Don't get bitten by the ASP

Application service providers (ASP) were billed as a panacea for all computing ills when they arrived on the scene a couple of years ago.

But with the hype starting to fade, there is scant evidence to suggest that such a sales model is going to revolutionise the IT world in the foreseeable future.

According to its proponents, the ASP model would provide the answer to the industry's skills shortage by supplying software on a rental basis. This would also enable customers to respond more quickly to fluctuations in the e-economy.

Large corporations could roll out applications quickly and deliver products to market before their rivals. Small fry startups, on the other hand, would gain access to heavy-duty IT applications that they would not otherwise be able to afford.

The model presumed that the internet, and other carrier networks, would provide access that was secure enough and cheap enough to make a compelling case for companies wanting to farm out the hosting and management of their software to a third party.

And by operating on a one-to-many basis, ASPs would be able to benefit from economies of scale and pass these cost savings on to customers. At the very least, users would be afforded predictable budgeting.

But the ASP model also promised that customers would be able to scale their applications up and down according to their needs, and pay for them on a metered basis like a true utility.

The utopia of companies being able to pay for their packages on a usage basis has not materialised anywhere in the industry, however. While the proposition of handing over the hosting and management of core applications to others has found favour in the US, pricing models are frequently based on traditional licence models.

According to analyst company IDC, US companies accounted for 90 per cent of the $150m spent on hosted and managed enterprise applications last year.

But Europe has been a lot slower to embrace the model, and after a number of false starts, suppliers in the region are now revising their claims about how and where ASPs can add the most value to businesses.

Users are starting to overcome their caution in entrusting core applications to third parties, however, and are now talking the language of service level agreements with their suppliers.

IDC sees both of these factors as a key influence, and predicts that the ASP market in Europe will grow from $15m in 1999 to $850m by 2004.

Putting the 'S' in ASP
Gary Pugh, data server marketing manager at Oracle, believes that the 'S' in ASP has been largely to blame for slow take-up so far. "Although the letter stands for service, most suppliers are delivering software," he said.

This may be because the ASP model has its roots in the early co-location services business, where suppliers managed basic hardware resources and network connections remotely. It was simply a logical next step to add software to that configuration.

But as Pugh pointed out, users need to bear in mind that the model is not about outsourcing their software hosting or even adopting a more relevant pricing model. The real issue is delivering benefits back to the business.

"The vast majority of [an] IT budget - up to 80 per cent - is consumed by infrastructure and doesn't deliver any benefit back to the business," he said. "A lot of time is spent maintaining code or adding patches to back-office code. Things like payroll and financial applications are not the crown jewels of a company's business."

By using ASPs for these services, companies can focus on the software and services that make up the remaining 20 per cent of their budget, and which provide a crucial competitive advantage. Pugh said that Oracle had done precisely this when it re-engineered itself to move to an ebusiness model.

The software house previously used 97 servers and 120 databases worldwide to host its email system, but now claims to have saved millions of dollars by rationalising them into one server running one database to host the function globally.

Tim Pickard, vice president of the ASP Industry Association, agrees that the ASP proposition should involve much more than simply providing applications on tap. "Classical infrastructures involve far more than legacy systems - it's about process and chemistry too," he said.

But he believes that into the long term companies will still want to manage their own infrastructure in-house and will use ASPs to add new functionality to their business.

Pickard envisages the scenario of a high street bank hosting financial packages for its small business customers so it can differentiate its brand. In this case, a sensible strategy once it had achieved its market objectives would be for the bank to bring its software and processes back in-house.

A rash of recent launches supports the theory that the ASP model works well with financial applications. Suppliers which have launched such a service include QSP and Anite Group, which are both marketing their wares to dotcoms. QSP identified a serious lack of financial management expertise within this group, evidenced clearly by the demise of Boo.com and the findings of its own research.

One of the attractions is that dotcoms can increase the amount of services and software they rent as the business grows. Some 86 per cent of the sample surveyed by QSP saw this as a priority across all business functions. Few dotcoms believed they would be able to support and maintain financial software on a global 24 x 7 basis, for example.

Modifying the model
But QSP's research also exposed one of the major flaws of the ASP model, and one for which no supplier has yet found an adequate answer. A 57 per cent majority of dotcoms interviewed said they would need to customise packages to accurately map them to their business processes. As a general rule, the higher the company's turnover, the more modification was deemed necessary.

QSP claims that many of its core applications are provided in template form, which enables individual user companies to undertake a degree of modification. The vendor also adds value by providing services around the core templates, which includes some integration work and training.

But no industry benchmark has yet been developed to quantify the amount of integration and customisation that would be necessary to optimise a given application over a three-year life cycle, for example.

Such a benchmark would potentially be as valuable to users as the 'PC total cost of ownership model' that research firm Gartner came up with in the 1990s. It found that support and enhancement costs way exceeded initial capital outlay.

Malachy Smith, QSP's group chief executive, conceded that the future direction of the ASP model will be dictated by market expectations. "The model is there to drive down costs and create a lower point of entry for adopters," he said.

But he knows that customers will not be prepared to pay high customisation costs and will probably want the traditional benchmark to apply, i.e. 20 per cent investment to provide 80 per cent functionality.

So the biggest benefit of the ASP model may yet prove to be enabling companies to outsource complete functions such as IT or finance systems, as long as they do not require bespoke applications.

One success story here is the Centre for High Performance Development, a company with 70 employees that specialises in training senior management. All of its business packages are now hosted by Esoft Global, and the company, having made the radical decision to do away with its IT department, has freed up its capital for core investments.

As Pickard said: "Budgeting, and making the best use of capital, are the key advantages of the rental model."