Umair Haque / Bubblegeneration

Thursday, December 09, 2004

Blogonomics

Om on the death - aka, relative invisibility to pro bloggers - of individual bloggers.

Here's the problem - people assume there are no switching costs in the blogosphere (ie info is cheap, there are many similar blogs, all that jazz). This is not the case at all - the truth is switching costs are high in the blogosphere. There's evidence for this: blogrolls tend to stay locked in. Once they're made, they're rarely (if ever) updated or switched.

Here's why: while reading any old blog is costless, searching for new blogs that are worth reading regularly is costly. We can call this search cost the price of a blog (there's more to the price of blog than this, but it'll do for now).

Now, once you do find a blog that meets your preferences, you tend to lock in, until (presumably) you've at least recouped the fixed investment of searching for the blog. Or, you lock in until a new blog offers you enough value to switch, which is the value of the first blog, plus any fixed cost left.

The point is that you lock in because the opportunity cost of reading the blogs you've already found is very low - the expected value of reading new blogs is simply not that high (ie, the probability that you'll find a new one as good as the one you're already reading is low, because search costs are high). So the second case - where you switch - is pretty rare.

Now, we can complicate this model with regret effects - the mental costs of dissonance, or processing new information that disagrees with information you already believe - which raises switching costs even further. But there's no need to do so at this point - this model already explains the incredibly lengthy time decay of the average blogroll.

What this means is that reputation effects are a key market structure of the blogosphere - because they slash search costs. You might not read bubblegen if only Om mentions it - but if Om, Battelle, Wired, and BoingBoing mention it, you'll probably give it a chance. The probability of this being a good blog (and thus expected value) jumps discontinuously - not linearly - with each repuational signal the market sends you.

But the odds of this happening are pretty low - even if your blog is great, sending this many signals is combinatorially complex. It doesn't (and won't) happen often. What all this really means is that the blogosphere will ossify - it will get and stay locked into winner-take-all dynamics. Especially as the number of bloggers increases massively more than the number of readers.

So what we really need, I've thought for a long time, are new mechanisms to create churn in the blogosphere - not simply positive feedback ones like Blogdex and Technorati (and traditional media, blog awards, advertising, etc), but things like reverse syndication - and newer ones we haven't thought of yet. These are search cost slashers - or expected value multipliers, if you like. It's an absolutely huge gap in the blogosphere just waiting to be plugged.

If these mechanisms aren't created, blogonomics dictates that pro bloggers will be read more than personal bloggers - because they're more likely to generate reputation effects, and take advantage of positive feedback mechs. Believe it or not, this was part of the motivation for Blogversations.

-- umair // 1:08 PM // 5 comments

Comments:

All i know is that i found this blog via fred wilson's blog and i'm so damn glad i did. You do an amazing job offering incredible insight into important issues. You have won me over for sure.

# //

eric goldstein // 4:08 AM

Have you looked at findory.com ?

# //

Murray // 3:49 AM

Nice site!
[url=http://pvyxuvfy.com/bjey/vqhx.html]My homepage[/url] | [url=http://rlutcque.com/ugwq/nrol.html]Cool site[/url]

# //

Anonymous // 8:18 AM

Thank you!
http://pvyxuvfy.com/bjey/vqhx.html | http://mdliyxyh.com/niyk/jcyp.html

# //

Anonymous // 8:20 AM

Practically prescient given today's hyper-emphasis on the reputation of "referrals" by the major search engines. A year and a half later and you're right on with this post.

# //

Melanie Phung // 10:47 PM