July 5th, 2014

_The Long Tail_, Chris Anderson

Subtitled: Why the Future of Business is Selling Less of More
Also: Longer A New Chapter The Long Tail of Marketing

Anderson's book is a decade old, just in time to enjoy a wave of people saying, hey, that turned out to be All Wrong. Guess what? The people saying that have an idea of what is in this book that maps very poorly to what I read in this book. Anderson has seen plenty of it, and part of the last section or so is devoted to, here are the ways that people tend to misunderstand what I am saying.

Here are some samples of Long Tail Is Wrong:


This is actually the one that convinced me to finally read the book. I had figured, when it was originally published, that I never needed to read it, because I understood how it worked, having been on a rocket launched by the long tail since the Spring of 1996. There wasn't really anything in the book that surprised me.

Shatzkin points to very self-serving piece by a digital small publisher that argues that the long tail doesn't pay because their catalog sells a few things really really well and everything else barely moves if at all. Which is actually what the long tail predicts, so you sort of have to wonder whether we read the same book.

Shatzkin also points to a piece of data from the Canadian book market which, if true, is sort of interesting. It doesn't prove anything wrong about Anderson however. (Anderson never said you couldn't find where the tail went to zero -- he just noted that it wasn't showing up yet. Perhaps Canadian book production has found the intersection of the tail with zero.)

Here is another instance of Blockbusters vs. Long Tail, with the rhetorical frame that these two somehow don't fit together.


Let's go straight to Anderson for why this frame is wrong.

"By far the most common misperception is that the Long Tail predicts the end of hits. Not so. Hits are as much a part of a powerlaw distribution as are niches. What's dead is the monopoly [he emphasizes] of the hit." And by monopoly, he means that in the Bad Old World, the powerlaw distribution got chopped off unnaturally. Now, it is not chopped off, and Clever People and Companies can make money off the long tail, where there is lots of stuff to be moved and lots of money to be made, if you can solve the signal to noise issues. "For too long hits or products intended to be hits have had the stage to themselves, because only hit-centric companies had access to the retail channel, and the retail channel only had room for bestsellers. But now blockbusters must share the stage with a million niche products, and this will lead to a very different marketplace. Let me explain."

"There are essentially three kinds of hits." Basically, real ones selected by gatekeepers, fake ones pushed by gatekeepers ("think Garfield: A Tail of Two Kitties") and bottom-up real hits (he doesn't mention 50SOG, but it really springs to mind here). And this is indeed the world in which we live. Anita Elberse is basically condemning strategy number two, when she emphasizes why people producing content for theatrical release or network TV must absolutely go big or go home: faked up hits no longer work at all. The article frames it as long tail is a myth, but that's just a bad frame. Elberse and Anderson are 100% in agreement that the second kind of hit is dead and gone, and they may even agree on what killed it, altho they probably give that phenomenon different names.

"Another common question was whether the theory meant that obscure producers could now expect to get rich. Sadly it's not as simple as that." This would be the macrame jeans or Marcello Vena being all sad and thinking it is somehow Amazon's fault that his specialty catalog is not moving well. Macrame jeans and the Vena catalog can move at least a little in the long tail world, which is more than can be said for 3rd quarter of the 20th century (well, he's in Italy, and they've always been small firm, so *shrug*). But the long tail never said there would be a lot of customers for those jeans or those books -- just that the customers and producers could now actually find each other. Specifically, Anderson emphasizes that the businesses that benefit from exploitation of the long tail are customers (who experience greater choice) and aggregators (like Amazon) who can find margin in connecting customers and producers in the long tail. Producers in the long tail are going to stay poor. "For producers, Long Tail benefits are not primarily about direct revenues. Google Ad-sense on the average blog will generate risible returns, and the average band on MySpace probably won't sell enough CDs to pay back their recording costs, much less quit their day jobs." He goes on to point out other possible benefits of microcelebrity: you get a job, people come to your gigs, etc.

There are numerous places in the book where I went, hey, that's not quite right. In the main part of the book, he actually does a terrible job on the business models appropriate to the long tail (mostly by omission), and he makes some assumptions about cultural defaults (he mentions the 30 minute TV show, but he could have as easily identified the 300 page novel, or the 45 minute album, or wtf) being arbitrary and/or imposed by previous technology and therefore likely to go away. His advice on how companies should interact with bloggers is fairly iffy (altho if I had read this when it first came out, I probably wouldn't have noticed that, because I hadn't yet hit the point where authors/developers/wtf saying, Thank You! in the comments creeped me out yet). Using Bonnie McKee in the way that he does is especially hysterical, given what happened to her career in the years since the book was first published, and the career of the band which he uses as a contrast. Both the early success of the latter and the later success of the former say more about the ages of the performers and their natural audience than anything else, and the ultimate success of both of them does speak to the greater forgiveness of our music culture vs. the Days of Vinyl.

Speaking of which, I went on a little tirade a couple nights ago about how annoying it is when people talk about producer/performer trends in music without thinking about the installed music systems (tube vs. transistor, pre fuel crisis vinyl vs. post fuel crisis vinyl, 8 track's arrival in the studio, the compact cassette, etc.) of the purchasing audience. The music industry, believe it or not, is made up of a bunch of super savvy people. They recognize that the Art You Make Had Better Sound Good When It Gets Home. My example was, if everyone has vinyl systems with tubes on them, _of course_ you're going to get non-stop walls of violins. Duh. R. pointed out that AM radio tended to reinforce that as well. Once you have transistor systems, FM radio and hi-fi systems in general, you're going to get more exploration of deeper and more detailed bass, high-hat will actually survive, etc. All of which is simultaneously relevant (technology defines the limits of the possible, and not just in terms of connecting consumers and producers) and irrelevant (Anderson was focused on a particular piece of technology, the web-as-catalog).

One of the rules Anderson did supply in the book directly addressed gatekeeper decisions: put it all out there/don't choose for the customer. Let them choose. "Rule 7: Think "and," not "or." He gives examples like more colors, more retail channels, multiple endings to a movie on DVD, more subtitled languages, more screen options, more cuts for ratings. Also, "Rule 8: Trust the market to do your job." He makes a big deal about gatekeepers/pre-filtering vs. customer preference/post-filtering. It all sounds really great, and if it's versions of a blog post or whatever, and your readers are your friends, it is Teh Awesome. But we've seen with Kickstarter projects that forgiveness for problems in the V1 product does have some limits.

A while ago, someone had this great set of pictures titled, "Let's go to the Record Store." "Let's go to the Video Store." "Let's go to the Book Store." It was meant to show how music sales moved online, putting Tower etc. out of business, movies moved online, putting Blockbuster out of business and Amazon plus a financial crisis put Borders out of business and downsized Barnes & Noble extensively. It was brilliant -- I wish I'd saved the link -- because it also illustrated the high water mark of suburban big box media retailing which we may never attain again. You cannot seriously claim that The Long Tail (as a statistical description of the self-similar, powerlaw distribution of sales of any given kind of item) isn't real. If you do, you're innumerate, and that's okay. I don't make fun of the disabled because we all have our issues (I have more than many people do).

When people _do_ claim Hits Haven't Gone Away, well, they're right. Anderson maybe _hoped_ the hits would get smaller, and to the extent that hope seeped into the book (and at points, it really did), well, he's been disappointed, at least in terms of profit-making enterprises. But it's not clear that that was ever Anderson's focus. He's one of the lead sheep, unlike me, so he's really in it for the Cool, and the Discovery and cultural and reputational stuff like that. I'm a second rank sheep; I want to know how it's going to pay.

But those sheep back there, baaing along, claiming that the world is the same chopped off, hits only, no niche world of the 20th century?

Bwah ha ha ha ha.


I was wasting time on lifehacker the other day, when I ran across a recommendation to bring your EZPass transponder with you on vacation and use it in a rental car as it will save you money. I went, hmmm. R. always says you can't do this, and yet lifehacker says that Consumer Reports is giving this advice. NJ and Mass EZ Pass sites both agree that you can do this, and seem to think you should add the rental plate to your account BEFORE driving in it and remove it 2 days _after_ you return the vehicle.

Let the nerd odyssey begin.

Question 1: SunPass isn't compatible with EZPass yet, is it? No, it is not. Is it about to become compatible with EZPass? If you are a total sap and believe in deadlines imposed by Congress in MAP-21 it is. I Am Not a Sap.


Question 2: Can you get a SunPass transponder from out of state and, if you do, what kind of costs are you incurring (one time fee or annual fee) and how does that compare to what the rental car company offers? Turns out if you rent a car for at least three days in Florida, like, ever, it'll probably cover the transponder cost. Or really, really close. 4, for sure. At this point, it's a matter of whether you will remember to bring the transponder with you or not.

Question 3: Why AREN'T SunPass and EZPass compatible? See: Transcore lawsuit, patent exhaustion, etc.

Question 4: So why exactly is everyone out west other than California using 6C? And does it have a chance at becoming a national solution? Because it's way cheaper and at least as good (worse is better!!!), it's open source, and it isn't the installed base underlying either Sunpass or EZPass. Which in the bizarro world of trying to figure out what National Standard to adopt is actually a _plus_.

Question 5: What went wrong with Title 21? Dude, they wrote the law wrong. Surprise.

And now that I've gotten the worst of that out of my system, here's a brief summary of toll road growth.

We used to have lots of privately built and maintained roads, because they were ludicrously expensive to build and use. Then we had privately built and maintained rail roads, which were cheaper, but still quite expensive. Then we grew our federal government and Wilson sort of nationalized the operation of the rail roads For the Duration because the rail roads and nascent trucking industry were not able to move men and materiel fast enough. Some of the industry participants (at very high levels) were kind of hoping Wilson would carry on with that, but he unwound it. The trucking industry continued to develop, and we had a series of presidents build a bunch of highways (yeah, I know you think the freeway system was our first national highway system, but you're wrong and I don't even want to get into it), culminating in the Ike era interstate system which we now take for granted and allow to fall apart because raising the gas tax is Impossible.

Why is raising the gas tax impossible? Unclear. But more importantly, we'd have to raise it _a lot_ and keep raising it (in real terms -- just indexing it to inflation is not going to get it done, altho obvs that would help) if we wanted to rebuild the kind of highway fund that the gas tax used to effortlessly build. And _that_ is because fuel efficiency is really improving, and we have a lot of people running Volts and Priuses and other stuff, and it is only going to get more so, which means all of those vehicles will cause wear and tear on the highways without contributing to their upkeep through the proxy of buying gallons of the black stuff.

We _could_ fund the highways (already are, actually) through the general fund, at the national, state and local levels. But if we would like to maintain the idea that highways are somewhat/mostly funded through a use tax or a use tax proxy, then tolls are a fairly obvious solution. And tolls are attractive now because we can collect them without making cars stop (or even slow down much) and we can make them variable to Encourage Desired Traffic Patterns. AKA, yes, you can drive through Seattle on I-5, but we will be asking you to pony up to do so, type of thing (not that that has happened, but I think it is _a_ foreseeable future, if not _the_ foreseeable future). Toll collection can also be used to fund transit (and we had already established the precedent of using gas tax money to fund transit and bikeways and sidewalks and so forth, so it's not even a huge change), thus reducing the need for additional lane-miles and maybe even keeping the wear and tear down.

Tolls have been in the news. Christie, in NJ, got into a lot of trouble over what appears to have been politically motivated punitive bridge access lane closures. The details were uncovered in the course of an investigation into an unpopular toll hike (honestly, has there ever been a _popular_ increase in tolls?). In the PacNW, there is a proposal to toll the I-90 bridge to complete the access lanes to 520, to balance traffic, and probably ultimately to provide a funding stream for general road maintenance and development. It looks likely to happen, with the traditional sop to Rich People on Mercer Island Who Complain Loudly (already WA DOT has said they won't toll island-to-one-side, only if you go all the way across in one day, in response to big turnout at a Mercer Island public hearing; I vividly remember how Mercer Island wound up with that gold plated lid on I-90, so this is part of how the game is played out there). Meanwhile, here in Land of EZPass, there is an effort to modify (and reintroduce) tolling along the Mass Pike and possibly later other roads. Whenever I hear the amounts charged (or see them on my statement) I try, again, to understand how anyone who might ever contemplate taking surface streets to avoid these tolls can possibly afford to live in Massachusetts. I always fail. I kept trying to understand why New Hampshire wouldn't adopt EZPass, too, and after I left, they did.

As I read the work product of IBTTA, I started out absolutely incensed at them and their attitude towards Congress and MAP-21 and blah blah bleeping blah. They seemed like the all time worst trade association ever, and I can really get going on trade associations behaving badly. But when I started to read about the Transcore lawsuit, and realized that while IBTTA's approach to coming up with a national standard for toll readers and transponders SEEMED utterly craven and embarrassingly ridiculous, it had actually been very effective in terms of pressuring Mark IV/Kapsch into making public/granting a perpetual license (subject to some constraints -- it's basically copy-left altho they call it something else) their protocols. By defining a hypothetical National Standard as being open source, they produced enough incentive to make the biggest player (slightly over half of all tolls collected in the US, by dollar amount) open up.

Which I have to respect.

It's early days, yet, for tolling in the US. By the time I am too old to drive and people young enough to be my grandchildren are old enough to convince the court to make me turn over my car keys (yeah, look, I know that we won't be using keys for starting cars then -- I'm not an idiot), we will have sort of forgotten that there is an element in our vehicle that does calculations and charges us money when we drive around. It will be that automatic and that normal, and we will view scofflaws approximately the way we currently look upon shoplifters and tax evaders: an indication of bad character, and when caught, you must pay up, but we probably won't throw you in jail unless you are a little too clever and systematic about it, or running a ring of people doing it for profit.

At least, that's _a_ future, if not _the_ future. And to be clear, I am not at all predicting private highways. Not even remotely. The opposite of true.

ETA: Anyone care to take a guess at when map programs (in cars, on phones, etc., such as google maps, tom tom, etc.) will show you total toll cost so you can compare two routes that include tolls on the basis of cost as well as time/distance? Currently, they usually offer options like, "avoid toll roads"; I'm specifically looking for something along the lines of fast food menus which show calorie counts along with the dollar cost.

ETAYA: If you don't like the idea of funding highway maintenance through gas tax increases, tolls, or the general fund, I'd be interested to hear of alternative funding ideas. NOT funding it would be akin to the French deciding that defending the Franc towards the end of the 19th century was a lot more important than defending France. Not a Good Idea.