A few weeks ago, as rumors that Satya Nadella would be named Microsoft CEO gained prominence, a friend and I were discussing who would replace him as head of Cloud and Enterprise. Our conclusion? Scott Guthrie. So I wasn’t particularly surprised when Scott was named interim EVP of Cloud and Enterprise, though I found the interim designation a little odd. I’m going to explore Scott’s appointment and this “interim” thing.
Scott is very popular with customers, and developers in particular, both because of his long association with Developer Division but more specifically because of his interaction style. He’s a heavy blogger, and a technical blogger at that. He had his own conference, Mix. He’d owned very popular technologies, like .NET. And once he started working on the Azure App Platform he made Azure something developers could really look at, instead of some confusing and seemingly useless Microsoft initiative. This has often led to customers calling for him to take on roles well beyond what most insiders would have considered him ready for.
So how can I on one hand consider the Cloud and Enterprise position a huge stretch for Scott, and on the other think he’s the best candidate for the position? Simple, he is the one with the vision and leadership to make Azure a success. There are other great, some more senior, some perhaps better, engineering CVPs amongst his peer group. But none of them are as likely to pull off the Cloud transition as Scott is. And that, as we say, is the high-order bit.
So let me dig a little more deeply into the problem here. Scott’s historical scope is rather narrow. He didn’t run DevDiv, he ran the .NET Runtime part of it. The Azure App Platform effort was also relatively constrained, and a very similar area. Most recently he’s been running Azure Program Management, spectacularly from everything I can tell, which covers far more ground. But it isn’t that large from a people management perspective. I don’t know the largest organization Scott has led, but it is probably under 1000 people. He’s also never owned a diverse portfolio of products. And I don’t think he has significant experience managing other CVPs, some of whom are actually SVPs by level. Put simply, running Cloud and Enterprise isn’t just a move up for Scott it is really a double move up. Both in product/service scope and organizational complexity.
Now if I’m Satya and I know Scott is my guy for moving the Cloud efforts forward then I’m going to want to set him up for success. And putting him in charge of Cloud and Enterprise as it exists today is probably setting Scott up for failure. So I know I need to make more changes, but I’m not ready to do it yet. Thus the interim EVP designation.
There are four things Satya may be considering in terms of the Cloud and Enterprise organization and how he could set Scott up for success:
The first possibility for Satya, and the one that most managers turn to in this situation, is to look for things you can take off of Scott’s plate without really changing the mission. For example, the Global Development Centers (GDC) report into Cloud and Enterprise through DevDiv head Soma Somasegar. There is no strategic reason for the GDCs to be in Cloud and Enterprise, though I think Soma really takes on all the overhead so it wouldn’t take much off Scott’s plate. Of course, you could move DevDiv out of Cloud and Enterprise too. And if developers weren’t deeply in Scott’s blood that would make a lot of sense. A more dramatic change would be to take Windows Server and move it into the Windows organization, though you then risk overloading Terry Myerson. Anyway I’m not making recommendations, just throwing out illustrative examples.
The second possibility for Satya would be to split Cloud and Enterprise in a far more dramatic way, leaving Scott running part of it and a new EVP reporting to Satya for the remainder. This would require a lot of effort as you tried to tease apart the current organization in a way that both strategically made sense and avoided creating non-functioning organizations. The Windows Server and System Management organization separates from the Azure organization relatively easily, but the Data Platform Division has feet in both camps.
The third possibility would be to have Scott run Cloud and Enterprise but move away from him having multiple triads reporting to him. That is, re-insert a management level below Scott with a (S/)CVP running Windows Server and System Management, one running Azure, and one running Data Platform. That reduces the number of directs Scott has to manage as well as letting him offload more responsibilities to area owners.
The fourth possibility is to maintain a Cloud and Enterprise organization with a different internal structure, perhaps along Cloud and Enterprise lines, with Scott owning the Cloud organization and someone else owning Enterprise. Both would report to an EVP who reported to Satya.
By giving Scott the interim designation Satya has set things up so that further near-term changes won’t seem like they were done because Scott failed. He was just a caretaker until they decided on a permanent structure. Of course it could take months to sort all this out, in which case Scott could prove himself ready for the Cloud and Enterprise EVP role and they’ll just drop the “interim” designation.
Most likely there are further changes afoot, but I have no idea if they are modest ones designed to help Scott succeed or more substantial ones based on not wanting to push him too far too fast. Or, they could be changes based on Satya having a different organizational viewpoint from Steve. Or ones that Scott himself initiates (e.g., who says Scott wants to pay attention to everything that he inherited with Cloud and Enterprise).
Scott is a good guy and he was ready for a bigger role. I want him to succeed, and hope Satya and Scott are working on making sure his role is defined in a way that he has a good shot at it!
Great analysis, Hal, as usual, thank you! My 2 cents: ScottGu is great figurehead for Cloud and Enterprise and they could not have found a person more popular with customers. Keeping him in charge and adding a management layer below (option 3) seems like a no brainer to me.