Dear Meg Whitman… Some unsolicited advice on HP’s PC future

The passing of Steve Jobs has me thinking about another Silicon Valley icon whose soul has left this world. I'm speaking, of course, about Hewlett-Packard. 

Hewlett-Packard's CEO Meg Whitman said last week that she wants to make a decision on the fate of HP's personal systems division by the end of the month. Speaking at Fortune magazine's Most Powerful Women Summit, Whitman said that the fate of the PC group is a decision "I want to make much faster than my predecessor," according to a report by Bloomberg Businessweek.

Not that Whitman is going to listen to me, but I'd like to offer some unsolicited advice: it's time to get out of the PC business. If you still have to ask the question of what to do with the PC business despite being the market leader globally in PC sales, then get out now and sell the division to someone who cares.

I'm no HP hater. The very first computer (or at least the first real computer) that inspired unbridled techno lust in me was an HP Apollo PA-RISC workstation. I had a similar reaction not too long ago to an HP Z800 workstation. HP can still make machines that make an engineer's pulse quicken. The question is whether they have the soul left to do it—or more properly, whether the board of HP still has a pulse.

A question of identity

The real question Whitman has to answer is bigger than what to do with HP's personal systems group—a business unit that still somehow sells more desktops and laptops than anyone else in the world. That question is what kind of company HP is now, and what it should become.

It's not clear how much of that decision is in Whitman's hands at this point, given that HP's board seems to have signed some sort of corporate suicide pact. When HP's board hired Leo Apotheker last year, it seemed pretty clear what they wanted to be: they wanted to be more like IBM, which shed its PC business, aggressively downsized, and focused on big iron, big software, and big services. 

Since 1999—the year that HP's last engineer CEO, Lew Platt, retired—the company has acquired and cast off businesses with such frequency that almost the only thing that HP now has in common with the company Bill Hewlett and Dave Packard founded in their legendary Palo Alto garage is the name. But perhaps the biggest change has been the acquisition of Electronic Data Systems, the IT managed services business born out of General Motors. Since that acquisition, HP's services unit has changed the very nature the company.

The cold, hard numbers

In the quarter ending in July of this year,  more than a third of HP's earnings before taxes came from services (for those who care, $1.2 billion in earnings on $9 billion in net revenue). If Whitman is judging purely on profitability, the personal systems group (which earned $567 million on revenues of  $9.6 billion) is not exactly a winner. The company earned more than that just selling ink and toner.

To stay competitive in the PC business, HP has cut corners. Most of its systems, both enterprise and personal, are made by contract manufacturers. 

Sure, nearly everyone else in the PC business does now, including Apple. But other companies apparently do a better job of managing quality. HP pays contract manufacturers to make PCs—particularly laptops—that come in fifth place in reliability. And they're barely hanging onto that position—Samsung is within reach of pushing them down the stack further.

HP still holds a commanding lead in global market share for PC sales. But its edge and share are slipping.  In the US, the company's share of new PC sales has dropped 2.5 percent in the last year. HP isn't even in the top five in PC sales in China, a market that just passed the US in size in the last year, and one that Lenovo dominates.

Then there's the tablet business. Mark Hurd saw an opportunity in mobile devices when HP acquired Palm last year. But with HP killing off the TouchPad, its Windows tablet seemingly going nowhere fast, and kicking almost all of its internal mobile expertise to the curb, the company has divorced itself from the fastest emerging market for personal computing.

HP's PC business is hardly doomed. In the right hands, it's still a viable business, even with shrinking margins. But it's not clear that HP has the skill sets, or the desire, to raise its game and make it as profitable as it could be—mostly because the big margins of enterprise servers, software, and services look like easier money. That was what was behind Apotheker's drive to turn HP into a company that looked like IBM.

What HP can't be: IBM

IBM's model for success is based on how well the company has balanced its enterprise hardware, software, and services businesses. They all feed into each other: "Like DB2? You'll love DB2 on Power. And let us send some consultants over to help you move everything to DB2 on Power." Unfortunately, if HP wants to be IBM, it's missing some significant pieces.

First, it's missing enterprise software. While HP has some assets in IT systems and network management software—such as HP's Open View unit—it has killed just about every other developer platform or general-purpose enterprise software line it had over the last 10 years. Just before the Compaq acquisition, for example, HP bought a middleware company—Bluestone—and then promptly tossed it out in the trash; what's left of Bluestone is now part of Red Hat's JBoss unit. Ironically, right after tossing HP Middleware, HP signed on to provide enterprise support for JBoss.

The pieces that HP has acquired recently aren't exactly the base to build a big enterprise software business on. In February, HP acquired Vertica, a big-data analytics company founded by MIT's Michael Stonebraker. And then there was the nearly $12 billion acquisition of Autonomy and its voice and unstructured data search and analytics technology.

On the server side, dependence on other companies' enterprise software is creating some headaches. Oracle has dropped support for HP-UX on Intel Itanium, much to HP's chagrin. And while HP is still in second place in the Unix server market, ahead of Oracle's SPARC, its market share and server revenue is slipping.

So, to pull together the parts required to become IBM-like, HP would have to make even more acquisitions. And that's apparently not in the cards. Reuters reported on October 5 that Whitman said the Autonomy buy was "certainly the end of big acquisitions."

Meg's Choice

So, with the "be IBM" dream off the table, Whitman's options are limited. The question is whether personal computers fit with any of them, other than as a way to sell more printers for which to sell more ink, or as part of big managed services offerings. The options come down to:

1. Dig in and reaffirm HP's commitment to the PC market—at least until the next management change.
2. Spin off the personal systems division, as HP did with Agilent, and let it rise or fall as a separate business. Agilent, which HP shed in 1999, is on track to be a $5 billion business this year, and continues to innovate.
3. Sell the division off, either to internal managers through a leveraged buyout or to the contract manufacturer that already makes the systems. IBM went the latter route with Lenovo, and that company is doing a better job at producing reliable hardware than HP is.
4. Pull the plug completely, and sell the intellectual property off—or auction it on eBay.

Given the state of HP's management, I find myself hoping for either the second or third options. Anyone could be a better steward of the PC business than HP can manage right now—anyone, that is, that doesn't have to answer to the most dysfunctional corporate board on the planet.

Listing image by Photograph by Meg Whitman for Governor