Last decade, we really saw the difference between how a pyramid executive hierarchy could perform against a company made up of silos, or pillars. This was the battle between Microsoft and Apple. What is interesting is that if you go back a decade to when Microsoft was kicking Apple’s butt, Microsoft was organized into a pyramid hierarchy and Apple had fragmented into a loose set of divisions that weren’t competitive. Jobs seemed to understand this better than Gates did, flipping to Microsoft’s old model while Microsoft flipped to IBM’s or GE’s.
IBM was the quintessential silo company, but something funny happened last decade: It got rid of or collapsed the silos. Suddenly, it is a hierarchical firm again, leaving HP, which modeled itself after the old IBM, as one of the few remaining umbrella companies in the technology industry. While HP is clearly off its highs, Sony is the example of the worst umbrella company, with a decade during which the CEO of the firm didn’t seem to have any real authority.
With Microsoft abandoning the silo structure and HP getting three new board members this month, I think we are in for some changes in both companies, and I think the contrast in organizational structure is at the core of both firms’ problems. You might conclude, given my examples, that the silo approach is one I’m going to argue against, and generally, you’d be right. EMC, though, has showcased that there is a way to go down this path that might actually work.
Let’s explore tech company structure this week.
Silos = Infighting
Like a lot of bad policies (forced ranking being one of the worst), I tend to blame the move to silos on GE and Jack Welch. But, in the tech industry, I could just as easily point to IBM, which moved to a similar methodology while still dominant and before this approach to corporate governance and executive structure became far more common in the tech industry in the ‘80s and ‘90s. The structure appeared heavily favored out of major business schools and it appeared well founded. The idea was to create nearly fully functioning businesses within a complex company. Each business would have a certain amount of autonomy, and its own staff functions (marketing, finance, etc.). The belief was that this would make each far more agile because decisions would be made far closer to the customer.
It was a good idea but the problem was that each of these businesses then had to effectively give some or all their revenue to the corporate executive team, who would redistribute it in a way virtually none of the executives would think was fair. This created a win/lose scenario, where the general managers of these units focused on hurting each other, or at least avoided helping, because they were in competition for funding and people. They weren’t really independent; they were more like kids on an allowance managed by parents who often had their own very different agenda.
In Sony, you saw this play out when the PC group partnered more tightly with Microsoft’s Xbox team than it did PlayStation, and when the media content group required so much DRM protection on Sony media hardware that it was pretty much unusable in the early parts of last decade (which is how Apple beat Sony decisively with the iPod). In Microsoft, you saw Office catapult Windows versions through XP, then work against them from Vista on, and an Xbox effort that crippled PC gaming.
The mistakes were in two areas: First, the silos weren’t truly independent, and then the focus was pulled too far from corporate needs to favor those of the divisions, with catastrophic results.
Microsoft’s Apple Moment
Steve Ballmer rightly saw that the company structure wasn’t working. He has reformed Microsoft to be more hierarchical, effectively eliminating much of the divisional conflict. Granted, until he gets rid of forced ranking, he’ll still have inefficiencies at the employee level, but this should stop Xbox and Office Groups from hurting Windows and Microsoft from being its own worst enemy. Which begs the question: Could HP go the same way, particularly given that IBM preceded Microsoft into making this move?
EMC suggests a different path — one that shows that silos can be done right.
EMC, with properties ranging from RSA, VMware, VCE, and most recently Pivotal, pushed the silo approach to its more effective conclusion. Each of these properties is a part of the EMC family of firms, but they aren’t on an allowance. Each is a fully functioning company, in some cases wholly owned (RSA), in some cases semi-public (VMware), and in some cases owned with partners (VCE, Pivotal). This structure appears to provide the anticipated benefits of a silo structure without many of the problems. The resulting unit functions better than close partnerships, but not as well as a fully integrated firm, and likely works best for units that don’t need to be closely aligned.
Wrapping Up: HP’s Possible Lesson
HP has pretty much replaced its board, by my count, for about the fourth time in a decade. This new board would do well to look at the lessons taught by Apple, IBM, Microsoft, and particularly EMC to provide a faster way to recovery for the company. Silos just aren’t working for anyone in the traditional sense, at least not in the technology market, and many technology firms are successfully implementing alternatives. Dell is even moving to go private. We’ll know later this week if Dell’s move will be successful. But one thing is clear: HP should change its current structure into one that has a better chance of success in order to advance to its full potential.
Organizational structure is now being used as a competitive tool, with EMC still the most creative. In the end, the moves by IBM, EMC and Microsoft make these companies more agile, more focused and responsive to customers, and more able to anticipate competitor moves largely because they aren’t primarily focused on inside politics. That’s certainly a good thing for both customers and investors.