Not, of course, that Salesforce.com defines itself as a CRM provider. According to Gloster, Salesforce.com is actually an "online platform company" - CRM simply being one application that is delivered by the platform they provide. This statement takes us to the heart of what seems special about Salesforce.com. Potentially, it has to be said, very special.

Whilst the traditional CRM market is dominated by providers whose sophisticated offerings come attached with a whole raft of costly infrastructure requirements, Salesforce.com appear to be successfully pioneering a delivery model which - much as it may owe to the ASP model - has the potential to transform the entire software industry.

"For us," claims Gloster, "it's all about sharing the risk with the customer. We don't get paid unless we deliver." Gloster is comfortable describing the company he works for as a utility company. Just as a client of an electricity provider flicks a switch and pays to access the electricity he or she requires, so a salesforce.com user can pay to use only the services or parts of an application that they need.

The Salesforce.com platform is built to allow users to access applications through a gateway, rather than requiring them to both license and install software. Any PC or handheld device with an internet connection can make use of the platform and users can also work in offline mode, enabling on-the-road salespeople to synchronise their data when convenient.

Scalability is a key benefit of the model. Salesforce.com can cater for teams as small as 5 and scale up to an unlimited number of corporate users. "It's a simple process to grow your CRM system along with your business," claims Gloster. "You only need to pay for the extra headroom that your salesforce requires if they actually need it."

This flexibility is key. Previously, Mike Dodd, VP at Giga Information Group has shrewdly identified a key flaw in the on demand offering of some vendors - pointing out that users also need to be able to reduce their demand levels, as well as raise them. With Salesforce.com, the utility computing ideal moves closer. Gloster points out that a key sales-point for them is that companies only need to give 30 days notice to end the relationship. He actually claims much of their new business has been won by companies turning to Salesforce.com as a temporary measure, whilst they await the completion of a project to install a CRM solution such as Siebel. "Once they've tried us out, they tend to stay." Gloster declares.

"It's an old adage, but you don't buy a car without test driving it first. You certainly don't buy one after watching a man take you through its technical prowess on a laptop. So why do companies buy CRM on the same principle?"

Initially, this accommodating model opens the door to the small and medium businesses that cannot afford the large associated costs of a traditional high-end CRM solution. However, as Salesforce.com has entered the fray for large corporation business, the market has continued to react favourably. Commentators have bracketed Salesforce.com with internet behemoths eBay and Google, whose IPO would arguably kick-start the bleeding tech sector into life. Equally, the growing revenue curve (it doubled in 2002, to reach an estimated $60m) appears to leave Salesforce.com with no current need for a cash injection (playing Larry Ellison at his own game and Marc Benioff taking himself back to Oracle would be somewhat out of reach for now).

It is clear that the utility computing model, when applied to software, has potentially revolutionary effects, exposing incumbent market leaders to a difficult decision: if they will not ignore the on demand trend (and it is increasingly obvious they cannot afford to), how will they to migrate to a fundamentally different revenue model?

Computer Associates recently faced a similar dilemma and made the difficult decision to alter their payment structure. In doing so, they suffered a great deal of short-term pain, as their share price fell significantly. However, their recent recovery suggests the move paid off. In the opinion of this columnist, it is a question of when, not if, the wider software industry begins to pay significant attention to these trends.