Diane Greene, Part 2: Launching Google Cloud
By: Terah Lyons
When Diane Greene got the call to step up as the founding CEO of Google Cloud in 2015, she had already been a director on Google’s board for almost five years. She invested her career in effectively creating the market for mainstream virtualization with her pioneering work leading VMware starting in the late 90s. She then sold her latest company to Google–Bebop–which built a new software development technology for writing cloud applications.
Diane’s appointment to Google Cloud was seen as a strong signal that Google was making a commitment to position itself on the offensive, facing off against competitor cloud giants Amazon and Microsoft, which Google was significantly trailing behind. When Diane’s role was announced, Larry Page, Google’s co-founder and chief executive of its parent company, Alphabet, called her new post a “huge new responsibility at Google.” It was its own massive entrepreneurial endeavor, housed within one of the fastest-growing companies on the market.
It was also a sign that Google was interested in pivoting from its consumer heritage–where the company’s main focus had been rooted in delivering online services such as search and email to individual customers–towards the enterprise market. No easy task for a company as big as Google, standing on the edges of what was then a swiftly-expanding enterprise cloud market. As the first leader of Google Cloud, Diane was faced with making structural changes to the business, and cultural changes to an organization that had grown up with a significant consumer emphasis. She led their strategic investment in AI technologies and found a successful pathway to the ultimate customer of cloud services in any large enterprise: the CEO.
I sat down with Diane to draw on her career-spanning insight for the AI Founder’s Playbook. This conversation summary is part two of two, focused on her experience leading Google Cloud starting in 2015. The first installment, about her time as the co-founder and CEO of VMware, is here.
Terah Lyons: You were brought into Google to help them make their first foray into the Cloud–as the founding CEO of Google Cloud. This was its own massive entrepreneurial endeavor. Tell us about the environment at Google back then and what challenges and opportunities you faced when you joined. What conditions did you think were necessary for the organization to succeed from the outset?
Diane Greene: I was on the board of Google for about five years before starting that role. At the time, I was pushing them to commercialize their infrastructure because they had one of the world's biggest clouds, but it was for internal use. They had a big customer in Snapchat, and they had a very opinionated way of running applications called Google App Engine, which kind of did everything for the customer, which Snapchat liked. At the time, they were the biggest customer.
I was helping them recruit someone to set up a cloud and go to commercial enterprise. They kept asking me to do it. We almost hired someone that I thought was very good, and they backed out of it. At that point they bought my company [Bebop], and I came in. It was December 2015.
At the time I came in, they didn't have an enterprise division. They had engineering. Sales was off with the ads. Marketing was off with the marketing people. They were just missing whole groups. They had no enterprise production customers. Google Apps, as they called it then, was not part of it. Before I came in, I said, "Look, we have to bring it all together under one division. It's never going to work otherwise." Google is a totally amazing company. I mean, it was such a privilege to be there and be a part of it all. They served over a billion consumers.
TL: One of the biggest objectives of that job, it seemed, was to turn a very large organization with consumer muscle into an enterprise giant. Consumer and enterprise DNA are so different–how was the conversion process? How did you approach that challenge?
DG: Consumer is very different from enterprise. I came in maybe naively thinking, "Okay, we'll just do this." We brought everything together. We had to bring in a lot of new leaders. Then, we had to re-construct [the whole org]. It was like flying the plane while building out the wings because we had to move so fast. We almost lost Snapchat when I first got there just because we didn't have any kind of audit logs. We didn't have any kind of security. There were no user keys. We didn't have proper networking. We didn't have proper storage for how they did it in the enterprise.
We set to work on all that in engineering. Fortunately, Google came with a phenomenal engineering team, but we were just missing all the functionality you need to be able to sell.
Google Apps was consumer at first. We rebranded it. G Suite made it enterprise because our reasoning was anything you do for the enterprise helps the consumer because it was mostly about hardening [the infrastructure] and making it more secure, which are needs that consumers have too. The changes served both markets.
TL: Besides the technical gaps – which sound like they were significant – what were the other practical or cultural capabilities you needed to build?
DG: Google didn't spend time with customers the way you do in enterprise. At the time, spending time with customers was considered boring. Fortunately, I have always really enjoyed doing it. I loved it at VMware. I came into Google with that mindset. It didn't bother me because I knew how to do it well, I had fun doing it.
But then we bumped into business model issues. The first issue was the contracts you do with the consumer–at Google they were all about protecting Google, as they had to be, but you try giving those same contracts to JP Morgan or Morgan Stanley or Goldman Sachs, I mean, are you kidding me? The contract thing was horrendous, how all the financials worked. This was a subscription business that we were building from scratch. You build up your cohort over time and that's how you get the growing revenue. That was a completely different model for Google. Then, we were the first professional services group at the company, and fortunately, I was able to get the guy that built it at VMware to come in and do that with us.
The person who at the time was running engineering said, "I think I'm working almost harder than when I did when we were founded,”--and he was employee number eight at Google. We were just [running] flat out, but it was one of the most exciting things. VMware to me had been very exciting–but the scale of Google was just amazing. It was really a grand adventure.
This was a subscription business that we were building from scratch. You build up your cohort over time and that's how you get the growing revenue. That was a completely different model for Google.
TL: What other priorities did you set for the business while you were there? Was there anything you faced that was just totally unexpected?
DG: We didn't have an AI group–so I brought Fei-Fei Li in to run AI. Then, when she went back to Stanford, I recruited Andrew Moore as the Director of AI, and he's still there. We recruited a lot of really excellent senior talent. While I was there, we also worked to develop AI principles. I got very involved in that because Google – and I – cared a lot about how AI was used. We didn't have a set of ways to think about what was okay and what wasn't okay.
The thing I didn't know to do was invest more time educating everyone internally. I was like, We’ve got to move so fast. I was just moving so quickly with my team and the division that I needed to spend more time externally so that we could partner with the rest of Google better. The legal and finance people would understand the business better, when we wanted to do acquisitions, they would be more sensible to the rest of the company. I just didn't invest the time. I think we could have moved a lot faster had I done that, but it was complicated.
TL: Well, talk about the layers of responsibility and stakeholder management that it must have entailed to be successful. You're growing this entirely basically new business unit on top of an existing one inside of this behemoth of an organization–all in the face of tremendous competition, moving fast. A big challenge.
DG: In three years, we went from about a billion [dollar] run rate–just a handful of customers–to an eight billion run rate. And a lot of enterprise customers in production. We build out verticals. That was something we became really clear about. It was a very technical sales process. You needed people to be there at the outset, to help the customer adopt the technology and go through their digital transformation. We set up an office of the CTO. We had a bunch of people that were CTOs at other companies come into this one group. They could go in and talk at a high level with a customer about what it meant to move to the cloud and how to do it. Similar to many sales situations in an AI company, I'm sure.
TL: How did you structure your approach to customer acquisition during that period? Who were your customers when you went to sell?
DG: We created verticals because domains like health were different from manufacturing, different from media. That vertical strategy really paid off. Then, we had a lead for each of the verticals that could generate the leads and connections into those areas–that was an adventure. I really worked. I worked harder than I worked with VMware.
I worked harder than I worked with VMware.
The customer category was a really pleasant surprise: What I found with the cloud was when you went into a company, it was such a major strategic move for them. There was sort of an existential need to do it, that you got to really sit down and talk about the company's strategy with the CEO. We didn't get to do that at VMware. That was a lot of fun.
When I was at VMware, we’d sit down with the chief architect, and sometimes the CFO because of the money savings–but it wasn’t a discussion about the fundamental strategy of the company. I would often spend time with the CEO of all these huge Fortune 1000 companies. In fact, I brought Google Cloud to Davos as a sales tactic. It was a pretty effective strategy.
TL: Thanks, Diane. This was a real pleasure.