Slashdot Mirror
Doubts Multiply About the "Long Tail"
fruey sends in a New Scientist analysis of the many second thoughts about the Long Tail theory. It summarizes four studies that show, in different markets, that the tail is both flatter and thinner than originally supposed, and that blockbusters are not going away in those markets — they are getting bigger. It's theorized that widely used collaborative filtering software is magnifying the winners' share of the various pies, and peer influence is a large contributor to consumer behavior.
I stopped buying CDs when the music companies started sueing their own customers.
:-)
So I'm not even part of the fat root
If I had an Ass, I'd call it Fanny Bottom, then I could slap my Ass; Fanny Bottom, on the Arse.
What is the long tail? The summary, and TFA (I skimmed it so maybe I missed this) seem to indicate that the long tail theory means the more obscure stuff will be more popular. I thought it simply meant that you could make money off the obscure stuff when your distribution costs went to zero (because of the Internet). Am I missing something, or does the article interpret the idea of the long tail incorrectly?
This post cannot be rebroadcast without the express written constent of Major League Baseball.
OTOH, it is clear that the way to make a lot of money as a retailer is to have the popular options at a low price, a la Wal Mart. As the article suggests, though, have a few niche items can make money, if you can get people to pay for therm.
Here is what I think. Stocking an item, even a virtual item, incurs a cost. So if one stocks a million different items, and only one thousand different items moves a week, then there are costs not being covered on an even yearly basis. One way to make this work it mark up items a lot, as in the used book store, wher a book can be bought for a dime and sold for a dollar.
To me the long tail means either selling in a high markup niche market, or having the ability to control costs so you can sell less of more different things and actually draw a profit. In either case, the long tail in going to be chanllenged in times that force are going to force some fiscal responsibility, and people are going to focus on necessities and cheap luxuries.
"She's a scientist and a lesbian. She's not going to let it slide." Orphan Black
There's a certain segment of the population that enjoys finding obscure stuff. I'm not sure of the size, but I'd guess it's about the same size as the introverted segment of the population, around 25%, as the two behaviors are somewhat correlated. (i.e. folks who actively stray from the herd socially are more likely to express interest in consumables that are different from the herd's preferences) So, given that assumption, the drivers of a long tail market, "funky stuff seekers," could be overpowered by the general population.
On top of that, it's my guess that researchers in this area, as researchers in most areas, tend to be "funky stuff seekers" themselves. (I mean, it's their job to search and speculate on the edge of their field of study.) So, right off the bat, there's a bit of inherent bias in interpreting the effect of their cohort on the market. In other words, they're following the non-herd herd. :)
The habit (of wanting what everybody else wants) is not waning as selection and availability increase. The habit is reinforced by the increasing channels by which we can see what everybody else wants and adjust our own wants to that.
The Internet, while logistically making the long tail feasible, is socially making blockbusters bigger.
Go green: turn off your refrigerator.
... also known (in a simplified form) as "Zipf's Law". The most popular word occurs twice as often as the second popular, three times the third popular, four times the fourth etc.pp.
If you look at the 10% most popular words in a table, you will notice that the absolute share of the upper 10% increases, if you count longer and longer texts with more and more individual words.
With 10 different words you get, that the first 10% make up ~34% of all words.
With 100 different words you get already ~56,5% share for the first 10% of words.
With 1000 words we have 69,2% share for the first 10%, with 10000 we come to ~76,5% share.
The idea was not only that the internet would enable sellers to tap into this market more effectively (through lower costs) but also that this would in turn lead to the tail becoming yet larger (both because increasing number of items sold means more potential sales, and because these rarer items may pick up additional momentum and sell more copies, thus giving more events with higher frequencies) as compared to the "blockbuster" events, which suffer due to increased choice.
More specifically, the issue is not "stuff one person will buy won't outsell stuff millions of people will buy" but rather "will many items purchased by a few people exceed the sales of a few items purchased by many people", given improved choice. The long tail hypothesis as presented in the Wired article says yes (or, at least, that the few-selling items will gain ground in the scenario presented by the internet) but evidence in this article says no. Perhaps still not the most shocking discovery in the world, but certainly a bit more than the trite discovery you make it out to be.
Also, as an aside, IQ is a very bad choice to illustrate the "long tail" because IQ is by definition normally distributed, and thus has a relatively short tail, as such things go - a long tail distribution is typically characterised by a significant slowdown in falloff as you move further from the median.
I'm not sure how convincing the article's evidences are. For example, a study that says 0.4% of an online catalog are responsible for 80% of sales sounds damning. But where is the traditional sales outlet with 13 million items for us to compare it to?
I believe Wal-Mart stocks about 5000 CDs, or about 50,000 songs. That's roughly equivalent to the 52000 songs that constituted 80% of the online retailer's sales. But in Wal-Mart's case, 50,000 songs generate 100% of sales. So the dredges of that online catalog do generate a large number of sales.
If a song has a, say, 98% chance of only generating two or three plays (and zero sales), it's probably worthwhile to keep it in the catalog, just in case it breaks out. But to me, such numbers don't show that the site carries 98% crap. It's more likely that the site just isn't making the songs discoverable enough, or that they're not providing much incentive to explore.
If I were running such a site, and I wanted to get a better understanding of the hidden side of my catalog, I might offer random shoppers the opportunity to listen to five random-but-obscure songs (perhaps filtered by genre), rate them, and get a free download in return. They could also buy the songs at a discount if they liked it enough. That might give me more knowledge, which I could use to make better recommendations.
There might also be a way to stagger prices so that sales-free songs go for trivial sums ($.10 for the first purchaser, $.12 for the second, etc.), just to encourage dumpster diving.
The point is, the tools can be crafted so as to encourage the Long Tail phenomenon, or to undermine it. If the site doesn't have at least a "listen to something totally random" button, don't expect much out of the tail.
You want the truthiness? You can't handle the truthiness!
Yea no. That's better than say Disney and their offensive attempts at disabling parts of your remote during their previews, but it's still a far cry from ideal.
Remember when DVDs first came out? You could watch previews, if you actively looked for them on the disc. They were often sequestered in the special features menu. Which is ideal because trailers for future movies have a shelf life of a month or two. Whereas the movie you purchased is meant to last for years. That was the whole point of buying it in the first place. Once DVDs got a foothold in the market companies began forcing previews down your throat. They don't care that you really like this movie and intend on watching it multiple times over the years. They just want to make sure you see the new product they have coming out.
The new product which, if you're at all of discerning taste, is likely to be garbage. Garbage that parades around in front of you again and again each time you try to play your beloved movie. Anymore I just rip movies to my media server immediately after I get them. Then the next time I want to play the movie, it just starts. It starts up and I'm just there, in the movie. The whole point of the exercise in the first place. Every once in a while I just place the disc in the tray and play it the old fashioned way. And often I'm amazed how much useless bullshit I have to flip through to get to the fucking meat.
You can call it nitpicking. It's a very small amount of your time they're wasting. Any particular instance of it is insignificant. But think about the aggregate amount of time and energy wasted on such things over the course of the lifetimes of all consumers. For something that moves as fast as the modern world does, those little moments are more precious than people tend to realize. Especially when it is something that ultimately benefits neither party (spent, worthlessly old advertising). I paid you. Stop fucking with me and let me enjoy my purchase.