March 11, 1998 was a banner day for Hewlett-Packard Co. At a trade show in Los Angeles, CEO Lewis Platt announced Electronic World, which HP officials touted as a "far-reaching strategy to harness the utility of the Internet with the company's full breadth of solutions."
A year later, E-World was dead, usurped by a new initiative called e-services. Hewlett-Packard had replaced more than just the E-World moniker, however. It also had killed or significantly repositioned key projects once considered critical to the strategy. In the process, HP alienated customers and partners that had bought into the company's vision.
The about-face points to a larger problem for the nearly $40 billion computer manufacturer. E-World was doomed by a decentralized culture of autonomous product divisions concerned more with selling hardware and preserving partnerships than supporting corporatewide strategic initiatives.
E-services, said an HP executive who requested anonymity, "may very well be a train wreck in the long history of train wrecks."
HP is going to great lengths to make sure history doesn't repeat itself. Carly Fiorina, a former Lucent Technologies Inc. executive who took over as HP's CEO in August, has promised to break down the walls between HP's product divisions to foster common long-term goals that realize the company's Internet vision. (HP officials said Fiorina and Ann Livermore, president of Enterprise Computing Solutions, were unavailable to comment for this article.) This week, the company will host its first e-services developers conference to drive home that message.
Stakes are high
HP is in a pitched battle with IBM, Sun Microsystems Inc., Compaq Computer Corp. and other major computer companies remaking themselves as providers of solutions comprising hardware, software and services for the Internet economy.
So far, HP has struggled with that transformation. Software has often been a second-class citizen at HP, existing primarily as a means to sell more hardware. Over the years, executives have been quick to dismantle or narrow the focus of software projects that didn't yield hardware sales or meet unrealistic revenue goals, according to several current and former HP executives interviewed for this story.
Sources close to the situation said the company continues to grapple with uncoordinated product and sales efforts, lack of understanding of software solutions needs, and infighting.
Adding to the problem of hardware and service groups' failure to support HP's own software, sources said, has been top management's reluctance to tie the disparate groups together more closely.
"It's a highly competitive atmosphere that is extremely political," said Dave McGoveran, a longtime HP consultant and principal at Alternative Technologies, in Boulder Creek, Calif. "When you have to be aware of whose toes you're stepping on, how can you succeed?"
When E-World debuted, HP singled out three products as critical to the initiative: OpenPix, a family of technologies for high-quality Web-based printing and viewing of images; Changengine, a workflow automation product; and VerSecure, a technology for strong encryption. All three initiatives are either dead or significantly scaled back from their original objectives.
HP killed OpenPix this spring, despite a string of product rollouts and customer deployments that were well under way. HP dismantled OpenPix because it "didn't meet revenue goals," said Joe Beyers, vice president of HP's Internet Business Unit, in Palo Alto, Calif. HP has since spread some of the technology's components throughout the Internet Business Unit and the Inkjet Imaging Solutions Division.
Changengine nearly suffered the same fate. HP executives decided to kill the product along with OpenPix this spring, before a Changengine manager made a last-ditch plea to save the technology. It was subsequently transferred to the Communications Industry Business Unit and repositioned as a component of the company's enterprise application integration solution a far cry from its stature as "a foundation for customer e-business solutions," as a 1997 HP press release referred to it.
John Dillon, worldwide sales and marketing manager for Changengine, maintains that the product still has strong support throughout HP. "There's a lot of history and baggage, but all the groups are behind Changengine and e-services," Dillon said. He acknowledged, however, that as part of CIBU, "we are not an integral part of e-services, and that's one thing we need to improve on."
Developers in a bind
HP's failure with Changengine, which Dillon admits was "immature" when it debuted, stymied developers that had signed deals to use the technology.
"It's very frustrating to play musical chairs with a company that one month says one thing and the next month does another," said Kevin Kilroy, CEO of Bluestone Software Inc.
HP paid Bluestone, of Mount Laurel, N.J., $1 million last year to develop adapters for Changengine that would enable it to plug into middleware and Web and enterprise applications. As part of the deal, HP agreed to write an API to work with Bluestone's adapter.
Bluestone held up its end of the bargain, delivering its adapters as promised, but HP never finished its work. Bluestone never received any royalties from Changengine sales.
"HP's software organization is bureaucracy at its finest," Kilroy said. "They could have done other things with Changengine, but because they're afraid they'd compete with one of their database partners, they decide not to do anything."
The partnership fizzled, even though HP and Bluestone had held discussions in 1998 about a possible merger, sources said. In September, HP signed up Bluestone competitor Active Software Inc. to develop the same adapters that Bluestone had originally been hired to create. "Bluestone provided a certain set of solutions, but it was time to think bigger and make it core to the whole business at HP," HP's Beyers said.
Not the only one
Bluestone is not the only partner to fall out of favor with HP. In the spring of 1998, Teamware, the London-based subsidiary of Fujitsu Ltd., agreed to develop a graphical programming tool, called ProcessWise Workbench, for Changengine. The deal called for Teamware to receive a percentage of revenues from sales of Changengine, sources said.
But because Changengine sales were so low, Teamware never realized significant revenues and dropped its Changengine development work in March.
Changengine 3.0, released earlier this year, includes Fujitsu's technology, but the break with Teamware contributed to HP's missing a planned October delivery date for Changengine 4.0. The upgrade is now due next month.
Changengine was hamstrung by more than just immature technology and broken partnerships, according to a source familiar with the situation. HP's professional services group didn't want to market Changengine because it might conflict with deals the group was trying to put together with other companies, the source said.
"It's a battle," the source said. "We can't make [software] things happen without hardware or services. Every time I wanted their support, I got push-back."
Internal lack of support
The conflict grew to the point where HP software executives threatened last year to partner with IBM Global Services on HP's software initiatives in response to the lack of support from their own services organization.
The executives never carried through on their threat, but observers said they believe the internal conflicts remain. "The structure constantly beats down the software [divisions] in the company," said one recently departed software executive. "It's a system that grinds you down and out.
"There are all these software project cancellations, and as a result, a lot of skepticism," the former executive continued. "I'd be giving a presentation on a new [software product] to a customer, and they'd ask me when I was going to cancel it."
HP executives admit they've had trouble bringing their Internet vision into focusuntil recently. "E-World was not a coordinated business plan," said Nick Earle, chief marketing officer of e-services. "People were right to question our capability."
Said Beyers, "The difference now is there is real meat around the [e-services] strategy."
This week, HP will host its first e-services developers conference, where it will discuss pilot programs and its intention to make e-speak APIs available under open source by year's end. E-speak, which HP announced in May, is an object-oriented framework, written in Java, for developing and deploying e-ser vices.
More spending on software
As evidence of the company's commitment to software, Beyers said HP has tripled spending on its software business since early 1998. The exact figure, along with HP's software revenues, are a well-kept secret. HP does not segment software sales from the rest of its enterprise computing revenues, as does IBM.
To be sure, HP has had its share of successful software. Beyers boasts that 70 percent of all Web sites worldwide are managed with OpenView, and he said that sales of the Praesidium line of security products have grown about 800 percent since this time last year. He also cited growing interest in Web QoS, a quality-of-service technology that prioritizes Internet customers by placing them in a queue.
In addition, MC/ServiceGuard, a high-availability solution for clustering and failover, is ringing up sales of $100 million annually, said Roberto Medrano, general manager of e-services.
Organizationally, HP continues to look for the right formula for its myriad Internet divisions (see chart, left). It's in the process of forming yet another organization, under Earle, to oversee HP's products, services and consulting efforts.
The group will be charged with forging alliances worth at least tens of millions of dollars to HP. It will explore additional partnerships in which it will provide hardware and software in exchange for an equity stake or a cut of the company's revenues (see story, Page 20).
"Right now, we're not using all of our strengths to win the big deals," Earle said.
List of failures
In the path of these success stories, however, lies a string of failed projects. In addition to OpenPix, the company has pulled the plug on SmartService, a process engine for telecommunications, which was killed last year, and Chain Reaction, a middleware message broker that was killed late last year, shortly before its scheduled delivery date.
HP's VerSecure encryption technology has suffered more at the hands of a changing marketplace than by any of HP's moves. With the U.S. government easing restrictions on exporting encryption, HP "is not pushing [VerSecure] as a technology into the marketplace but rearchitecting it for content management over the Internet," Beyers said.
One of HP's current partners has a simple explanation for the company's missteps.
"HP is not a software com pany," said Susan Locker, CEO of Coda Enterprises Inc., of Mountain View, Calif., which works with HP on products based on Changengine. "The people who come into the software divisions have mostly worked in hardware, and they're not problem solvers."
"[HP has] had a checkered past, riding on the brand loyalty related to hardware," said Tom Willmott, president of Aberdeen Group, based in Boston. "They've just begun to get their act together."
In August 1998, HP hired Aber deen to analyze the company's strengths and weaknesses in the enterprise space. Aberdeen's report, "Hewlett-Packard: Failing to Execute on Its Enterprise-Trusted Supplier Strategy," sharply criticized the company's past software follies and questioned HP's commitment to enterprise customers and software projects. Its conclusion: HP "must quickly make some very difficult and significant changes in the way it services enterprise customers."
In some areas, HP appears to have listened. It has invested in middleware companies such as BEA Systems Inc. and Broad Vision Inc. and reconstructed its partnering strategy to be more focused on a handful of companies.
Many observers, both inside and outside the company, said HP still has a lot of work to do.
"HP's software efforts have historically been rooted in a division that is not tuned to the needs of the market," McGov eran said. "Carly has to create an organization that understands that innovation ... cuts across boundaries. If she can't do that, HP doesn't have a chance at becoming an enterprise software player."