The ‘Long Tail’ of Expert Systems

Seemingly everyone is talking about ”The Long Tail”, a term that

describes a booming new economic model. The Long Tail, a phrase coined by

Chris Anderson of Wired magazine, is originally a statistical expression,

and here describes a population of low-selling products that, when

combined, can equal or even vastly outnumber the bestsellers.

Business models like Amazon.com or Netflix demonstrate the potency of the

Long Tail. As Anderson reports, blockbuster titles remain popular

mainstays for both companies, but low-cost distribution via the Internet

has enabled consumers to exercise their highly individualistic tastes.

And these specialized titles — quirky though they might be — outsell

the blockbusters when added together.

For example, the average Barnes & Noble store stocks 130,000 titles.

”Yet more than half of Amazon’s book sales come from outside its top

130,000 titles,” Anderson writes. The cost of producing, distributing

and delivering these countless products has plummeted, making The Long

Tail a pillar of the new economy.

Now everyone — mainstream pundits, bloggers, boardroom chiefs and water

cooler quarterbacks — is dissecting the Long Tail phenomenon. In his

Wired article, Anderson focused on the media and entertainment

industries. But the Long Tail effect also is evident in other fields,

like interactive software.

Recently, I had an ah-ha moment at the Caltech/MIT Enterprise Forum,

during a talk by serial entrepreneur Eric Pulier.

With software in mind, it struck me that the decreased cost of

manufacturing and delivery allows for the existence of niche products

that formerly didn’t have the economics to justify not only the marketing

and distribution costs, but the design and production costs, as well.

And this isn’t new. We saw this 25 years ago.

Before VisiCalc (the first spreadsheet available for PCs), what was the

cost to develop and distribute a custom accounting application? VisiCalc

created a Long Tail for an entire class of applications that wouldn’t

normally have enough market demand to justify development, marketing and

distribution. Developers could create small spreadsheet applications and

distribute them. That trend continues to this day — to such a degree

that people wouldn’t imagine paying top dollar for something spreadsheets

can do. However, even with all the bells and whistles of today’s Excel

and Open Office Calc, they really don’t handle complexity that well.

Fast-forward from the ’80s to the ’90s, when everybody and his/her

brother was learning HTML and earning money creating static Websites.

In the commercial world, only people willing to spend a healthy sum of

money could afford a professional-looking Website with the latest design

and layout techniques — certainly not a mom-and-pop shop. Now, a high

quality Website is in reach for everyone. Numerous WYSIWG HTML editors

and full-blown online content management systems, like CrownPeak’s CMS,

mean you don’t need to know HTML, CSS or how to set up a server.

A business can be as far out on the tail as it likes.

So what about dynamic, interactive software applications?

My old friend Dr. Rich Volpe, who manages the Mars Rover Software effort,

lamented recently that the ”working unit of software construction is

still the single line code.” Due to the unique constraints of space

flight software, that may be true for some time for folks like Volpe. But

in many corners of the software world, that is changing. Costs are lower

because of common building blocks like Oracle, PostgreSQL, WebLogic,

Jboss, Hibernate, Spring, and Apache.

But people are still writing a lot of code.

Software Rock Stars

So who are the Dave Matthews Band and Britney Spears of the software

world?

I submit that it’s Turbo Tax Online. You might call it a work process

tool or you might call it an expert system, but it really is a category

of its own. It’s the Tivo of tax — a truly exceptional product that has

single-handedly decreased American anxiety and Prozac use while

increasing quality family and golf time.

But what about the niche products — the individual obscure and specialty

titles so precious to so few, yet purchased by so many. Where is the Long

Tail for interactive software?

Few pieces of software need sufficient market to justify the tens of

millions that Quicken spent developing Turbo Tax. There now are

companies developing products so configurable that there is little or

zero custom code being written for each application, and these

sophisticated interactive Web applications can embed amazingly complex

logic, calculations, and content. You might call them expert systems for

the rest of us — or Turbo Tax for everything else, if you like.

Since these are tool-based platforms like VisiCalc, Excel, or a content

management system, interactive applications can be built without writing

a single line of code and offered over the Web for a fraction of the cost

of a custom application. That means all of those applications out on the

tail — the ones that didn’t have the size to justify the cost of custom

applications — now have a market.

Let’s try a case study to illustrate.

Say you are a doctor with a growing practice who wants to determine the

feasibility and costs of opening your own outpatient surgery center.

Certainly it’s a complex decision, but the market probably isn’t large

enough for a software company to invest millions to develop an

application just for you. But working with the client’s experts, that

very application was built by one non-tech person in 12-person weeks and

has since been used by more than 100 physicians groups.

When was the last time you designed, built, tested, and deployed a

complex software application in 12 person-weeks?

And there’s even a bonus.

The dirty little secret of software is not the purchase price, but what

it costs to modify when your business changes next week. There is old

adage about boats: ”A boat is a hole in the ocean into which you pour

money.” The same is true of custom software applications. Modifying

these applications is convenient and affordable because they do not

require changing code.

Ready for one more case study? This one might take a bit more

imagination.

Imagine you are a rural sheriff whose state just passed a law that

requires you to process permit applications. Odds are against you

digesting a hundred pages of legalese to comply with the new system. It’s

also unlikely someone can build you an affordable application to

implement the new law within, say, 12 weeks, before the citizens are at

your door.

Yet that exact application was built and deployed in two person-weeks at

a price even state government could love. And since it is ”software as a

service”, it was live to 88 county sheriffs the same day it was

finished.

Yes, life is sweet on the Long Tail of interactive software.

Mark Long is senior vice president of technology for PortBlue Corp., which provides a system for building Web-based expert applications. (www.portblue.com)