Category Archives: long tail

WSJ/HBR: A Manufactured Long Tail Debate?

A friend of mine forwarded Lee Gomes’s WSJ article questioning Chris Anderson‘s Long Tail theory, referencing a recent Harvard Business Review article by Anita Elberse.

Setting aside Lee’s original Long Tail skepticism, possibly because it suggested the blogosphere presented an opportunity to impact journalists at the head (e.g. WSJ columnists?); I emailed the following thoughts to my friend. This feels like an attempt by Anita to manufacture some Long Tail controversy to benefit her research/writing, much of her observations actually match what Chris suggested in The Long Tail.

Did I miss something? Any thoughts?

From: Dan Rua []
Sent: Wednesday, July 02, 2008 10:19 AM
Subject: RE: Long Tail questioned
Great article. I read Chris Anderson’s response earlier this week. I felt he was overly gracious about her attack on the Long Tail, but he did share some nuggets such as: “So in the data she cites, the head of the online music market represents 32% of the all plays, and the tail represents 68%. That’s certainly no challenge to the Long Tail theory; indeed, it’s even more tail-heavy than the data I cited in my book (probably because I used a more generous estimate of 50,000 tracks for Wal-Mart’s inventory).”

I’d share that Lee’s WSJ summation doesn’t match what I got out of reading Long Tail: “The Web is clearly changing cultural consumption patterns, but those changes don’t seem to involve the sort of drastic flattening of demand curves predicted by the Long Tail.”

I may have missed Long Tail’s punchline, but I didn’t read the Long Tail to predict a drastic flattening of demand curves or that hits wouldn’t continue to be valuable. Rather, I read it to admit a head and tail exists, but the historically inaccessible (due to physical costs etc.) long-tail can become lucrative because of online zero/low costs of manufacturing/distribution. Google’s $5B+ long-tail adsense business is a very concrete example of that opportunity.

It’s interesting that the opening “aha” example in the HBR article focuses on physical goods/expense decision-making: Grand Central Publishing spent $7M marketing their top book titles and $650K marketing their other titles, and the top titles were more profitable. That example doesn’t even match Long Tail theory which would have suggested spending almost zero marketing the long-tail, but making it available digitally for consumers to find/discover.

Despite the HBR article’s attempt at Long Tail controversy, it’s nice to see that it ends with advice matching what I got out of The Long Tail in the first place. For example: “1. If the goal is to cater to your heavy customers, broaden your assortment with more niche products. 2. Strictly manage the costs of offering products that will rarely sell. If possible, use online networks to construct creative models in which you incur no costs unless the customer actually initiates a transaction.”

I think the combination of Long Tail and this article suggests: 1) there is opportunity in the head and 2) there is opportunity in the tail. I’ll buy that.


UPDATE 07/03: Erik over at TechCrunch came away with a very similar analysis shortly after mine. In fact, his summation sounds strangely familiar:

“In the end, Elberse presents a false dichotomy. The choice is not head or tail. It’s both.”

Related posts via Techmeme: Matt Rosoff, Michael Masnick, Matt Asay, Alan Patrick, Jackson West, Slashdot, Coolfer, David Utter
Related images: the long tail, chris anderson, anita elberse, lee gomes, wall street journal

GuessNow: Can Predictive Markets Be Fun?

I’ve always been intrigued by the potential of predictive markets: speculative markets created for the purpose of making predictions. That’s why I was excited when Delray Beach, FL-based GuessNow contacted me about sponsoring an FVB review via PPP Direct. It was also timely because I’m following up my completion of Chris Anderson’s The Long Tail, with James Surowiecki’s The Wisdom of Crowds.

The broad idea behind predictive markets is that large populations of people, who stand to benefit from accurate predictions, can become an engine for predicting future events. If you read the wikipedia article I linked above, you’ll find that there is some controversy around the accuracy of results and various approaches to received optimal results. has an interesting management team, with John Ferber leading the charge as CEO. John, with his brother Scott, previously founded before selling it to AOL in 2004. I like serial entrepreneurs. I like the connection of advertising minds to predictive markets, because I’ve seen too many companies pursue “cool ideas” like artificial intelligence, behavioral modelling or predictive markets with business models as a secondary concern. I also like that the site feels a bit more engaging/fun than you might expect from a predictive market.

Turning to the site itself, I thought John’s incentive system was interesting. Instead of a pure stock market type of system with various prices for different outcomes, GN implemented a point system. Specifically, users can earn points for answering questions correctly (more points for correct, fewer points for incorrect), answering questions early (more points for early, fewer points for later), and avoiding group think (more points for correct answers going against the crowd). They also have a bonus point system for site participation and advertiser offers, but I don’t entirely understand the “bonus” section of the site — that section feels more like rewarding site behavior and CPA advertising than predictive.

Points are then redeemed for cash, according to a “point value” decided by the total Prize Pool for a month divided by the total number of points awarded in that month. For example, if 500,000 points are awarded in a month with a $5,000 Prize Pool, then each point is worth $.01. If you earned 1,000 points that month, then your points are worth $10.00. I believe a similar calculation happens for the Bonus Prize Pool and bonus points.

They have a good set of questions, including topic areas such as:

Some of the questions I’ve answered include:

The model is pretty flexible. In addition to predictive questions, I also noticed trivia-type questions (e.g. “name the state that…”) and survey-type questions (see hybrid car question above). It’s not clear these are necessary to keep people engaged for predictions, but I can see them opening monetization options.

A few of my suggestions include:
1) I loved some of the higher level data concepts such as accuracy ratios, friction and confidence levels — find ways to share that data and reward publicly on these;
2) I know the “Shocking New Video” ads are probably prompted/related to your Miss Internet Pageant 2007, but they could be a tad risque for the diverse demographic good predictions will require; and
3) I may have missed it, but I couldn’t find where to compare past group predictions with past actual results — that is the question everyone has about such markets and there has to be some data you can share, probably great linkbait.

And, lastly, I’d be remiss if I didn’t mention GN’s points-based affiliate program and blog. I don’t know GN’s funding status, but this review has prompted me to dig a little deeper. Thanks for reaching out to me guys!

(sponsored post)