Sunday, August 17, 2008

Buzz Comix is Bogus

It's hard to tell whether Buzz Comix and similar voting sites hurt or help a comic. I lean toward the view that anything that appears to be one thing but is actually another is probably harmful, or at least, is not worth the bother.

Buzz is one of several "voting list" sites online, and recently completed a major overhaul. The new site is graphically wrenching and confusing to navigate.

Regardless of the design, the site is all about the voting standings. Like other voting sites, such as Zuda, people have found ways to game the system.

The violations of democracy are quite open, and I joined in one for a week as part of my test drive of the site to see if I could level the playing field for my comic. I joined a vote-trading group of five people, and we all (presumably) voted for each other every day.

With the added help of some friends, I was able to work my way from about 150 to about 106 over a week, before the daily votes for those ahead of me exceeded what I could muster, and I began to slip. At that point I would have to add at least five dependable voters just to keep even, and ten or more to crack the top 100, and though I wanted to reach a higher rank and report what sort of traffic I got, it simply wasn't worth the effort.

The other way the system is gamed is with voting incentives. A comic urges readers to vote, and in return they get to see a preview page, get some animated wallpaper, or even a mere sketch. Ninety-percent of the comics in the top ten on a recent day were offering these premiums, and the one who wasn't told me he used to do it and recommended I do it if I wanted to be in the top ten, but decided a while back he didn't have time.

So what does Buzz do for a webcomic's traffic? My allies during my ten days or so of testing were scattered from ranks 60 to 210. Some of the higher ranked members of the group compared analytics data, and no one could find any trace of traffic originating from Buzz.

I assume that some people are getting traffic or they wouldn't participate. Because Buzz presents itself as a legitimate polling of comic readers, visitors will naturally inspect the top ten, who have the privilege of having their own banner ads showing constantly beneath their names. While I don't think a single webcomic author in the top ten is trying to deceive anyone in any way, I question this reading of their popularity relative to other participants. 

Traffic data -- which is sometimes off -- reveals that things are out of whack. Big name comics that are overtaken by low-traffic titles. Here is data, collected from the usual sources. First number is voting rank, number after the name is approximate average daily visitors:

Buzz Comix

1 - Phoenix Requiem - 77K
2 - Misfile - 222K
3 - Fey Winds - lacks own site
4 - Multiplex - 35K
5 - Earthsong - 6K
6 - Goblins - 8K
7 - Project ROL - under 1K
8 - Shadowgirls - 16K
9 - Joe Loves Crappy Movies - 22K
10 The Gods of Arr-Kelaan - 10K

Top Web Comics
1- Phoenix Requiem - 77K
2 - Two Kinds - 6K
3 - Twisted Kaiju Theater - 5K - 10K+ (spans multiple sites)
4 - Girl Genius - 194K
5 - Girly - 44K
6 - Flipside - 6K
7 - Multiplex - 35K
8 - The Noob - 56K
9 - Cru the Dwarf*- lacks own site
10 - Earthsong - 6K

The top ten obviously come from premiums and reader motivation. 

If Buzz and TWC want to truly reflect popularity, I think they'd have to eliminate voting in favor of crawling analytics data on traffic. This would make them into something so different from what they are that it's a start-from-scratch proposition. And do we really need Nielson ratings for webcomics? Such ratings traditionally benefit people other than creators.

A few problems would arise. Smaller, newer comics would have no data unless they shared their private analytics. User participation would be effectively wiped out, though I'm not sure a lot of people spend time voting for comics unless they have a stake.

The worst issue is it would further entrench the most successful comics as the objects of attention of potential readers. The most popular comics are not the best comics: they are a mix of best, best-connected, luckiest, most broadly appealing, most associated with pop culture themes, best in a genre, award winners, sexy/verboten, gimmicky and/or easy to read.

Possible counterweights include letting members vote for their ten favorites, with revisions possible. This might bring attention to up-and-comers. The list could also be sliced into genres, as it is on Buzz. (Buzz's genre list may be troubled -- comics that should appear on it by my calculations do not, and owner Andrew Gomez did not respond to my email). When there is a main list, however, genre lists get secondary attention.

In the end, what appears to be important is your top ten, rather than what a site pretends is the top ten. A site that connects people based on their top five or ten might be interesting in that you could see what people who overlap with you are reading that you're not.
(I do note that the people-who-liked-this-liked-that list on Piperka, still listed as experimental, hasn't attracted much enthusiasm.)

I can't say what the best approach to voting sites should be. I can say what the best approach to fake voting sites should be: stop pretending to be what you're not, or fix the system.