I have just finished reading Chris Anderson’s (Editor of Wired) excellent book - “The Long Tail” and it really opened my mind to the profound change in economics that is taking place online and how this will affect main stream business. The Long Tail premise is that the 80/20 rule no longer holds because some businesses may sell more individual or “one off” products in TOTAL than the usual top sellers -ie the top 20% of items sold (the hits)account for less than all the one off items. In this case a very large number of people buy a small amount of unique products. Anderson’s premise is that when there is unlimited choice this is the typical demand profile. His point also is that this is the way consumers ultimately like to buy. In other words hits still matter - but not so much anymore. The details are well explained in Wikipedia. Examples of The Long Tail are the usual suspects - Amazon, Google, ITunes, increasingly cable TV and so on. There are many of them in many different categories.

Here are some reasons why this is a very important concept for mainstream IT:
1 IT is the real enabler of The Long Tail - in the majority of cases its digital content or web sites that are globally available to large numbers of people
2 Business models invariably follow the rule that best sellers should account for the bulk of inventory at any one time - now, that may not be true. Shelf space is no longer the limiting factor - inventory is virtualized.
3 Any traditional business is facing a consumer population becoming increasingly used to unlimited choice - this could be a cumulative effect -that is the norm - what can they do about it?
4 If IT is the enabler in the first place - how will corporations use IT to address the problem? Isn’t that the logical way to deal with it?

One question may be:-”why worry? These sites have been around for a while. Well one reason is that there are more of these sites now - they live in almost every category. Anderson’s book analyses well the economics and the incredible profitability of “infinite variety selling” by using someone else to provide it - and even more amazingly peer production such as open source or the wikinomics argument [getting someone else to buld it for you].

IT management essentially builds, operates and sells technology services to the Business. Their business clients when they read books like The Long Tail will be starting to ask questions like
” Where is my Long Tail?”
To IT :-”How will you help me build one?”

Anderson provides some good suggestions on the business side of this equation - its the downstream implications I am interested in. For IT the impact may be the following -

- Service/Application Focus. If in the book category alone there is infinite variety, its difficult to offer such variety across many market segments. The same goes for delivery - sharing code may save costs but is difficult across lines of business all determined to offer a lot in a short time frame. As with The Long Tail itself - one size does not fit all - IT will probably need more than ever to focus on supporting specific lines of business with highly customized needs. Standardization will not be easy.

- In the services business most options/extras are just process changes - application code needs to be built to be modified quickly - so The Long Tail needs to be planned for. Anything that is difficult to change probably needs to be changed. Flexibility should be the new yardstick. A well thought through SOA strategy is an absolute “must”.

- Federation technology - this allows a web site/or application to allow users to securely log in in one place but gain access to others. This is an excellent ennabling short cut that can lead to ultimate variety on the cheap!

- Speed. Flexible code is one thing - but speed to deployment is another. Infrastructure teams used to infrequent changes need to consider much faster and frequent releases. Change becomes the norm.

- Teamwork - all of the above means a lot of stress for IT teams in general, creating greater demands for both the business, IT development and infrastructure teams to seamlessly drive change into variable offerings much faster than they have before.

The last point about teamwork is probably the most important because without it good design will not be executed effectively - having a common computing model that does not rely on any one “tribes’ methodology” is a very important mechanism to support teamwork.

As a very interesting final note The Long Tail as a distribution curve also applies to the use of words in a particular language. But just because we all tend to use a smaller number of words more frequently does not mean that we understand them by the same meanings. A “tree” can mean many different things especially if you are an Active Directory specialist - frames of reference and different mental models play a significant role in how people use words - common visual models are an excellent bridging mechanism