Quelle: CIO, USA
We may never know what killed the dinosaurs. Comets crashing to earth? Ice creeping down from the poles? Smart little mammals running circles around them? But we will know what killed the giant enterprise application suites that ruled the last years of the 20th century: smart, little point applications. They've already begun to eat the lunch of those lumbering enterprise suites, and if the trend continues(and there's no reason it shouldn't now that Web services is here), in a few years CIOs won't be buying any more tightly coupled, multimillion-dollar enterprise systems from just one vendor.
Case in point: When FleetBoston Financial last summer decided it wanted to automate the process of identifying potential customers for new products, the $13.3 billion Boston-based financial services company received two proposals. One came from CRM giant Siebel Systems, from which Fleet had purchased millions of dollars' worth of software licenses in 2000. The other was from MarketSoft, a smaller$100 million vendor whose product focuses solely on the kind of lead management FleetBoston was looking to do. And despite the fact that Fleet had scores of unused Siebel licenses, the bank went with MarketSoft's more targeted product, opting for a best-of-breed lead management solution.
"We evaluated the two systems for functionality, cost and speed of implementation,'' explains Ann Christensen, the former executive vice president of customer and sales management in Fleet Boston's consumer banking group, who made the decision. "We found that Siebel had, as you would expect, a broader set of functionality than we needed and didn't go nearly as deep as we needed."
Many companies are making similar choices by seeking out specific point applications and backing away from enterprise application suites. This is a major shift from two years ago when everyone was investing in ERP and CRM packages - the bigger the better. There are numerous reasons for this sea change. Point applications cost considerably less than enterprise suites and often offer more flexibility and deeper functionality. They are easier to implement and upgrade. And, paradoxically, it is sometimes easier to integrate a point app with an already installed suite than it is to integrate new components from the same vendor's suite. For example, Fleet found that the MarketSoft application, according to Christensen, "integrated with Siebel systems better than Siebel systems integrated with Siebel systems."
Siebel and the other big software vendors dispute that contention. They argue that the long-term cost of ownership will actually be higher if a company goes with point applications because of the cost of integration and managing the multiple vendor relationships.
However, Web services is promising to make the process of connecting best-of-breed applications simpler and less costly than it has been in the past, when complicated interfaces were required to integrate solutions from different vendors. In fact, some say Web services may lead to the Holy Grail, allowing companies to link best-of-breed apps for a seamless and much more flexible enterprise wide system. Even enterprise application suite vendors such as Oracle, SAP and Siebel are making their modules compatible with a Web services infrastructure.
"There's a strong backlash emerging against these large enterprise application suites, particularly in these tough economic times," says John Hagel, a Burlingame, Calif.-based management consultant and author of the recently published book Out of the Box: Strategies for Achieving Profits Today and Growth Tomorrow Through Web Services. "Today, CIOs are looking for low costs, short lead times and near-term business impact, which is the polar opposite of the enterprise application suite proposition."
The Case Against Enterprise Suites
Neil A. Hastie, CIO of TruServ, the Chicago-based parent company of the True Value hardware chain, was never a big believer in enterprise application packages. "I was the anti-one-size-fits-all ERP guy before it was in vogue," boasts Hastie. "I still don't know of any ERP system that's totally integrated. There's no one suite that will solve everything for everyone."
Hastie got religion after spending $10 million on a 1998 Oracle ERP implementation at Fel-Pro, the gasket manufacturing company acquired by Southfield, Mich.-based Federal Mogul. Hastie found that Oracle's order management system couldn't efficiently handle the heavy transaction volume at Fel-Pro. "It wasn't vast or flexible enough," says Hastie. Then he found that Oracle's warehouse management system wasn't specialized enough to handle Fel-Pro's distribution function; it didn't integrate well with Fel-Pro's materials handling equipment or scanners. Hastie used several different distribution packages, one of which he built in-house and another that was a bolt-on to Oracle.
"What Oracle was really good at was the manufacturing shop floor and financials," Hastie says. "They were just OK at order management. So we had to find another application to bolt on." (Oracle says that since that time it has introduced a completely new order fulfillment solution, 11i, that includes configuration, advanced pricing, order management, transportation and warehouse management capabilities.)
When Hastie took over as CIO at TruServ in 1999, he applied the lessons he learned at Fel-Pro. No ERP system exists at the $2.4billion member-owned hardware cooperative. The enterprise is running on a suite of applications. They include The D&B Corp. for accounting; a "simple down-and-dirty" general ledger to record financials; a distribution system from EXE Technologies; a homegrown order management system; and what he calls a "data warehouse on steroids" from Business Objects that offers CRM-like analytics, which he assembled himself. This mix-and-match approach may be a little more work in terms of maintenance and integration (TruServ uses Microsoft BizTalk Server and the Oracle Integration toolset to connect everything), but Hastie prefers it to the alternative. "You either have to write some of your own applications in-house or get a package and modify it," Hastie explains. "And like all modifications of enterprise application software, that prevents you from being able to upgrade easily."
Hastie says that as cash has gotten tight, CIOs are getting back to basics. "If you look at the amount of mega millions that CIOs poured into ERP systems, you have to wonder how much further could we have gotten from the true core applications that a company runs on," Hastie says. "And now that IT budgets are getting crunched, those core applications are becoming the focus again."
"We had gotten into a trend where companies were buying massive amounts of stuff, signing a contract for a five-year project, and buying it all up front," says Jim Shepherd, senior vice president of enterprise applications for Boston-based AMR Research. He notes that some CIOs are still paying upward of 15 percent a year in maintenance on licenses they will never be able to use.
"Today, people are buying in much smaller increments than ever before, and they're not signing a contract for the entire suite of applications or buying up 50,000 seats," Shepherd adds. "Those kinds of projects are a thing of the past."
How to Avoid the Upgrade Headaches
When Rich Clow joined Deutsche Bank in New York City as vice president of eConsulting in June 2000, the financial services company had already decided to standardize on Siebel for CRM; it had invested $10million on several implementations. So in late 2001 when Clow was looking for a tool that could analyze customer profitability, Siebel was first through the door with a new tool it had acquired from another company. But Deutsche Bank had implemented half a dozen different versions of Siebel CRM in various business units, and all of that software would have to be upgraded before Siebel's new profitability analytics tool could be installed.
Instead, Clow and his team decided to invest more than $1 million in Alphablox, a privately held, 100-employee business intelligence application vendor based in Mountain View, Calif. Clow found that Alphablox offered a Web-based analytics system capable of connecting the data housed in the existing Siebel systems with the bank's profit and loss data in a way that would allow employees to analyze each banking customer's profitability. "We decided it was easier to use Alphablox than to integrate those different Siebel systems implementations," Clow says.
Clow believes that big CRM vendors are less than valuable to CIOs outside their core competencies (which in Siebel's case, according to Clow, is sales-force automation), though the pressure is often there from the business side to stick with the same vendor. "From a business perspective, particularly in a large global conglomerate, once people have had success with a certain product, contracts are signed and discounts offered, it often becomes simpler to stick with the same vendor even if it isn't the best solution," Clow explains. "But there's a point of diminishing return as you continue to invest in a single vendor beyond their sweet spot."
Siebel admits that although its Inquire module could have provided visibility into profitability, it could not have delivered the kind of profitability engine Deutsche Bank required. "It wasn't a core capability or function of that solution," says David Carter, general manager of Siebel Finance in San Mateo, Calif. "You can always write additional code to make it work, but it wasn't an exact fit."
Last summer, Clow joined New York City-based Citi Cards as vice president of technology architecture and strategy. The credit card company, Clow explains, has no CRM suite but is about to close on a large deal with a small customer data model company to handle customer management. "We needed to add tools around contact management, but we decided to look at small, on-the-cusp types of companies, like small e-mail companies or business performance software vendors," says Clow. Citi Cards will be investing tens of millions of dollars on the customer management project on the assumption that it will pay for itself in less than a year - something few big CRM suites can claim. Smaller vendors are now providing best-of-breed functionality and open architecture that can easily integrate with other solutions- including CRM suites - using standards like J2EE or simple object access protocol, says Clow.
Christensen was facing a similar situation at Fleet when she discovered that the bank would have had to upgrade the entire organization to Siebel 7.0 in order to use the vendor's lead management tool. And Fleet had greatly customized the earlier release, making the upgrade option an expensive proposition. In contrast, an application processing interface was all that was required to connect the MarketSoft point application to the existing Siebel systems package.
So despite the fact that Fleet, the seventh largest financial holding company in the United States, had purchased a lot of extra licenses from Siebel, Christensen made the decision to go with a small, targeted application. Fleet invested just over a million dollars on the MarketSoft lead management product. The navigational tool operated on rules that each of Fleet's business unit managers could customize and change over time. In contrast, the Siebel lead management module was based on rules hard-coded into the system, as they are in most enterprise suites. If Christensen had needed changes or customization, she would have had to call in a programmer.
Furthermore, the MarketSoft product required less business process change than Siebel's and had the potential to pay for itself in a year, at a time when Fleet was hurting financially (revenue was down 27 percent and profits were down 76 percent in 2001), says Christensen, now a freelance CRM consultant in Andover, Mass. "I was able to target the investment to match the revenue and enable a best-of-breed solution versus going with a CRM suite that had anymore business processes that I'd have to implement or take out, and I didn't want to do that," she explains. Using MarketSoft "was an example of a laser-focused implementation to meet the economic need."
Siebel executives say Fleet would not have gone in that direction if there had been a match between the company's enterprise objectives and business decisions further down the line. "Fleet had made a huge commitment to Siebel at an executive level, but that didn't translate down," says Siebel's Carter. "At the business level, they were focused on a very narrow piece of functionality - they were looking specifically for lead management and not a broader enterprise solution."
Most CIOs, however, are searching for just these kinds of lower risk, quicker return investments. "[CIOs are] all looking for limited risk investments with payback in a year's time," Christensen says. "Today, CIOs are looking at ways to invest only in what they need to deliver fairly certain results."
The Web Services Glue
CIOs who have the benefit of a clean slate can avoid some of the big three-letter application suites from the get-go. At Cubist Pharmaceuticals, a Lexington, Mass.-based biotech company that develops antimicrobial drugs that fight life-threatening hospital infections, IT Director Kelly Schmitz is creating a flexible architecture that will support the company's current financial systems and also its future sales, data warehousing, clinical trial management and CRM systems. And he sees Web services - not enterprise application suites - as the way to accomplish that.
Cubist does use Oracle's 11i ERP suite for financials, human resources and some manufacturing functions, Schmitz says, but the implementation will stop short of sales-force automation. "It's not that we're dissatisfied with Oracle's solution; it's just that their sales solution doesn't work for pharma-based sales," Schmitz says. "It's more geared toward the sale of widgets." In contrast to sales effort sat a manufacturing company that sells products to individuals, Cubist's sales staff tries to get hospitals excited about the potential benefits of its investigational drug Cidecin (for which Cubist filed a new drug application with the Food and Drug Administration at the end of 2002) - from getting them involved in clinical trials to selling them the drug once it's approved.
Because Cubist is more concerned with its relationships with entire hospitals than individual doctors, Schmitz is looking for an account management rather than contact management tool. Oracle's offering in this area, which focuses on the analysis of person-to-person encounters, would not, in Schmitz's estimation, help Cubist track its success in selling hospital staff on the value of its drug. It's a situation that arises from time to time, according to John Wookey, Oracle's senior vice president of application development. "Our applications won't always fit the specific needs of every customer," he admits.
The challenge for Schmitz will be in integrating an account management app with various outside sources of competitive sales data and hospital-specific sales numbers, such as IMS and Verispan, as well as applications to be introduced in the next year to automate clinical trial management, medical affairs and marketing campaign management. Web services, Schmitz says, is designed to standardize the method by which all of these applications exchange data and provide platform-independent EAI.
"I have seven new point applications I need to roll out in the next year that will need to talk to each other. Web services can tie together applications that were never made to interact with each other," explains Schmitz, who is using a Web services platform from Reston, Va.-based Dimension Data for a pilot application. Schmitz had success last year using Web services to tie Microsoft Outlook's resource scheduling function to Cubist's conference room phones to track meetings and make better use of the companies' limited meeting space. He notes that a Web services approach tackles integration issues at about 75 percent of the cost of traditional middleware.
The standards-based Web services approach also makes scalability less of an issue than it is with the big enterprise application solution. Schmitz previously worked at PerkinElmer Life Sciences, an Oracle pharmaceutical shop, where he witnessed big problems (and even bigger costs) as the enterprise acquired more companies and tried to integrate a large number of customized applications. "Being in a much smaller company now with a huge potential for growth, Web services is very attractive," he says. "We can take a modular approach to delivering systems and simply plug in applications as they are completed, knowing that Web services will allow data exchange in real-time."
Compared with the traditional method of connecting point solutions via an interface, which must be done from scratch each time a new application is introduced, or implementing a big suite of applications that requires a CIO to rip out existing applications, Web services can be a less complex and costly proposition, Schmitz says. Once one application has been exposed as Web services, those services can be accessed in a standard way by any other application and on any other platform. Thus, Web services allows IT executives to leverage what's already in place.
"One of the values people are seeing in Web services technology is that, unlike enterprise application software, it's an overlay on existing technology," Hagel says. "There's a much smaller investment required than when you're forcing people to implement all new applications."
Web services standards are still evolving and security remains an issue, but CIOs may be willing to make that trade-off for increased flexibility. "It's relatively new technology, but something like Web services is very attractive," says Steve Morelli, senior vice president of strategic planning, business development and IT for San Francisco-based Del Monte.
Schmitz predicts that more CIOs will eventually move to the component-based systems made possible by Web services. "As people begin to build out on an enterprise wide scale using best-of-breed technology in every category, they're going to achieve a level of reliability, flexibility and speed that the monolithic systems will never be able to achieve," he says.