Software as a Service

How far can the pay-as-you-go approach take you? Proponents say the road less traveled is about to become jammed.
Norm AlsterJuly 19, 2005

In what has become a time-honored tradition, IT upstarts delight in branding their more-established competitors as dinosaurs. If you’re a new hardware company, your machines are faster. Networking? Your gear pushes through vastly more data. Software? You don’t even sell software, you sell “solutions,” usually branded with some combination of three letters that lays claim to a new technological frontier, one that your company will define and dominate. Get on board, Mr. Customer, or be left in the dust, amid all those dinosaur bones.

This overzealous marketing hasn’t escaped the notice of IT pundits, and terms such as “vaporware,” “hype cycle,” and “trough of disillusionment” are now part of the lexicon. But that hasn’t chastened IT companies, most of which are as exuberant as ever. In the case of those at the forefront of a concept known as “software as a service” (SaaS), in fact, the claims that a new era is dawning surpass pretty much any previous hype cycle you can think of.

“Current market leaders in enterprise software will fade into oblivion,” says Greg Gianforte, president, chair, and CEO of RightNow Technologies, which develops and hosts customer-service-management software. “I firmly believe that we’re in the transition era from client/server computing to on-demand computing,” predicts Phil Robinson, senior vice president for global marketing at “In three to five years, all applications will be delivered this way.”

Drive Business Strategy and Growth

Drive Business Strategy and Growth

Learn how NetSuite Financial Management allows you to quickly and easily model what-if scenarios and generate reports.

By “this way” Robinson means that customers will no longer buy software to run on their own computers — an approach that Zach Nelson, CEO of NetSuite, which delivers Web-based enterprise software to small and midsize businesses, dubs “do-it-yourself Stone Age software.” Instead, customers will essentially rent software and access it via the Internet, without having to worry about how to manage it. Most people are already familiar with this concept: your home Web access probably entails a monthly fee to AOL, your phone or cable company, or some other type of Internet service provider (ISP). And while you may have had to install software on your PC to use the service, the vast majority of the computing horsepower needed to make things happen resides at the ISP’s offices, with the cost of keeping the service up and running spread across all subscribers.

Could all this bravado be warranted? Whether you call it software as a service, on-demand computing, or something else (it’s a close cousin of utility computing—see “The Meter System,” Summer 2004), the idea of renting versus buying is not new. It harkens back decades to the time-sharing services on mainframes, and more recently to the application service provider (ASP) market that enjoyed a boom-bust cycle in lockstep with the dot-com era.

At the very least, predictions of the imminent demise of the business software giants seem premature. Hosted software still represents only around 5 percent of total software revenue, and most (though not all) of its success thus far has hinged on niche applications and smaller customers that simply can’t afford to buy enterprise software outright.

But it’s worth noting that most successful technology attacks begin at the low end, and there is ample evidence that SaaS, though not yet the all-purpose solution its proponents claim it to be, is broadening its appeal and shaking off the problems that plagued its first iterations.

Some of the success of the SaaS model has to be attributed to, which has managed to not only take a healthy bite out of the market share of traditional CRM vendor Siebel Systems, but has also proved to be a relentless proselytizer for the SaaS approach. Indeed, Siebel was forced to respond with its own hosted software offerings, and in late 2003 spent roughly $70 million to acquire hosted CRM vendor UpShot. Today, even Siebel sings in the SaaS choir. “All enterprise-application companies are going to have to modify their delivery strategy to offer a hosted, as well as an on-premise, solution,” says Bruce Cleveland, senior vice president and general manager at Siebel’s CRM OnDemand unit.

Investors, which have been largely skeptical of technology for the past five years, have also endorsed the concept of hosted software: both and RightNow Technologies executed successful IPOs last year. And two bellwethers of the ASP business, USinternetworking and Corio, have finally managed to either eke into the black (USi) or be acquired by a software giant (Corio was acquired by IBM in March, in a move seen as key in allowing IBM to quickly offer small and midsize businesses a hosted option).

In total, worldwide spending on SaaS exceeded $4 billion last year and is expected to grow to $10.7 billion by 2009, according to IDC. In an IT environment where many vendors are learning to live with single-digit growth, that’s an eye-popping CAGR of 21 percent, a figure that leaps to 40 percent when looking at major services companies such as IBM. (Meanwhile, independent ASPs will see revenue grow about 16 percent, which may explain why many believe the IBM-Corio deal will usher in a spate of acquisitions — unlike actual dinosaurs, the metaphorical kind have proved time and again that deep pockets provide a substantial Darwinian advantage.)

SaaS comes in two basic flavors. “Net-native” or “Web-native” companies develop their own applications and design them to be run via the Web for a multitude of clients. Other companies don’t develop their own programs, but simply take over the management of applications that customers have already licensed from traditional vendors. The latter approach is how most ASPs operate today; customers still buy software outright, but arrange to have it run from remote data centers, saving on infrastructure, manpower, and other support costs. Currently, the market is divided about equally between vendors that have developed their software expressly for the Net and those that simply host standard programs in their own data centers, according to IDC analyst Amy Konary. But she believes the Web-native approach will eventually win out.

Whether developed specifically for remote hosting or simply hosted to eliminate the cost of acquiring and managing infrastructure, the appeal of on-demand software lies in its speedy deployment, low up-front costs, and overall promise of savings. When a customer licenses software for on-premises use, it pays an up-front fee, along with ongoing yearly maintenance costs that typically run to 15 percent or more of the initial licensing fee. Additionally, the customer must acquire the hardware and supporting systems (such as a database) to run the application, and will most likely need to employ pricey consultants to get things up and running. That makes for hefty overhead, but companies have usually been willing to pay it because they feel it buys them maximum control.

Estimates of savings from SaaS vary widely.’s Robinson claims a hosted solution can be 80 or even 90 percent cheaper than a traditionally licensed software package. But analysts warn that initial cost advantages can diminish over time as monthly subscription fees pile up. That’s one reason Rebecca Wettemann, vice president of research at Nucleus Research, emphasizes that subscribers must continue to monitor hosted applications to ensure that they are optimized for changing corporate needs. “Don’t just assume that it’s better because it’s cheaper up front, and that it doesn’t have to be evaluated and managed on an ongoing basis.”

But even with those caveats it’s hard to deny that up-front savings can be substantial. The cost of implementing conventional enterprise software is typically four to five times the cost of the original license, says analyst Phil Wainewright of Summit Strategies. So even when the cost of a license and a two-year on-demand contract are equal, the SaaS approach does appear to be the large bargain that Robinson and others claim it to be. Little wonder then that SaaS vendors are extremely aggressive in touting the superior ROI of their services. Indeed, Wettemann reported that a Nucleus Research study found that just 40 to 50 percent of customers with licensed CRM applications achieved positive ROI. By contrast, 82 percent of companies using hosted CRM saw positive returns.

But the SaaS model does face some major hurdles — technology, for one. As Andy Stern, CEO of USi, explains, software written expressly for the Web has thus far appealed chiefly to small and midsize businesses. “Today, most of the Net-native products aren’t fully capable enough for a billion-dollar company,” he says, adding that they tend to lack the rich feature sets and customization capabilities that large enterprise customers often require. They can also be difficult or impossible to integrate with other applications and to modify as internal business processes change. And the programs aren’t dubbed “Net-native” for nothing: while the Internet is a very low-cost way to provide access and essentially share one program among many customers, it also creates a host of security concerns. “This is an issue for us,” says Kim Perdikou, CIO at Juniper Networks, which does subscribe to some SaaS services. “I’m concerned about the security of customer and company data, that it not be copied or interfered with or compromised in any way” (see “A Prescription for Subscription,” at the end of this story).

Right for Everyone?

David Brooks, director of CRM at Magma Design, an electronics design automation firm and customer, says that “there was some concern initially that security would be an issue.” But confidence in its own IT staff combined with the belief that has plenty of incentive to address security issues (“If they had a breach, they’d probably go out of business,” Brooks says) assuaged the company. As for the ability to customize the program, Brooks adds, “ gives us up to 95 percent of what we want to do, and I’ll trade off the last 5 percent of customization capability to be able to sit back and relax on weekends when they do the upgrades.”

Fabrice Cancre, CEO of U.S. operations at R/D Tech, a producer of test equipment used by many industries for defect analysis, says that one way around the integration problem is to subscribe to SaaS software that is already integrated. His company is a client of NetSuite, whose product suite combines finance, CRM, inventory management, and other components. Cancre noted that he has 125 users on the software system but “zero IT staff supporting it. This gives a small but fast-growing company like ours a chance to use software that we otherwise couldn’t afford. Even a second-tier ERP package would cost half a million dollars just to get started.”

One key question is whether large customers will ultimately side with Cancre’s view and decide that a monthly subscription model makes more sense. Siebel’s Cleveland says he’s fielding more and more requests for OnDemand from larger clients. “Customers are telling us that they’re intrigued,” he says. While Siebel CRM OnDemand has reported strong revenue growth in recent quarters, on-demand users at large companies represent only about one half of 1 percent of the company’s total user base. Some analysts suggest that large software companies face a major hurdle in moving to an on-demand model: how will Wall Street react when massive one-time licensing fees suddenly give way to monthly subscriptions? That would have a profound impact on financial statements, at least in the short term, and for public companies, the short term is the only reality that counts.

One major advantage to the SaaS model, according to IDC’s Konary, is that it aligns the interest of vendor and customer. With traditional licenses, she says, “there’s no strong vested interest from the vendor perspective to make sure the customer is successful.” Once a package is sold, she says, “the vendor is not coming back until it has something else to sell.” SaaS vendors, by contrast, must continue to justify their ongoing monthly fees. “They aren’t just throwing software over the fence,” she says. “They have to make sure the customer is happy with it month after month.” In the long run, that alignment of interests could prove to be hosted software’s strongest advantage.

A Prescription for Subscription

The remote hosting of software applications offers users many benefits. These include lower up-front costs, speedy deployment, free maintenance and upgrades, and perhaps best of all, no need for pricey know-it-all consultants. But there are risks in embracing software as a service (SaaS). Here are three key considerations that can’t be overlooked:

Service-level agreements. What level of uptime will a remote host guarantee, and how quickly will it respond to a problem? Many customers insist on negotiating guaranteed uptime. If a provider, for example, falls short of 99.9 percent (or some other agreed-on level of) uptime, there should be some negotiated penalty to compensate the customer.

Data security. Although SaaS vendors invariably emphasize the resources they devote to security, many customers remain uncomfortable with their employee and customer data flying over the Internet, not to mention potentially residing on the same data-center server as their rivals’. “Look at security. Do the due diligence. Make sure the vendor has the right premises and that protecting your data is its top concern,” counsels David Brooks, director of CRM at Magma Design. Juniper Networks CIO Kim Perdikou insists on modifying SaaS contracts so that she has the right to do periodic security audits.

Customization capabilities. Does your changing business process require more tweaking of the software than a remote host can — or is willing to — provide? Or do you need (or plan) to integrate the capabilities of the software with in-house applications or other software that you may subscribe to? Either one of those can be a deal-breaker. —N.A.