Fixing HP: The Back Story

With the recent news surrounding Autonomy, it isn’t hard to see HP as some kind of train wreck. But much of HP’s ongoing business remains relatively strong and profitable. I believe the core of HP’s problem is instability at the top, but even this is more a symptom than a cause. In my opinion, the cause lies with successive boards who thought tactically to address strategic HP problems and who put speed over quality. Let me explain.

It all started with Carly Fiorina.

Fixing “Boring”

For most of HP’s initial life through the 1990s, HP was a “boring” company. It had few exciting products, but it had solid financial performance. Its only real hit was a calculator. And even though folks lined up to see its Sojourn PC, it was easily overshadowed by industry darlings such as Sony and Microsoft. At that point, even Apple was getting more interest from investors and consumers.

In the late 90s, HP’s board decided the firm needed a shake up, and they brought on board Carly Fiorina from Lucent to do it. Now, in my opinion, this was a fundamental mistake. What HP had was largely a marketing and consumer product problem. The company wasn’t going under — it just wasn’t exciting consumers or investors. It didn’t need to be fixed operationally, it needed an image makeover and at least one hit product that, unlike the Sojourn, folks could afford.

This meant they needed a new CMO and likely a new head or a new product manager for their PC division. Instead, they got a new CEO, and they picked badly.

You see, over the past decade or so, there had been two catastrophic acquisition failures in the computer industry: IBM’s acquisition of ROLM and AT&T’s acquisition of NCR. Many felt these failures made it clear that telephony people couldn’t run computer companies and that computer people couldn’t run telephony companies. In addition, marketing people generally don’t have the operational background to run corporations. Fiorina was a marketing executive from a telephony company. On paper, she couldn’t be successful, and the board was fixing the HP problem at the wrong level.

Ironically, I believe Fiorina might have actually done well at the right level because she seemed to intuitively understand HP’s need for marketing and hit products. But running all of HP was beyond her. Add to this that she made herself executive chairman, which effectively made the board her subordinate, and the board couldn’t manage her. It wasn’t that she failed; it was that no one put in that position with her background could succeed.

Fixing Ops

Fiorina was clearly having problems guiding HP, and she moved to acquire Compaq without the approval of powerful board members, resulting in a proxy fight. Ironically, this proxy fight forced her and the CEO from Compaq, an ops specialist (Michael Capellas), to put in place one of the strongest integration plans ever developed, and the acquisition was successful. The conclusion was that HP needed a strong ops person. Capellas appeared ideal. But many believe Fiorina rightly saw him as a threat to her job and in her executive chairman position couldn’t be stopped from forcing him out of the firm.

She was subsequently fired, and Mark Hurd from NCR was hired to replace her–another bad choice, in my opinion. The issue wasn’t that Hurd didn’t have the right pedigree; it was that his skill was largely cutting costs. Fiorina had already cut the company heavily and had broken employee loyalty. Basically, she had isolated herself from the rank and file. While HP needed an ops person, it needed someone who could foster loyalty and rebuild the company. Instead, it got someone who had a history of being hated by his employees (Hurd reportedly had his tires slashed at NCR) and who would have been more appropriate for a firm that was failing and needed to be packaged for a quick sale.

So again, rather than getting a good COO or a builder to fix an ops problem, HP got a CEO who was a cutter. While this did improve HP’s financials, it came at the expense of HP’s long-term survival. And it restored the initial problem that HP was suffering from — the lack of exciting products.

Repeating the Problem

Many credit Hurd with brilliantly creating a situation where, by effectively throwing the then HP chairman under a bus, he could become executive chairman. This repeated the Fiorina mistake and effectively made him his own boss. Many thought his history at NCR suggested a tendency to abuse power, and he allegedly tried to get HP to fund a mistress. The situation ended catastrophically with his sudden departure.

At that point, yet another HP board was faced with hiring a new CEO. Once again, they were faced with a nearly identical problem: they needed to restore excitement to HP. And in my opinion, they approached this problem even more catastrophically.

The Oracle Problem

You see, Oracle had through partnership effectively served as HP’s software arm in competition with IBM. But Oracle decided to get a divorce. Oracle bought Sun, a failing hardware company, in order to do so. Oracle then hired Hurd to run this near disaster. Suddenly, HP was forced into competition with both IBM and one-time partner Oracle, which was now led in part by HP’s last CEO.

The HP board was likely a bit panicked. But instead of hiring a competent head of software (Hurd had gutted HP software during his tenure) and then rebuilding that unit under a new CEO qualified to run HP, they once again over-corrected. They hired Leo Apotheker, the failed CEO of SAP. He might have been fine running a software division but wasn’t at all suited to run a firm that was mostly hardware.

This decision was almost immediately catastrophic. First Oracle moved to have Apotheker become a key witness in their ongoing litigation against SAP. Second, many of the financial analysts and investors believed that he was a bad choice to run HP. And third, he was put in a position of both having to dodge Oracle’s attempts to get him into court and having to execute a bold strategy to create an HP software division.

This situation had all of the earmarks of a massive disaster, and that is clearly what resulted. HP’s Palm acquisition failed. And it now appears the Autonomy acquisition was rushed and can be viewed as a failure as well. Apotheker may go down in history as doing more damage to a company in a short time than any other CEO in history. But the fault wasn’t his; it was the fault of those who selected him.

Wrapping Up: The Lasting Lesson

The problem with HP wasn’t the CEOs. The problem was with their selection.

In fact, the selection was almost always overkill. Where they needed a scalpel, they used a table saw. First, they needed stronger marketing and more compelling consumer products; they over-corrected with a new and mismatched CEO. When they needed ops, instead of a strong COO, they replaced the CEO again. And when they needed a software head, they once again replaced the head of the company with a person with that specialty. It wasn’t that the board didn’t understand the problem; it was that they forgot to protect the majority of the firm, which wasn’t broken and over-corrected. In short, they constantly traded one problem for another set of problems that were worse.

I believe Whitman is actually a good match for what they need at the top of HP. But she inherited the combined mess that HP’s boards and CEOs had created over a decade of false starts and boneheaded staffing decisions. She now has the marketing and product problem of Fiorina, the ops problem of Hurd and the software problem of Apotheker. She can turn the company around, but only if they systematically fix each of these components. Given HP’s history, it seems more likely the firm will replace her — and that would simply continue this cycle.

Whitman’s path should be similar to Gerstner’s at IBM. Fix HP’s image first. Then put division heads in place that can execute and slim the company down for combat.

HP can be fixed. It just needs to be fixed systematically and in parts. In short, if the body is sick you don’t repeatedly replace the head, largely because that will end badly; instead, you focus on the sick parts. And that is how you’d fix HP.