How to Turn Microsoft Around: A Primer, Part 1

Share it on Twitter  
Share it on Facebook  
Share it on Linked in  

One of the first things I read this morning as I opened my eyes to face the first week of 2010 was a piece in InformationWeek on the seven things Microsoft Must Do in 2010. Each was a good idea, but each -- and this isn't unusual -- addressed symptoms of an undisclosed core problem and didn't try to determine if there was a central cause for the mistakes, assuming they are mistakes, in the first place.


I'm calling this post a primer because while this should likely be a book, I'm going to cover the issues in blog form, so it will be light and relatively quick. This piece dovetails somewhat with my final post for 2009, dealing with Apple, Google and Microsoft CEOs.


Avoiding the Whack-a-Mole Approach


Going after the symptoms isn't unusual. Often, a new chief executive comes in and plays a few years of whack-a-mole, pounding on highly visible problems, then either sells the company, stabilizes it or gets replaced. Rarely does he or she ever restore the company to the greatness it once enjoyed. To do that, you have to put your mole-whacking stick down and take time to study the company for a while. Few are given this kind of time and those that are generally focus, like our fictional whack-a-mole CEO, on the more topical moles. Whacking moles also tends to be less effort and much more fun.


Since I'd like this decade to be a bit different, I'm going to approach this topic in another way. Instead of whacking moles, let's instead see what's getting them to pop up out of their holes in the first place.


Microsoft, the Good Years: Defined Customers


If we were trying to analyze a problem in a car or a person, it would be good to look back at when it was working well, compare that to what is going on now, and see what changes are degrading performance.


Microsoft seemed to peak in 1995. Its big break was when IBM licensed DOS for a pittance and Microsoft helped create the PC OEMs that today make up the majority of PC sales. The customers were well defined (they were these original equipment manufacturers that had effectively outsourced software to Microsoft). A side lesson was learned by IBM, in that by buying DOS so cheaply, IBM devolved from Microsoft's most important to least important partner, setting the groundwork for their eventual breakup. (I wonder if IBM retains this lesson.)


This OEM relationship made Microsoft the vendor and the OEMs the customer, with clearly defined roles and little conflict. Microsoft Office was, particularly with the peak 1995 launch, subordinated to Windows, and even though Office often was sold directly to companies and consumers, it was seen as a PC enhancement, strengthening the OEM/Microsoft bonds. PC turnover went from eight to 10 years to two to three years over the '90s, and this was a lucrative market for everyone.

Microsoft, the Bad Years: Servers and Confusion


Unlike with PCs, many of the OEMs had a server business that consisted of software largely subordinated to hardware that was either proprietary or UNIX-based. Microsoft rolled out Windows NT Server in direct competition with these platforms. I can recall one meeting I had as an analyst for Giga with Hewlett-Packard's senior management; they wanted me to try to explain to Microsoft why they couldn't do with Windows NT what Microsoft wanted because HP had too much invested in HPUX. We clearly had a fractured relationship because no vendor should be forcing its customer to do something they don't want to do, yet this was the problem HP was describing.


What the OEMs needed, because they were badly fragmented, was a centralized UNIX. But Microsoft provided something different -- a product that came from inside Microsoft that appeared to completely ignore its primary customers, the OEMs. Conflict resulted. Microsoft then went around the OEMs and created demand for this product with the OEMs' customers, the IT businesses with which the OEMs dealt. The OEMs effectively went from being customers of Microsoft to being its vendors and competitors, but the vendor relationship (building servers for Windows Server) was never finalized and remains largely in flux. In short, the vendors should have become ODMs, which design and manufacture products for another company, with the servers carrying Microsoft's brand, because Microsoft, not the OEMs, was calling the key shots.


Many customers also had wanted a better UNIX. This disconnect between what the OEMs wanted, what the customers wanted, and a customer/vendor relationship in flux created a fertile breading ground for an alternative offering, and Linux was born. Still, this wasn't ideal because it, too, fragmented. It didn't throw off much profit or marketing and seemed to fuel outsourcing, which left the door open for Google and its public cloud, Android and ChromeOS efforts. Microsoft and industry growth slowed substantially on all fronts.


Office separated from Windows and ran just fine on older hardware. Windows Vista and Windows 7 were sold as a way to extend the life of older hardware. Xbox competes with PCs for dollars, and Zune abandoned the OEM approach for a consumer-centric one. This showcases a company caught, and partially stuck, in a transition from one that is a vendor to others to one that is a solutions provider itself.


Customer/Vendor Confusion


Therefore, I maintain that at the core of Microsoft's problem is role confusion between Microsoft and its large partners. For example, during the Windows Vista launch, Intel expressed concern that its imbedded graphics processors wouldn't comply with the "Vista Capable" logo program. It was treated like a customer and the program was adjusted, resulting in breakage and litigation. Had Intel been treated like another vendor, which is what it actually was, Intel would have had to make the adjustment and no litigation should have resulted.


This doesn't suggest that vendors should be ignored, only that they be made subordinate to customer needs and that related conflicts be resolved favoring the customer. Whoever the "customer" is. And Microsoft is in flux with regard to who that customer is.


Wrapping Up


In my next post, I'll write about how to fix this, but I leave you with a brief definition of what I think the problem is.


Microsoft has four groups that are being treated in balance, but that should be in a hierarchy. They consist of end users, OEMs, IT and critical partners (Intel, EMC, Cisco, AMD, Nvidia, etc.). The resulting confusion over ranking results in products that are not really focused on any one of these groups, largely incomplete or counter-strategic from the perspective of any of them, and that generally underperform their market potential. Apple, Acer, IBM and EMC all seem to have a much tighter focus on who their customers are and appear better able to expand their companies and maintain higher customer satisfaction as a result. Microsoft has one question it needs to answer: Who is your primary customer? Its answer of "all of the above" is currently getting the company a "C-minus" on the test.