Experience suggests that pinning down the economics of the Internet is no easy task. Even as momentum builds around the so-called ‘Web 2.0', the next generation of dot coms still has a worrying propensity to think technology first and business model second – if at all.

So it is heartening that Chris Anderson, editor-in-chief of technology glossy Wired magazine, has defined the way the Internet affects markets with a forthcoming book, The Long Tail. While most high-street music shops only have room to stock the biggest-selling artists, online stores like iTunes can serve those customers outside the top of the demand curve – potentially a third of the total market. The Internet's lower distribution and inventory costs mean a long tail of niche tastes can be exploited.

Search engines are a classic example. While the most popular search terms are keyed in many thousands of times more than the least popular, they make up only a small proportion of the total number of searches, says Joe Kraus, a founder of Excite.com, a portal that fell victim to the dot com bust in 2001. Now CEO of wiki vendor JotSpot, he says Excite "didn't figure out how to make a business out of 97% of our traffic volume". Likewise, he says, "in the software business, the traditional focus has been on dozens of markets of millions instead of millions of markets of dozens."

Anderson illustrated his theory with hosted CRM provider Salesforce.com's new AppExchange, a marketplace to which users and software vendors post applications they have created and picking up others to plug into their own Salesforce platform. It brings together developers and consumers of niche products, smoothing development, deployment and distribution.

Anderson argues that traditional "monolithic" software application vendors cannot hope to appeal to every last employee, but try, resulting in ‘feature bloat': "The lowest common denominator is not always the best product." This analysis might suggest that the giant of enterprise applications, SAP, sits at the opposite end of the scale. But the company is trying to address the long tail with its NetWeaver platform, to allow users to compose applications using web services, claiming it allows better data integration than hosted applications.

SAP, however, is not wedded to the web delivery model. Instead, it is betting that enterprise IT's future lies with service-oriented architectures (SOAs). Bill Wohl, SAP's VP of products and solutions, says NetWeaver creates a "common language of business" for companies to express their process requirements and convert them into technology.

NetWeaver tools enable business users, without much training or any programming experience, to compose appli- cations using diagrams in a graphical user interface. Wohl glosses over the complexity of migrating to the underlying SOA, but he argues this satisfies integration and customisation demands that Salesforce's model cannot. "Customers want total control over their data."

In spite of their differences in execution, both companies' visions amount to an apparent validation of Anderson's long tail theory: easily deployed applications, shaped directly by users to their individual needs. What nobody dares mention – in relation to SAP, Salesforce or even iTunes – is the contradiction with the initial promise of the Internet to open up customer choice. All three business models are founded on locking in customers to the vendor's platform – even if that is sweetened by a surrounding ‘vibrant' community.