Invent a new acronym, and the IT vendor community will flock to it like moths to a naked flame. The application service provider (ASP) moniker is no exception, and while no single vendor has yet been burned, it is a difficult proposition for potential customers to understand. So, despite the pre-existence of the software rental market and well-worn arguments for outsourcing non-business critical applications to a third party, actual user references remain few and far between.
The concept of ASP is defined in a recent in-depth report by Lehman Brothers as "delivering computing resources via a one to many model; owning the resources being delivered; taking responsibility for the delivery; and acting as a single point of contact". The idea is that users rent software applications through online links with a service provider, avoiding the upfront costs of purchasing applications and the headache of implementing them. Certainly vendors seem to believe that the concept is proven--and most major companies have announced an ASP strategy. But vendors have so far failed to provide potential buyers with compelling evidence of how ASPs can address their business needs and convince them that security and other issues have been adequately addressed.
None of this prevents the analysts churning out estimates that the market is about to explode. Recently some of these predictions have become industry jokes. For instance, Deloitte Consulting reckons the ASP market in the US will be worth $48.5 billion by 2003--but of that figure, only $27 billion represents applications. IDC believes that the worldwide market will grow to $7.8 billion by 2004, lumping collaborative applications, groupware and office applications into one pot. Ovum claims that 66 percent of British and German businesses plan to use an ASP in the next 18 months. By 2004, it says the European ASP market will be worth $8.7billion and the worldwide market $44 billion. By 2006, it puts the worldwide market at $132 billion, with the US market worth $83 billion.
Regardless of whose figures you read, they all have great expectations of the ASP market. All analysts agree that the ASP model best suits small and medium-sized businesses, with a turnover up to $250 million and fewer than 1,000 employees. But, to date, there is little evidence that the vendors one might expect to have the most to gain--ERP vendors--are profiting from this.
Phil Wainewright, independent ASP analyst, blames the vendors for not making the position clear: "Vendors have still not understood that the operative words in the acronym are 'application' and 'service'," he says. In Wainewright's view, most of the action involves little more than relocating hardware, using the same pricing model as in a straight licensing sale, adding charges for service provision and then calling it an ASP offering.
This is the model used by Oracle with its Business OnLine (BOL) offering. SAP's model is similar. Both insist they are committed to ASP, with BOL vice-president of marketing Terry Rollo maintaining the Oracle commitment to supply 50 percent of applications as an ASP service within a couple of years. Quite how it plans to reach that point is a mystery and, so far, the subject has been a huge turn-off in Europe. Oracle UK cannot come up with a single reference site, or even any indication of whether sales have been made. SAP is in a similar position. At last month's user conference, it was unable to field one customer worldwide to appear under the ASP banner.
This need not mean that as a model, ASP is stillborn. It comes down to figuring out what the market looks like and what can be expected. PeopleSoft vice-president of e-business Deepak Gupta is charged with turning its ASP strategy into revenue dollars. He says PeopleSoft, which itself announced its first UK partner last month, will still sell a license, but the difference comes in the implementation and service. "PeopleSoft has a tradition of customer service. We're taking that and building on it to provide a superior offering." Andy Simms, PeopleSoft's European ASP director, believes the key driver comes from IT departments under pressure to manage complexity. "There are numerous demands on IT at the moment, and an ASP style of offering provides an escape route," he says.
So far, PeopleSoft has landed five customers in the US from its direct operation, and claims in excess of 100 worldwide through partnering arrangements--including Corio. Gupta believes the ASP model could be delivering up to 50 percent of the company's revenue in three years' time, but in Europe, Simms says the market is more complex. "You have to recognise that each country will have different adoption rates."
Nor has the best method of delivering the ASP offering been decided on. At first, software vendors like Oracle thought they could go it alone, but later back-pedaled when they realised the way forward was through partnerships. Internet service providers and telecoms companies like BT would like to see the ASP model take off because it generates network traffic. These organisations are keen to partner, but they are not presenting a clear picture to potential customers for each category of software that would fall into the ASP definition.
BT, for instance, will claim it provides everything an enterprise might need, including Office software and Lotus Notes groupware, and is spending millions on marketing its offerings in the area. BT product strategy and alliances manager Angus Fox says: "We've been trialling a number of initiatives over the first quarter." But it is difficult to see a compelling case for outsourcing Office products unless the enterprise has a substantial investment in leased line connections, because these applications require intense application access, an untenable position if using a time-cost-related connection like ISDN or modem dial-up.
In the short term, it will be the vendors' pricing model that will attract customers. Even here the potential for confusion is high. Customers centering on the licensing model may not see any advantage to buying an in-house application license. But, as Gartner has been arguing for years, total cost of ownership is only marginally impacted by software license costs. The really large costs are in lifetime support and maintenance. However, many enterprises still experience great difficulty in making sense of this cost element, partly because the desktop PC is a notoriously difficult asset to cost and manage. Few enterprises look at the real cost of managing applications that require not only support from the software vendor, but also internal systems and database support.
Gartner itself has said the ASP model allows enterprises to make between 20 and 40 percent savings on application costs, based on a realistic estimate of internal systems support charges. This has yet to be proven in the real world, but it is not unreasonable to predict some savings. One way it can be achieved is through near to "vanilla" implementation. Although it goes against the business process implementation model of the ERP software vendors and restricts them to providing templates developed on the back of industry solutions, this allows the software vendor to quote a reasonably low licensing and implementation budget. Corio says this runs counter to its strategy of providing tailored solutions made up of best-in-class components, but even it does not want to customise too much. "We must give customers what they need, but in the context of an offering we can replicate," says Corio founder and chief strategist Jonathan Lee.
A near-standard product that can be shared may keep costs down, but the idea is proving difficult to realise. From a software licensing perspective, ASPs are finding application vendors reluctant to support such an arrangement since they make their money from licensing per customer, per seat. In the model an ASP would like, seats would be added over time, reducing the initial revenue stream for the application vendor.
Customers also want to know they can go and see "their" hardware at a secure data centre and are reluctant to share databases. They perceive that shared services means a loss of security, which is not strictly true provided the controls are adequate. But, in the ASP game, perception is everything.
The ASP model has much to commend it, provided that management keeps an open mind as to what can be delivered. The fact that vendors like Oracle and PeopleSoft want to see significant revenue streams from this source is encouraging. But there is still a pressing need to convince the market hat the ASP model is worth taking to heart. Pricing and deal models have yet to be refined, and fears about security have not been removed. Perhaps when that happens, customers will be more easily persuaded.