Outsourcing /

What went wrong with ASPs?

Application service providers were supposed to be the next big thing, but the market got off to a rocky start and many vendors went bust. Now a new generation of ASPs, joined by traditional software vendors like Microsoft and Oracle, is promising to deliver on the 'software as a service' concept.

By Jennifer Mears
Network World, 10/29/01

From the start, the premise was simple: to deliver hosted applications over the network to customers. Application service providers jumped into the market in droves to cash in on what was expected to be a gold mine. During the ASP frenzy of 1999, Dataquest predicted ASP revenue would top $22 billion by 2003. But hitting it big as an ASP wasn't so easy.

"Everybody expected this to be instantly adopted, that customers would just stop dead in their tracks and say, "Wow! This is what I've been missing all my life!' - like a revolutionary kind of thing," says Laurie McCabe, vice president and service director at Summit Strategies. "But very few things are revolutionary. Most things are evolutionary. And what we're seeing is the evolution of the industry."

Evolution means change, and in the ASP market, change means consolidation. Dozens of ASPs are struggling or going out of business. Those failures - including the Intel and SAP venture Pandesic, and more recently, ASP pioneers Breakaway Solutions and FutureLink - are making an already skittish customer base even more wary.

Gone but not forgotten
Loads of ASP information and research

However, at the same time, hundreds of companies are happily using hosted applications. And although the ASP market hasn't come close to reaching the lofty initial projections, it's still strong. According to Gartner, ASPs pulled in $2.1 billion in 2000, and worldwide ASP revenue should hit $7.4 billion by 2003.

The ASP industry is experiencing growing pains common to any emerging technology as IT executives slowly warm to the idea of changing the way they do business. The difference is that instead of providing a tangible product, ASPs are delivering a service that is somewhat difficult to define. Add to that the inherent complexities of remotely delivering applications designed for a client-server environment, and the troubles quickly mount. Nevertheless, the notion of software as a service is here to stay. The question is which vendors will figure out the winning way to deliver that service.

Growing pains

ASPs emerged during the dot-com boom when capital was easy to get and business plans consisted of little more than a "build it and they will come" philosophy. But demand didn't take off quickly enough, and many ASPs were stung by their lack of focus.

"Most ASPs tried to be everything to everybody," says Lew Hollerbach, managing director of service providers at Aberdeen Group. "You cannot go into business these days and have the sort of "throw anything against the wall and see if it sticks' approach. It just doesn't work. You have to have focus."

For example, Pandesic was highly touted because of its founders Intel and SAP, but when Pandesic couldn't make the transition from delivering e-commerce applications to dot-coms to servicing larger companies, the venture folded.

Such failures have scared many firms away from considering ASPs. But IT executives who have successfully used ASPs advise their peers against ruling them out.

For example, National Semiconductor calls its ASP rollout a definite success, says Ulrich Seif, CIO for the Santa Clara manufacturer. Last year, National Semiconductor decided to shift its IT operations and thought ASPs could save money by offering scalability and reducing staffing needs.

"We wanted to explore if the ASP model was viable. To really test the model, we had to find out if we could put a key application like our finances on an ASP," Seif says.That worked OK, and Qwest Cyber.Solutions (QCS) now hosts National's 10,000-seat SAP human resources package and a SAP financial package that's accessed by about 500 users.

Another ASP convert recommends doing your due diligence to pick an ASP likely to stay afloat.

"Go into it with your eyes open. Evaluate the worst-case scenario. Come up with contingency plans and always monitor your applications," says Rodric O'Connor, vice president of technology at Putnam Lovell Securities in San Francisco.

O'Connor carefully checked the financials of the ASPs he considered, ensured that critical data he needed was synchronized daily into an internal database in case the vendor went out of business, set up redundant network connections between his four offices to offset any infrastructure problems and used VPNs and Secure Sockets Layer-encrypted Internet connections to access his applications. He uses AppShop for Oracle Financials, Salesforce.com for customer relationship management, BlueMatrix for research management and Jamcracker for human resources software from Employease, access control from iPass, and online collaboration tools from WebEx.

"The core financial arguments [for using ASPs] outweigh the fear," O'Connor says.

Technical difficulties

You can't blame potential ASP users for being cautious. In addition to some shaky business models, ASPs were facing difficult technical hurdles. "There were last-mile concerns, concerns about security over the Internet, transaction volumes over the Internet, things like middleware to write and maintain interfaces," says Rob Unger, CEO of Agilera. "It's not as easy as people might have thought it was going to be. But that's actually where the value to clients comes in, because in effect they'd have to manage all of that internally if they were going to do it themselves."

So while early ASPs targeted dot-coms and small businesses that required little integration and customization, and accessed their applications largely via the Internet, maturing vendors struggled with how to provide the same service to more demanding large corporations.

Corporations were worried about security, connectivity and integration, observers say. One thing that troubled them was delivery over the public network, says Mark Hoffman, application systems architect at Tufts Health Plan and a member of the ASP Industry Consortium's End User Council.

"There's the question of latency involved with the Internet cloud, because it's really kind of a black hole when it comes down to it. Even with a VPN, you can have a disruption of service; your service provider's router could fail," Hoffman says.

Even today, most large corporations connect to their ASPs, at least for data transfer, via private lines. For example, National Semiconductor is located close enough to a QCS data center that it accesses its SAP application via a T-3, "which is a 45-megabit connection that puts them right on the LAN," says Klaus Preussner, director of infrastructure for the chipmaker.

The proliferation of less-expensive bandwidth helped launch the ASP market because vendors could offer the pipes needed to deliver bandwidth-heavy enterprise applications, such as PeopleSoft and SAP, with high levels of performance. The proliferation of metropolitan-area Ethernet will only increase the quality of application delivery.

But bandwidth at a lower cost does nothing to help ASPs deal with the complexity of delivering applications designed for a client-server environment, another technical obstacle ASPs encountered.

"They were forcing a square peg in a round hole," says Summit Strategies' McCabe. "If you have applications that were designed to be put on a server by a company, used by one company, used with a Windows interface, let's say, the things you have to do to make that work as a Web-delivered service to many customers is pretty mind-boggling."

What's more, big companies didn't want the "vanilla" applications ASPs were set to offer, preferring customized software that could be integrated with their internal systems. While ASPs offer some degree of customization - QCS, for example, has prepackaged, industry-specific software suites - few are willing to spend the time and money it takes to offer high levels of customization. One exception is Corio, which now hosts Hitachi America's highly customized SAP R/3 software.

Another approach to customization is to choose an ASP that focuses on a vertical market. Larry Hancock, CEO of Altius, a health maintenance organization in Utah, was faced with upgrading his online claims and benefits software from Quality Care Solutions. He also had to meet stringent security guidelines set by the Health Insurance Portability and Accountability Act - all with an IT staff of 12.

By turning to healthcare, industry-focused ASP Trizetto, Hancock says he can continue to run applications customized to his needs and provide high-quality service to his 115,000 members, without worrying about internal IT.

"Somebody like us, we just can't keep up on it," he says. "When you have somebody like Trizetto that has the resources and is big enough to keep up on it, then there is a lot more value there than just some somebody hosting the application for you."

Delivering the data

Assembling the infrastructure necessary to deliver hosted applications isn't inexpensive. Most ASPs have found it difficult to establish the customer base and recurring revenue they need to offset debt.

Market research firm IDC released a return-on-investment study earlier this month showing that ASP usage provides a more than 400% return on investment over five years. Analyst Amy Mizoras says, "One of the things we're investigating is are the customers getting a [return on investment] at the expense of the ASP's own [return on investment]?"

Take USinternetworking, for example. One of the first ASPs when it launched in 1998, USi decided it needed to build its own data centers to offer the complete service it figured larger companies would demand. USi invested millions of dollars in its buildout, but customers were slow in coming.

USi instituted layoffs and decided not to build out any more data centers. Earlier this month, Bain Capital promised the ASP a $100 million equity investment that should carry USi to profitability. Still, USi executives say their decision to build their own data centers was the right one.

"Some people have suggested that you can provision an NT or Windows 2000 box, or a Unix box, with an Oracle database or a SQL database. Stick a router in front of it with a firewall behind it, build some bricks and some concrete and stick a cage in there, and you have all you really need to be an ASP in terms of the application stack," says Matt Howard, USi vice president of business development. "That's not true at all."

What's needed is tight integration between the application and all layers of the infrastructure - from servers, to network operating systems, to databases, to networks, he says. "It all starts with service. And to be in the service business you need to understand the total stack."

Infrastructure is something network executives need to take into account when considering ASPs. If your internal network isn't suitable for handling ASP-delivered applications, you may face big headaches such as systems integration and the need to install client software on each desktop.

"You need to make [the ASP] a strategic initiative," says National Semiconductor's Seif. "You've also got to have the right infrastructure behind it."

National uses a portal to deliver all of the applications accessed by its employees, customers and partners, eliminating the need to install clients on individual computers. That made moving to QCS easy. "When we moved to the ASP we only had to change one IP address," Seif says.

Software as a service

Vendors such as Microsoft, Sun and IBM are developing Web services that will let applications communicate with each other using XML-based languages. And PeopleSoft 8 and Siebel 7, for example, are designed to be delivered online.

What does this mean for pureplay ASPs? They'll have to find the niche where they can not only deliver applications but also add value, observers say. Those that don't are ripe for acquisition by telecom companies or large consultancies with the deep pockets to fund the top heavy business model.

"When we look, say, five years from now, the early ASPs will probably look as different from the software as service providers as Cro-Magnon did from humans," says Summit's McCabe. "It's got to change and adapt before it reaches its ideal state."

One thing analysts expect are more so-called Internet business service providers, such as Salesforce.com, Upshot, Portera and Blue Matrix, which write their own Web native applications to be delivered via the Internet on a monthly subscription basis.

There is no question the ASP market will thrive, observers say. But you probably won't know the players as ASPs. Software as a service, it seems, is what we'll call the next big thing.

"The word ASP will have to be retired. It's too damaged a name. To relaunch the word ASP, would be tantamount to relaunching ValuJet," says Tom DePasquale, CEO of Outtask, a firm that hosts business applications such as human resource management, CRM and sales force automation. "Software as a service is the entire model of the Microsoft .Net effort. It's very much the focus of Oracle. It's the focus of the leading software companies of the world. It's easy to argue the days of shipping CDs are over."