Two Decades Later, Microsoft Looks Like Big Blue Of Days Gone By
July 15, 2013 Timothy Prickett Morgan
An expected and sweeping reorganization at Microsoft last week probably has more than a few IBMers and IBM watchers having not only flashbacks, but also getting a good chuckle as CEO Steve Ballmer is facing his own John Akers moment.
So many incumbents in the IT racket have missed the boat on hyperscale web technology to drive search and other application services as well as the transition from PCs to smartphones and tablets as endpoint devices that it is laughable to single Microsoft out for scorn. But the fact remains that Microsoft has spent untold billions trying to build a search engine and cloud computing business and was even frustrated enough with its PC partners to launch its own Windows 8-powered Surface tablets–without making much of a dent in its competitors.
In case you have forgotten, John Akers is the boy wonder salesman at IBM who rose through the ranks at Big Blue to become its president, then its CEO and chairman, just as a recession was brewing in the late 1980s. This was also when proprietary minicomputers were really taking off and Unix systems were getting ready to upset the systems applecart. And ironically, this was also when IBM decided it could get away with a massive increase in the price of mainframe capacity, something that the market so aptly demonstrated was not the case when companies started to consolidate mainframe data centers and then, after looking at alternatives, starting up new workloads on non-mainframe machines. By 1993, IBM was in such a mess that it fired nearly half of its global workforce over the prior several years and brought in Louis Gerstner to get IBM to focus is research and development on solving real IT problems, got its various systems fiefdoms to share R&D and converge product lines, and ramped up services as a business. Gerstner and his protégé, Sam Palmisano, backed away from making disk drives, PCs, high-end printers, memory chips, point of sale gear, and a slew of other low-margin products.
Here we are in 2013, and IBM certainly has its issues. But still, the incumbents in the IT industry all aspire, depending on where they play, to be like IBM, Apple, or Google. Microsoft wants to be IBM on its systems platform, controlling every aspect of it and offering its wares for private consumption or as a rental on the Windows Azure public cloud, harkening back to the service bureau days of mainframes. It wants to be like Apple in creating snazzy and profitable end user devices and applications that run on them and its Azure cloud, and it wants to be Google by creating sophisticated and efficient data centers that can host its own applications for customers–such as Office365, SharePoint, Exhange Server, or SQL Server–or those created by end user customers.
And so, last week, embattled CEO Ballmer, who has been with company co-founder Bill Gates from nearly Day One at Microsoft, is making what many think will be his last stand with his “One Microsoft” realignment.
The eight fiefdoms that have defined Microsoft for so long are now gone, and Ballmer has created a single Operating Systems Engineering Group that will focus on every variation of Windows on every device. Say goodbye to the Server and Tools Business unit that has shepherded Windows Server and related systems software. All of the adjuncts that ride atop Windows Server are now over in the Cloud and Enterprise Engineering Group, which will not only do software for your data center but also design and build Microsoft’s own data centers. Office and other applications end up in the Applications and Services Engineering Group, while the Dynamics application suite remains a separate entity. The Devices and Studios Engineering Group will handle the design, supply chain, and manufacturing for the Xbox, Surface, and any other device Microsoft comes up with as well as games to run on PCs, tablets, smartphones, and game consoles.
I am not going to pretend I know the personalities that are in charge of the new divisions at Microsoft, or the ramifications of some of the exits of some key employees in the wake of the reorganization. Microsoft burns people out, just like IBM and Google and Amazon do, so I don’t put a lot of stock into such analysis. Every organization shifts executives around, and sometimes it is for cross-training and sometimes it is to get a fresh perspective in there to solve a problem. When people leave unexpectedly, this can be a different thing entirely, of course. And there was a little of that in the Microsoft reorganization last week.
As Gerstner did with IBM Research two decades ago, Ballmer is getting all of the new product development groups hooked into Microsoft Research, which does a lot of pie-in-the-sky stuff as well as more practical IT work. There is probably a mandate for Microsoft Research to not get too much pie in its sky, but the note from Ballmer to Microsoft employees didn’t say that explicitly.
You probably noticed something in the names of the four core groups: They are engineering groups. Meaning, those execs are now responsible for building great products, and doing it faster, and that is it. Marketing, business development, research, and finance all have their own cross-company groups, and that frees the executives in charge of making products from those four key engineering groups from worrying about how well or poorly those products are selling.
We will know if this was a good idea or not in about three or four years. I am skeptical of such changes, personally. I saw too many IBM reorganizations in the late 1980s and early 1990s to have much stock in them, and read about too many others that happened before then. I think this often has more to do with the personalities of the top brass at the company as opposed to what works or what does not. What I can tell you is that a great product happens, gets to market, and sells well–often despite the best efforts of the company to help it fail. For Exhibit A, I need only point out the AS/400, launched 25 years ago, as beloved as a red-headed stepchild and the last of the proprietary minis out there in the world with a reasonably healthy customer base and something that looks like a future.