet the rate still spirals.