Boy, Was That Fast!

Sardined into their seats and stacked three and four deep around the walls of a stuffy conference room in Boston, the sweating faithful gathered to hear Christopher McCleary, CEO of application service provider (ASP) startup USinternetworking Inc. (USi), tell them how ASPs were going to change everything.

It was 1999, and ASPs were the new new thing. Everyone hung on CEO McCleary's words. Until....

Maybe it was the lack of air in the room, but when McCleary announced that the Annapolis, Md.-based ASP wasn't going to charge a dime to implement its menu of complex rented software applications, the audience started to laugh. And it wasn't a scattered twitter here or there; it was a deep, rolling, communal belly laugh. Caught like Johnny Carson in the toils of a bad monologue, McCleary, who is now USi's chairman, did what Carson used to do; he turned on his material. "Look," he joked, "they tell me to say this stuff."

You couldn't blame McCleary. Last year, ASPs didn't have to admit that the costs for implementation (among other things) were hidden in their monthly software rental fees. They didn't have to prove much of anything, either--that they were making money or even in some cases that they had live customers. Most could float along on the hype generated by the ASP concept itself and its solid underlying logic of low cost and convenience for companies that either couldn't afford or didn't want to manage complicated applications like customer relationship management (CRM) or enterprise resource planning (ERP).

But that was then. This year, no one's floating and no one's laughing. The venture capital faucets are being turned off, and investors are beginning to clamour for a payback from an industry that generated a pitiful US $300 million in revenue last year, according to Framingham, Mass.-based research company International Data Corp. (a sister company to CIO's publisher, CXO Media)--less than a single big ASP client makes in a year. McCleary's company, which is touted by analysts as one of the better run ASPs, is still standing, but last year it lost US $103 million on revenues of US $35 million. (And much of that revenue came from--you guessed it--implementation consulting services.) USi's main competitor, San Carlos, Calif.-based Corio, lost US $45 million last year on revenues of just US $5.8 million.

The losses can be explained in part because ASPs spent a lot of money on infrastructure and marketing to gain their first customers--a building year, as they say in sports. But there's more to the struggles of the leading ASPs than that.

Just a few years after its birth, the ASP industry's original premise--that companies would be willing to rent big, complex software applications right off the shelf--is being criticised by the software analysts who describe the market and the venture capitalists who fund it. More significantly, businesses are resisting the core ASP sales pitch: A generic piece of software can work as well for a company that sells chemicals as it does for one that sells cars or computer chips, with no modifications or extra functionality built in to serve the particular quirks and needs of those industries. And with that resistance, the ASP business model, based on the economies of scale to be reaped by using the same basic software to serve many different customers at once, is showing significant leaks in logic and, most critically, money.

The predicted shakeout of the ASP market--which, according to Stamford, Conn., research company Gartner Group Inc., will be a swift and brutal 60 percent of the current herd of about 500 by the end of next year--has begun. And one of the acknowledged pack leaders, Pandesic, was the first to go.

The Price Was Wrong

Pandesic LLC in Sunnyvale, Calif., was established in 1997 by two of the three biggest, richest technology companies in the world: Germany's ERP software maker SAP AG and Silicon Valley-based chip giant Intel Corp. For the past three years, it has been renting e-commerce management software to online retailers. But on July 28, Pandesic's customers received an abrupt e-mail: "We are winding down our business," it said, "because we don't see a timely road to profitability due to slower than anticipated market acceptance of business-to-consumer electronic commerce solutions."

This was not a case of a couple of guys in a garage running out of borrowed money. This was a strategic decision made by a company whose six-member board was an even mix of SAP and Intel big shots. Pandesic exited a market, business-to-consumer retailing, that, three years ago, seemed ready to burst wide open.

It did. But it gushed red ink.

Although Pandesic blames its failure on the downturn in the Internet business-to-consumer retailing market this year, its demise contains more than a few warning signs for the rest of the ASP market.

Pandesic, which by all accounts had happy customers and working software, built its business model on a foundation of optimism that turned out to be misplaced. First, it pioneered the loss-leader approach to software installation fees--a bread-and-butter revenue source for traditional consulting companies--that its major competitors in the ASP industry also have embraced.

Pandesic chose not to charge its customers for such traditional consulting services as figuring out who would use the applications, deciding which business activities would be affected, installing software upgrades and hooking up to the company's network. Instead, it simply tried to bury those costs in a flat monthly fee and by taking a 2 percent cut of everything that customers sold over a Pandesic-powered website. (Most of its competitors now opt to charge customers a fee per user per month so that as application usage expands, the ASP can make more money.) Pandesic's revenue strategy had two unwholesome effects. First, it attracted the customers that Pandesic needed least: small, unproven dotcom startups with no revenues to speak of and therefore very little to lose. (Sources estimate that about 80 percent of Pandesic's customers were small dotcoms.) Second, among the customers Pandesic did want--big brick-and-mortar retailers that desired a Web presence--the scheme hit a sensitive nerve.

"Profit margins in retailing are just not that high [between 3 percent and 7 percent or less]," says Randy Covill, senior analyst for e-commerce strategies and applications for AMR Research in Boston. Cutting into 2 percent of that margin essentially meant Pandesic would be helping itself as much as 20 percent to 30 percent of profits. "To say you're going to take [that much] away is a hard sell," Covill says. Bigger retailers also resented the prospect of having an outsider's hand in their pockets. "Companies resist giving up a piece of their business in return for a service," adds Covill.

For Pandesic spokeswoman Paula Stout, the company's inability to attract big customers was simply a matter of wrong place, wrong time. With the booming economy, says Stout, most big companies believed they had the resources to build their e-commerce websites themselves. "Our No. 1 competitor was people who were building [the websites] in-house rather than outsourcing them," she says.

Though Pandesic's e-commerce engine was by far the most complete ASP package available on the market, it still could not be all things to all customers. "Let's say customers wanted 1,000 different functions and were willing to pay US $10 million to do it themselves," says Stout. "We'd say, 'We'll give you 900 of those functions and charge you US $1 million a year.'" But the big prospects wanted those last 100 features, says Stout, and, again given the economic good times, they were willing to spend the time and money to build them themselves.

The customers Pandesic did have called for more features in the software. But unlike traditional outsourcing companies that manage individual application portfolios for individual customers--and can charge a pretty penny to each customer for any changes in those portfolios--Pandesic couldn't charge for its time. Instead, its developers were left with the Sisyphean task of winnowing out a list of enhancements that would satisfy as many customers as possible while not bloating the software and bringing the company's hosting infrastructure to its knees. For example, Oshkosh, Wis.-based children's clothes maker OshKosh B'Gosh was one of a number of companies that wanted Pandesic to upgrade its software so that it would be possible to offer Web customers a free shipping option. "We wanted it in two weeks; they said six months," says OshKosh B'Gosh Vice President of MIS and CIO Jon Dell'Antonia. To its credit, Pandesic responded and built the new functionality well within the six month time frame it had promised Dell'Antonia and other customers.

But such responsiveness does not come without a cost. The proof is in the software itself: After just three years, Pandesic was on the fourth version of its software--as many incarnations as SAP has offered in the seven years since it first wrote R/3. Indeed, SAP gave up trying to make its core R/3 program all things to all people years ago and formed specialised task forces to customise the software to the specific needs of the oil, apparel, chemical and education markets, among others.

When Startups Stall

Three years ago, Pandesic could at least count on a steady stream of startup dotcoms that were prepared to forego a few bells and whistles to get up and running quickly. But lately, even that well has dried up. "Just try getting venture capital money for a pure-play Internet retailing dotcom startup today and you won't get 3 feet," declares Christopher Terry, president and CTO of Boca Raton, Fla.-based ASP HostLogic Inc., which serves the business-to-business market.

In a world of hype that heats up and cools down like a cheap toaster oven, the B2B market is now as hot as the B2C market is not. Pandesic never tried to rejigger its e-commerce engine to adapt to the needs of B2B commerce, which generally requires more feature-rich, specialised software highly adapted to the needs of a particular industry. "Our system was built to handle high volumes of low-cost transactions," says Stout. "For companies doing less than 100 transactions a day, it just wouldn't make sense." To adapt the engine for B2B would have required a major rewrite, she adds.

ASPs outside of Pandesic's narrow B2C realm claim immunity to the low-revenue customers and sagging market demand that did in Pandesic. Yet whether they focus on plugging gaps in big companies' application portfolios or on the lower volume, higher cost transactions of B2B e-commerce, these companies face the same problem Pandesic did: How to make their software specific enough to serve the individual needs of customers without allowing the code to grow out of control.

ASP Survival Guide

ASPs hoping to survive leaner times are going to need to specialise. "You won't see good ASPs going after 20 different companies in 20 different industries anymore," says David Boulanger, AMR Research Inc.'s service director of enterprise applications. "The good ones are going to fight to get a reference customer in a particular industry and then try to resell that solution to other companies in that industry. It won't be easy, but by the time you get to the eighth customer in an industry, it gets a little easier." He cites USi's recent specialisation in Internet-based CRM for the telecom industry as an example.

As ASPs move farther from their original mission and begin to create more specialised software and more individualised customer relationships, they become more vulnerable to competition from traditional providers like systems integrators on the one hand and big outsourcers on the other. Indeed, some of the major consulting houses such as New York City-based Ernst & Young LLP and PricewaterhouseCoopers LLP (PWC), and outsourcers like Electronic Data Systems Corp. in Plano, Texas, have already entered the market. Other consulting and outsourcing companies are poised quietly like vultures around the edges of the market, watching and waiting to see which ASP startups will bite the dust.

Defenders of the ASP market say that these outsiders will not be able to simply swoop in and snap up the orphaned customers. ASPs, they say, are developing skills that traditional outsourcers and systems integrators don't have. "The traditional outsourcing mentality is that you take a customer's applications and put them on autopilot for eight years," says Boulanger. "You make as few changes as possible and try to reduce your maintenance costs." Meanwhile, ASPs are driven by their subscription models to redesign and refresh their application technologies frequently with a creativity that most traditional outsourcing companies lack.

ASPs also have developed another advantage over traditional consulting companies--speed. Driven by their pricing schemes to minimise the costs of installing software, ASPs are getting good at being quick, installing complex ERP and CRM applications faster than the Big Five consulting outfits could ever dream of doing. "What ASPs are not doing is having a bunch of consultants sitting in a room for three months looking at their navels and drawing on a white board," says Boulanger.

"ASPs aren't just managing companies' technology for customers," says Traver Gruen-Kennedy, chairman of the ASP Industry Consortium in Wakefield, Mass. "These customers are looking for innovative solutions, and I think the innovation component is something that the traditional companies haven't fully understood yet."

Gruen-Kennedy says that this year marks the moment the ASP market moved from hype to reality. Many ASPs didn't get the funding they needed until late 1999 to build sufficient support and sales infrastructures, but "this year they are starting to make gains on wins and deployment," says Gruen-Kennedy. "In the fourth quarter [of 2000] we're going to see a real end user community emerge and get past all the experimental pilots."

Roadkill Recipes

The rented application market is here to stay. Every day, networks and the Internet become cheaper and more reliable vehicles for delivering access to applications. The staying power of the first generation of startup ASPs, however, is in doubt.

"The Goliaths like IBM (Corp.) and PWC are huge wild cards in this market," says Boulanger. "How will a Corio Inc. hold up against an IBM? Good question. And that's why I think some of the ASPs will become roadkill."

As many ASP pioneers are discovering, there is nothing inherently magical in the ASP concept that will free them from the need to tailor their rented applications to the individual needs of customers--especially large customers that want the ASP applications to hook into legacy databases and other software applications. In big companies, CIO's will expect their ASPs to also be adept systems integrators, which will hit the playing field towards traditional integrators, like the Big Five consulting companies that can adapt themselves to the ASP model.

ASPs that survive will rely not on the rental delivery mechanism itself but on the quality of the applications and the individualised service that customers receive. ASPs that can offer unique applications have a shot, as do those that can deliver standard applications with better speed, more industry specificity and more flexibility than traditional outsourcers or the Big Five consulting companies.

But that's a tall order, and time is already running out on the startups. If the demise of Pandesic is any indication of things to come--and it is--next year will be a bloody one for the ASP industry.

Executive Editor Christopher Koch wants to know whether you like ASPs. Contact him at

Hey, What About Us?!

When an ASP turns out the lights, its customers are left to curse the dark Dying technology companies usually leave plenty behind for scavenging competitors and customers. Many corporate mainframes are still running on 25-year-old software whose progenitors have long since passed from the annals of commerce. But as customers of expiring application service provider (ASP) Pandesic in Sunnyvale, Calif., are discovering, when an ASP goes under, it leaves not even a bleached-out server rack behind.

ASPs handle everything for their customers, from building the applications to maintaining them to serving them up on a network. That reduces costs and hassles for customers, but it also means that should a parting of the ways occur, it will be swift and complete--all the ASP has to do is snip a wire and all traces of customer's application and data simply disappear.

For companies that have made the ASP's software the core of their application infrastructures--and in the case of Pandesic's customers, of their businesses--that's a hole that won't be patched over easily. Nor can ASP customers call in their in-house IT cavalry to save them. Most companies use ASPs precisely because they do not want to pay internal IT staff to learn, maintain and support the applications. And even among those customers who have IS departments willing to help out, few have enough understanding of the ASP software to do so.

Once the wire is cut, accessing the ASP's software is impossible. Ed Vincent, CEO of New York City-based startup online retailer Inc., is kicking himself now--as are most of Pandesic's customers--because he did not put a clause into his contract giving him a copy of Pandesic's software if the ASP should ever shut down. "It simply never occurred to me," he says. "But it will be in the next contract I sign."

Vincent is one of the lucky ones. His staff created its own e-commerce infrastructure before going over to Pandesic. If Pandesic turns the lights out before he can find a new ASP to take its place, he can still rig something up.

Pandesic's larger customers, like children's clothing maker OshKosh B'Gosh in Oshkosh, Wis., aren't so lucky. OshKosh Vice President of MIS and CIO Jon Dell'Antonia went to Pandesic to save time and money, but now he's scrambling. Because OshKosh's business is larger and more complex than most dotcom startups, because it has legacy systems and warehouses to hook up, the job of switching over OshKosh's system to a new ASP will take time. "Six months is the minimum for making that kind of a transition," says Dell'Antonia. Although Pandesic has promised him and other customers that it will continue support through Jan. 31, 2001, Dell'Antonia is concerned that Pandesic will not be able to keep its programmers and support people from fleeing a sinking ship. "Frankly, they're in the middle of Silicon Valley," he says. "How long can they expect to hang on to people?"

Vincent has organised about 40 of Pandesic's customers into what he politely calls a "knowledge community." It teleconferences every Tuesday to exchange Pandesic death-watch news and share leads for replacement vendors. The idea is to present Pandesic with a unified front. So far, he says, there has been no shortage of vendors looking to take on Pandesic's customers.

That's why, like most customers, Vincent is left wondering why the plug was pulled on Pandesic so quickly. Out of all the struggling startups, the ASP with the deepest pockets went first. SAP has little to say. "This was not a decision that we were going to cut them off," says Bill Wohl, SAP's director of public relations. "This was a decision that Pandesic made about its own business."

"We thought that with those names behind it there would always be the appropriate financial backing, which of course proved not to be the case," Vincent says wryly.

Join the newsletter!


Sign up to gain exclusive access to email subscriptions, event invitations, competitions, giveaways, and much more.

Membership is free, and your security and privacy remain protected. View our privacy policy before signing up.

Error: Please check your email address.

More about AMR ResearchApplication TechnologiesASP Industry ConsortiumCorioElectronic Data SystemsErnst & YoungErnst & YoungGartnerGartnerHostLogicIBM AustraliaIntelOshkosh B'GoshPandesicPricewaterhouseCoopersPricewaterhouseCoopersPricewaterhouseCoopersSAP AustraliaUSinternetworkingWakefield

Show Comments