« January 2005 | Main | May 2005 »

March 09, 2005

The long tail of software. Millions of Markets of Dozens.

Hello blogging, my old friend…

You know you’ve been constipated in getting a post out the door when you start it with three caveats:

  1. This post is WAY long. For that I’m sorry. Some things just take a while to say.
  2. This post is not so much about entrepreneurship as it is about where I see opportunities for entrepreneurs to create new businesses and describing the general direction of my new company, JotSpot.
  3. This post is based on a presentation I’ve been giving for the last five months. If you want a copy of it, you can download it here. Download jotspot_long_tail_sw.ppt  

Excite and the Long Tail.

In Excite’s heyday, we were handling millions of searches a day. If you graphed the frequency of those searches (the Y axis being the number of times a query was asked per day and the X axis being the actual query listed in order of frequency) you got a power law curve that looked something like this (excuse my lame-ass powerpoint graphing...).


The most popular searches (things like Sex, MP3, and the bare midriff female singer du jour) were vastly more popular than the 1000th most popular search. For example, “sex” was on the order of 100,000 times more popular than the 1000th most popular search (whatever that was). Said another way, there were a handful of extraordinarily common queries and millions of far less popular queries.

In fact, the frequency of the average query was 1.2. That means if you wrote each of the millions of queries on a slip of paper, put them all in a fish bowl and grabbed one at random, there was a high likelihood that this query was asked only once during the day. Of ten-plus million queries a day, the average search was nearly unique.

The most interesting statistic however, was that while the top 10 searches were thousands of times more popular than the average search, these top-10 searches represented only 3% of our total volume. 97% of our traffic came from the “long tail” – queries asked a little over once a day.


Down the tubes

You know the real reason Excite went out of business? We couldn’t figure out how to make money from 97% of our traffic. We couldn’t figure out how to make money from the long tail – from those queries asked only once a day.

Overture figured it out, Google perfected it and we all know what happened from there. Those guys figured out something revolutionary -- the long tail of search was a advertising marketplace. But it wasn’t a traditional advertising marketplace like television, where a handful of large advertisers reached out to a handful of very large markets. It was a special kind of marketplace where small advertisers could reach small markets efficiently. It was and is a revolution to the traditional economics of advertising (where the cost of producing and distributing advertising requires large markets to justify the expenditure).

Search is a long tail business and that is the source of its power and profit.

Other Long Tail Businesses

Can I get a shout-out for Chris Anderson? For those of you who don’t know him, he’s the editor-in-chief of Wired Magazine and he wrote a great article called the Long Tail about six months ago.

Chris pointed to three other long-tail businesses (Rhapsody, Netflix and Amazon) as examples of the power of the tail.

Here was the first set of charts Chris showed.


Let’s look at the Amazon example. This graph shows that Amazon sells roughly 2.3M books and that the average Barnes and Noble retail store stocks 139,000 books. So, Amazon stocks roughly 2.2M more books that Barnes and Noble.

No surprise here. That’s the benefit of an online storefront. Massive inventories housed in ultra-low-rent areas that are fronted electronically.

The astonishing figure is the percent of sales that comes from the “long tail” of books (books that Amazon carries but that Barnes and Noble doesn’t).


57% of Amazon’s sales come from books you can’t even buy at a Barnes and Noble (to be fair, there is some skepticism around this number voiced here). This runs totally counter to the traditional 80/20 rule in retailing – that 80% of your sales come from 20% of your inventory. In Amazon’s case, 57% of their book revenue comes from 0% of Barnes and Nobles inventory.


To quote Gary Coleman, I can hear retailers across the globe saying,  “What you talking about Willis!?”

ITunes and the Long Tail

iTunes has over one million songs in it’s catalog. You know how many have been bought at least once?

Every one.

Completely counter to the traditional 80/20 rule, every iTunes song has been purchased at least once.

What this says to me is that the tradition 80/20 “rule” is more a function of consumers having their expectations set by the limits of physical inventory in retail stores than it is about real human nature. Amazon and iTunes are great examples.

So what?

What all these data points mean to me (and to most folks who are interested in long-tail stuff) is that the most interesting, transformative businesses that have been built over the last decade and that will be built over the next one are going to operate in and make money from the long tail.

Google, eBay, Amazon, Rhapsody, Netflix, iTunes. What do they all have in common? They all work the long tail and they’re all radically changing the dynamics of their more traditional businesses.

The Long Tail of Software

The long tail doesn’t just apply to music and movies. There’s a long tail for software as well. Here’s why.

The purpose of software in business is to support the way a business does business – from the way a business runs it’s hiring and firing to the way it orders materials to the way it tracks sales. In the market-speak that surrounds the technology business, the purpose of software in business is to support these “business processes”.

Let’s do some simple math. First, every business has multiple processes. Things like hiring, firing, selling, ordering, etc. Second, while some of these are pretty common in name from business to business (recruiting, for example), in practice, they are usually highly customized. Finally, there are simply a large number of processes that are either unique or that are common to millions of very small markets and therefore not traditionally worth the effort to buy software for (for example, the process by which an architecture firm communicates between it’s clients and the city planning office).

These three facts

  • every business has multiple processes
  • processes that are similar in name between businesses are actually often highly customized
  • there exist a large number of processes unique to millions of small clusters of industries.

means that there is a combinatorial explosion of process problems to solve and, it turns out, little software to actually support them.

Said another way, there is a long tail of very custom process problems that software is supposed to help businesses solve.

Inaccessible Tail

In the past, software’s long tail has been generally inaccessible because software has been

  • Too difficult to write
  • Too expensive to write and distribute
  • Too brittle or expensive to customize once deployed.

It just hasn’t been economical for someone to create a custom software company to help architecture firms.

That’s why, in the software business, the traditional focus has been on dozens of markets of millions instead of millions of markets of dozens. The traditional software model is to make software have enough features and address enough of a homogeneous market that you can sell millions of copies of the same software. In the past, that’s been the only way to make money.

How the software tail gets address today

The market doesn’t like a vacuum and people do solve their software needs in the long tail. They do it using two basic tools: Microsoft Excel and email. I’ve seen so many business that run on Excel+email. People build structured lists in Excel and then send them out over email for comments and updates – a list of people to hire, a list of deals they want to do with action items included, a list of features for the next product. Something like this


While normal users don’t think of it this way, what they’re really building is an long-tail application – a custom application, built by the end user and networked over email.

Doing better (warning, personal plug for JotSpot coming…)

Excel and email are the wrong tools for software in the tail and we all know it. It’s really easy to start with, which is fantastic, but it suffers from.

  1. Versionitis. We all know what happens with spreadsheets like these. You create it, you mail it to 10 people. One of them changes and mails it back out. Rinse, lather, repeat until everyone’s inbox is full of this thing and no one knows who has the latest version.
  2. Updates. You only know the sheet has changed is when someone emails it to you.
  3. No integration. What about the stuff that doesn’t fit in the grid? – the email and documents that go along with these spreadsheets?

That’s where JotSpot comes in. JotSpot is a company that is building a platform to make it easy and affordable to build long-tail software applications. To take those Excel spreadsheets and turn them into real web-based applications where you don’t have versionitis, where updates find you instead of you looking for them and where you can integrate data in your hard drive with data from the web, email and other applications.

Please, really Joe, wrap it up…

So, my tip for entrepreneurs? It’s all about the long tail. Whatever business your starting, think about how to serve millions of markets of dozens instead of dozens of markets of millions. Serving the head isn’t a bad strategy. You can build a great business. But, figure out how to serve the tail of your market efficiently and you’ve got a blockbuster.

March 9, 2005 | Permalink | Comments (50) | TrackBack