Another Fiddly Widget

Okay, this one plays music:

Yeah, it’s music I put together (except for “I Feel Loved” which is a remix of a Depeche Mode Song).

Let me know if the player presents issues. This is all for figuring out what stuff is useful/interesting on the site, of course.

“Big Posts” — How They Work

(Warning: Blog geekery in the extreme follows)

From time to time I’m asked how I’ve grown my readership here on the Whatever, and whether anyone else can do those same things to grow the readership of their own blogs. In the main it comes down to four separate factors, three of which I or anyone can control, and one I or anyone else cannot. The three factors that you can control are these:

1. Update frequency: Updating daily matters in terms of readership.
2. Enabling comments: People who comment feel attached to the site; people who don’t comment get updated content when they click through.
3. Quality of content: Putting in interesting stuff so people have a reason to click into the site daily.

I’m not going to talk about any of these in length right now. What I’m going to talk about is the factor you can’t control: Big Posts.

A big post, very simply, is a post that more than the usual number of people link to, thus bringing in an entirely different audience of readers. Most of these readers will be one-time readers — they click through to the link, see it, and click out, never to return — but some small proportion will root around, enjoy what they see (due to you working on the factors you can control), and put you on their daily reading list. Bang, you’ve got new readers.

Big posts can happen when one or more of the following conditions exist:

1. You write or create something unusually well-written about a current news event or other hot topic.
2. You do something unusually stupid and/or funny on your site.
3. You are linked to by one or more high-traffic sites (Fark, Slashdot, Digg, Boing Boing, Instapundit, Daily Kos, etc).

The first two of these will often have the effect of making your already-existing readers link to that particular entry, bringing in their native readers, some of whom may then also link and comment on their sites. This will result in a lot of links, with (probably) a relatively small number of people coming in through each link, but making up a large number of new readers in aggregate. I call this the “LiveJournal Effect,” because LiveJournal, thanks to its “friends lists” and community structure, is particularly good at creating cascades of many links among friends — you can get several generations of links from this sort of thing. LiveJournal is of course not the only place from which this can happen: MySpace can create a similar effect, as can informal online communities not affiliated with any particular blogging software — for example, my own community of science fiction writers and readers with blogs.

The third condition, of course, moves large numbers of people primarily from one or two sites. This phenomenon is well known — people talk of Web sites collapsing under the load after being “Farked” or “Slashdotted” or having an “Instalanche.” Very often, getting linked to a high-traffic site will also start an LJ Effect, so not only do you get the original flood of people from a high traffic site, you’ll also get a second wave of visitors from personal sites linking into the entry.

If we know the conditions that cause Big Posts, then why can’t we control them? Primarily because it’s not up to the original poster to decide what’s a “Big Post”: that comes from others. In the case of the LJ effect, a whole bunch of people have to decide the entry is worth linking to; in the case of the high-traffic sites, whoever is reponsible for putting posts on the front page of the site — usually a single person or a small group — gets to decide.

You can lobby some of these sites to pick up on a post you think it interesting (most of the high-traffic sites welcome link submissions, formally or informally), but there’s no assurance that the people running the site will agree. For example, I recently submitted a link to Boing Boing about my “Schadenfreude Pie,” because I figured it was a Boing Boing-y sort of thing. It’s not linked to on Boing Boing, however. I’ve queried Instapundit for links before and come up short as well. That’s the way it goes. However, Boing Boing and Instapundit have also linked to my site when when I didn’t solicit them for links. Which goes to my point: You never can tell.

(And of course, even if a high-traffic site links to you, it’s not a guarantee that you’ll get a rush of traffic, because there’s another group of decision-makers involved: The site’s readers, who have to decide whether the link seems interesting to them. This is necessarily also the case with an LJ Effect as well. In many respects a “Big Post” is the ultimate in the “Long Tail” phenomenon — it relies on many people making the decision to link, many people deciding to click, and then on another group deciding to link as well.)

I do think it’s possible to game the “Big Post” system somewhat, and generate a large amount of traffic, but I don’t think gaming the system will result in the desired effect of growing a site’s overall readership. As a hypothetical, I guarantee you that if I posted a picture of my wife naked (that’s the vulgar “naked,” as opposed to the arty “nude”), that this site would get (ahem) a fairly substantial traffic spike, both for the initial picture and then for the aftereffect of various people discussing what a scumbag I was for posting a picture of my naked wife purely to get a traffic spike for my site. But aside from my wife deservedly beating the crap out of me for posting that picture, I suspect that in the long run there would be no lasting effect, and that the site wouldn’t gain any new readers; the sort of people who are going to click through for a picture of my naked wife are not the sort who are likely to stick around for the usual bill of fare around here, and the people who are discussing what a scumbag I am are not likely to stick aound either. Indeed, in the long run I would suspect I would suffer a loss of readers, as some of the current regulars, disgusted by my actions, take off for parts unknown, never to be seen again. So gaming the system is not a good thing. Also, there will be no naked pictures of my wife. Sorry.

So that’s the theory behind a “Big Post.” How does it work in the real world? As it happens, I have two good examples of how a Big Post works, and how it affects your overall readership. Let’s start with one from September of 2005: My “Being Poor” post.

As of August of 2005, the Whatever was getting between 8,000 and 10,000 unique visits per day on weekdays (somewhat lower on weekends). 10,000 was the upper bound. The “Being Poor” piece was posted in September 3. For this piece, I didn’t request any links from high traffic sites; I just put it up and people began to link to it (NB: there was some initially higher traffic on 9/2 for a piece I’d written the day before). What you’re seeing here is the readership curve of a “LiveJournal Effect” — which is to say that while there’s a day where the readership spikes (September 7), there are a few days before and after where the readership of the site is significantly higher — the curve encompasses six days, from 9/3 through 9/9. This is the effect of new links being added as the piece filters through the blogosphere. After 9/10, readership returns to a lower readership plateau — but that plateau is higher than the previous readership level. Whereas in August of 2005, the readership was between 8k-10k, after the “Being Poor” piece the readership is between 10k and 15k, and never drops below 10k. What was previously the upper bound of the readership became the new lower bound.

That’s a “LJ Effect” curve. Now let’s look at how getting “Farked” can create a spike, as it did this month:

For August and the first nine days of September of 2006, the Whatever was averaging between 15,000 and 20,000 unique visits per weekday (and again, a bit lower on weekends). On September 10 through 12, the site experienced an “LJ Effect” when people started linking in regarding the entry in which I discussed my wife backing up a grabby drunk — and indeed 9/11 was the highest trafficked day the site had had up to then, with 26.6K unique visits (NB: Instapundit linked into all this, but his site didn’t drive most of the traffic — evidence that even a high-traffic site won’t always push lots of folks).

The wife incident would have been interesting in itself as a traffic mover, but then I taped bacon to my cat, and things got crazy. The event was linked to by Fark, and on 9/13, the site had in excess of 67,000 visitors, most of whom who wanted to see the miracle of the bacon-taped cat. Of those 67,000 visitors, 20K were “mine” — that is to say, my regular attendance — and 30K were from Fark. That leaves 17K coming from elsewhere: another “LJ Effect” that was swamped by the Fark spike. The next day had 56K visitors, of which only 6k were from Fark; subtracting my “own” 20k, that’s another 30k readers. Some of these were from other high-traffic sites, notably Metafilter, but the majority were from smaller sites and personal blogs; Fark readers and others pushing people through.

The spike is significant, but it’s also clear that the vast majority of “spiky” readers didn’t stick around; by 9/16 the BaconCat spike was over. But once again there’s a new readership plateau — whereas before the readership bounds were 15k-20k, since the BaconCat spike the new readership bound is 20k-25k, excluding weekends (but even those are up commensurately). Some of that new readership can probably be attributed to the LJ Effect just prior to the BaconCat spike, however.

This suggests two things to me: First, that for growing a readership on your site, it’s better to have an “LJ Effect” type of event than a “Farked” or “Slashdot” event; second, that a substantive post that is widely linked — in this case, “Being Poor” — is better for growing a readership than a silly/stupid post — in this case, BaconCat. I’ll also speculate, based on other “Big Posts” I’ve had here over the years, that there’s a correlation between substantive posts receiving an “LJ Effect” and silly/stupid posts getting the “Farked” effect. In the case of Fark in particular, of course, this is nearly axiomatic, since that site specializes in linking to goofy/idiotic/asinine things. But other high-traffic sites also seem to best drive traffic when they’re linking to some sort of “stunt” post.

Bear in mind, of course, that all this speculation is based only on my own experience here on this site; I leave it to others, possibly those gunning for advanced degrees, to do a more thorough examination of LJ Effects, Fark Spikes, and their overall effect on the growth of readership on personal blogs. However, I do suspect that my experience with these phenomena is not notably unusual.