Monday, January 24

Incentivizing Behavior: How Algorithms Kill Good Ideas

Opinions
The newest bright and shiny object at the center of social network news is undoubtedly Quora, which is a question and answer network that reached 164,00 unique visitors in December (and likely doubled in January). But all that might be for naught.

According to TechCrunch, Quora wants to develop an algorithm to measure online rank and user quality. Good luck with that. It's a mistake.

While algorithms can be useful, they also diminish the overall quality and value of any network where they are applied. Google understands this about human interaction. It's one of the reasons the popular search engine recently announced a harder look at search engine spam, beyond the changes that took place last May.

Much like photons and electrons in quantum mechanics, people behave differently when they are measured.

If you are familiar with quantum mechanics or wave-particle duality or the uncertainty principle, you already know they behave differently when they are observed and/or measured. People behave differently when they are observed and/or measured too.

And, the more people know about how they are measured, the more likely they will be influenced by the measurement. Even one of the first scoring systems widely adopted by marketers influenced their behavior as participants. As soon as Todd And's Power 150 was adopted, participants focused on improving the measures. Back then, it included things like Google Page Rank, Bloglines, and Technorati.

Since the so-called Power 150 move over to Advertising Age in 2007, it has undergone several overhauls, some good and some not so good. The new measures now include PostRank, Yahoo InLinks, Alexa Points, and Collective Intellect.

hamsterThroughout all those changes, an interesting phenomenon occurred with every addition and deletion: The usual list leaders would tumble down, sometimes as many as 30 places, before rising back up to their relative places. It's no secret why this occurred. It demonstrates a tactical shift in which measurements the list leaders concentrated tactical efforts.

The same phenomenon is occurring on Twitter. As Klout is adopted, participants adjust to accommodate the new measure of incentivized frequency because more tweets equal more Klout.

It also equals more spam, more auto messaging systems, more paid and fake followers, more conformity to popular opinion, and more time on the platform. Mike Judge already predicted the outcome of this path.

Why adding algorithms to networks changes the nature of the networks.

Quora originally started as a network where people ask and answer questions. The primary objective was to create a database of information. As such, the participants mostly attempted to provide the best possible answer to the specific questions being asked.

QuoraAs Quora adopts a ranking system, the objective will fundamentally change. Instead of answering a few questions by providing the best possible answers, many will be predisposed to answer more questions with whatever information is likely to generate the highest number of up votes. These lead to a fundamentally different answer.

As David Armano, senior vice president at Edelman Digital, pointed out after I questioned his post that placed even more qualitative merit on up votes: Quora isn't a database of information. He defines Quora as a marketplace of opinions.

Coincidentally, the supply of opinions outpaced the demand for the action some years ago. Mostly because, as a commodity, opinions are the only limitless product in a world searching for something increasingly rare. One good idea that works.
blog comments powered by Disqus
 

Blog Archive

Google+ Followers

by Rich Becker Copyright © 2010 Designed by Bie Blogger Template