by Ian S. Hayes

The Internet and ever-increasing bandwith have spawned a new method of delivering software functionality -- the application service provider (ASP) model. An ASP creatively combines aspects of software packages, outsourcing and Internet delivery, allowing customers to receive all of the benefits of using a software application without the associated burdens of supporting and operating it. Instead of purchasing software licenses and operating software on their own equipment, companies can now rent software applications directly from an ASP. For a fee, the ASP provides all of the services necessary to support and operate the software including network access, computer resources, maintenance, installation help and customer support.

First emerging in 1997, the ASP market is still in its infancy although analysts predict it will become a billion dollar industry in the next few years. Early pioneers like Usinternetworking and Breakaway Solutions may have led the way, but large players like Microsoft and Oracle have joined the fray in the hope of bringing ASP offerings into the mainstream. While vendors have been enamored with the ASP model from the outset, customers have taken longer to warm to the concept. Initial market resistance appears to be dying down, however, and there are indications that the ASP model is gaining wider acceptance in the buyer community. In a recent Cutter Consortium survey of 87 companies, 14% of the respondents reported that they were currently using some form of ASP services.

Companies have yet to flock to ASP services. Why? Many companies are accustomed to operating their own software, and the ASP model is just too foreign to contemplate. IT organizations are understandably hostile to ASP solutions out of job security fears. Company executives are leery of shifting control of critical software to outsiders, and are concerned about the security and privacy of sensitive, confidential data. Business users worry about application performance, responsiveness and the ability of the ASP to deliver what was promised.

Overcoming these concerns will be pivotal to the success, acceptance and eventual widespread usage of the ASP model. ASPs must address these worries by spelling out the services that they can and will provide, by offering guarantees or promises as to their performance, and then solidly delivering on the whole. In short, before ASPs gain firm footing in the buyer community, their service models need to evolve to a more customer-centric approach.

One way for an ASP to improve its customer focus is to offer more comprehensive, higher quality service-level agreements (SLAs). Even though many observers would say that SLAs exist primarily to benefit customers, they also offer many compelling benefits to an ASP. From an ASP's perspective, SLAs

• Quell the concerns of prospective buyers by showing that the ASP has the experience and understanding to perform the job as promised and giving customers a greater level of control over their purchased services,
• Serve as a marketing tool to attract buyers through strong performance and service guarantees and meaningful credit/penalty schemes, and
• Convince customers to stay by ensuring high-quality service

It is true that application functionality and the potential for cost savings may provide a customer with the impetus to buy, but only the quality of the services will get them to stay. A strong SLA ensures that the ASP will stay focused on what customers value most. It will serve to prevent poor service by expressing realistic and obtainable commitments and meaningful remedies if those commitments are not met. It will also provide the ASP with information that it can use to monitor, measure and improve its performance proactively.

Today, almost every ASP claims to use SLAs. In Cutter's recent survey, of the respondents using ASP services, 72% had an SLA in place with their provider while 28% did not. Given that these survey respondents included early purchasers of ASP services, many of whom purchased services before SLAs became more prevalent, this figure likely understates current SLA usage in the market as a whole. In the past six months alone, general awareness of SLAs and service commitments has taken center stage. Analysts and the press are covering the topic with increasing frequency, but vendors have been generally mum about the actual terms of their SLAs. Nevertheless, several organizations such as the ASP Industry Consortium and the ITAA have initiatives underway to create libraries of vendor SLAs and even to draft "standard" SLA terms.

SLAs will undoubtedly become increasingly important to the ASP industry, but will they focus on those items that customers truly value? And will customers have the clout and the expertise to negotiate the terms of an SLA? According to the Cutter survey, 87% of companies with SLAs did negotiate terms with their vendors, but 13% still signed the vendor's proffered, standard SLA. While the 87% figure seems high, the survey did not delve into the extent of the negotiations that took place, or the scope of the changes sought and obtained. Given that most SLAs in use today are very one-sided -- favoring the vendor -- it is likely that most companies did not obtain substantial changes to the terms initially offered by the ASP.

From the SLAs that we have examined, most of the service commitments currently being made by ASPs are rudimentary at best. Looking at a cross-section of SLAs in use today, the overall impression is that they are very weak from a customer point of view. Observations include:

• Few metrics are used. The ones selected tend to focus on performance issues such as network availability and hardware capacity utilization -- things that are easily measured and collected using performance monitoring tools. In the Cutter survey, of those companies with SLAs, 87% reported that their SLAs covered infrastructure performance issues such as availability and uptime. In contrast, the survey revealed that only 62% of those companies had commitments relating to the frequency of software releases and/or assurances to correct software defects within a specified period of time.
• Most performance commitments are vague or ambiguous. Many limit the ASP's risk by narrowly defining commitments to things under the ASP's "direct control," or by averaging outages over a period of time to stay within guaranteed levels.
• Remedies are weak. Penalties are generally confined to pro-rata credits for lost service, and many ASPs put the onus on the customer to discover and report problems within a short window of time to qualify for a credit. The ASP generally has no burden to notify customers of outages, and supply credits, if the customer does not encounter the outage itself.

Part of the reason that these SLAs are weak can be traced to the business roots of the ASPs themselves. Many ASPs have evolved from infrastructure providers and software companies, neither of whom is accustomed to offering, meeting and monitoring the types of service commitments that their customers may actually need. Infrastructure providers typically have a depth of technical expertise, and it shows in the SLA terms that they offer. They are quite adept at collecting and interpreting the more pure, easily measured technical statistics such as network uptime and availability. However, they often lack a broader understanding of the application functionality being offered, the customer business processes tied to that functionality, and the type and quality of services needed to support the applications themselves. In contrast, software companies do have a more intimate understanding of the software applications that they offer and their customers' related business practices. However, software companies rarely, if ever, provide any guarantees about the functionality or quality of their software (just take a look at a typical software license agreement) and often lack the customer service skills needed to support their applications in a real-time environment.

Existing SLAs may be weak, and ASP vendors may have to overcome their roots to offer more robust SLAs, but there is a trend underway to rectify the situation. As more ASPs realize that stronger SLAs are the key to greater market share, and as buyer sophistication increases, SLAs will undoubtedly evolve. What types of SLA trends will emerge?

• Increased use of "quality" metrics in an attempt to capture more than just performance issues. These may include the quality of service and support offered by the ASP, the quality of application functionality, responsiveness to defect correction and access to new functionality.
• Reliance on more business-oriented SLAs rather than technical metrics. For example, rather than guaranteeing blanket uptime, the ASP may promise to achieve 100% uptime during the customer's critical business hours. Or the customer may offer the ASP financial incentives (such as a percentage of online revenues) for superior processing of transactions.
• Meaningful remedy provisions, with penalties and rewards that can actually motivate an ASP. Some ASPs are already strengthening their remedy provisions and using them as a marketing tool. Navisite, a company that provides infrastructure services to ASPs and web hosting firms, recently announced that it would offer SLAs that proactively compensate customers for downtime beyond guaranteed levels. Navisite customers would not be obligated to report outages to receive a credit, and credits could range up to 100% of monthly fees, depending on the situation.
• One-stop-shopping customer support. Under current ASP business models, no ASP owns all components of its solution but instead relies on a series of partnerships to assemble a complete offering. For example, no ASP possesses its own web hosting infrastructure, but uses the backbone of companies such as Exodus, Interliant or Navisite. In addition, many ASPs do not even own the software that they host, but license it from a third party. While these configurations may make business sense to an ASP, they are largely irrelevant to a purchaser when it comes to service. Whether or not the ASP owns all aspects of its solution, the customer expects that the ASP will be responsible for the operation and quality of rented services from start to finish. When a customer experiences a problem, it wants to go to the ASP for resolution regardless of the "partner" that is ultimately at fault. With one-stop-shopping, the ASP agrees to provide support for all aspects of the service being provided, and may even pass along the guarantees made by its partners to its customers.

The ASP model and the ASP market may still be relatively immature, but one thing is clear -- the ultimate viability of the ASP market will be determined by the quality of the services provided. To identify and offer the level of quality needed by potential customers, ASPs must adopt a more customer-centric attitude. Until they do, the ASP market will be stuck in second gear.