CS183C Session 15: Diane Greene, VMware
For today’s class session, guest lecturer Jerry Chen interviewed his long-time colleague Diane Greene, the founder of VMware.
Jerry: “The smartest people I ever met worked at VMware.”
Diane and her husband Mendel founded the company in 1998, along with three of his graduate students. Diane led the company to $2 billion in revenues and 7,000 employees.
Diane is on the boards of Google, Intuit, Khan Academy and MIT.
Diane: “Jerry made VMware fun.”
Q: VMware was doubling in headcount every year. How do you keep the culture strong and hire the right people?
A: The hardest, most difficult hiring was in the very beginning. Convincing people to come was a big deal. Each person is a big deal. Then all of a sudden, you’re hiring 100 people a month, and it’s easier.
You can never drop your standards, and that’s why it gets easier. When you create a strong culture, when people come in and it’s fun and it speaks to them, it’s easier to get them.
We were competing with Google. We started off as friends, and then we realized that we were competing with each other for talent, and stopped talking with each other. When I joined the Google board, Larry and Sergey said we need to get Diane on the board because that’s the only company we lose people to.
People who had a hobby, or a family, they would join us. If they wanted a 24/7 experience at work, they went to Google. We got a lot done, but we didn’t have nearly the long hours of Google.
Q: Talk about the cultural immune system at VMware.
A: If you get a critical mass of really good people, when you bring someone in, and they’re not keeping up, they self-elect to leave. You have to watch and help someone leave if they’re the wrong person, but that’s the exception. I remember hiring a VP of Engineering, and three weeks in we had an offsite. He didn’t say much. He came to my office and said, “I really couldn’t keep up.” He didn’t have the confidence, and he left.
Q: How do you balancing promoting from within and hiring from outside?
A: Whenever I went to hire the first of something, I went for the most talented person I could find. I didn’t necessarily go for someone who had the big brand. Jason Mar was this smart Stanford grad who had run on the track team. He came in as an individual contributor, and he scaled up professional services into 1,200 people, and ran it as a profit center. We hired this guy who had been a pilot in the Navy — always hire a military person if you can; they’re so disciplined, especially for Sales — he built out Sales Engineering, and he’s still there.
Q: I remember when VMware fit in a single lunch room. When you’re a public company, how do you maintain the communication at scale?
A: I was very shy and self-conscious, and got nervous in front of people. When you give a talk every week in front of a growing company, I got over that fear.
We had a staff meeting every Monday. By 9 PM on Sunday night, you need to get me a report on what’s going on in your group that others need to know about, or what you need from people. Then I’d write my own and send it to everybody on my staff, and they could share it with their staff.
Every staff meeting, I invited the founders. They were individual contributors in the Engineering organization, and people had questions, the people could ask them.
It turns out that my staff members started asking their staff to write up their reports by Friday night, and so on.
Student Question: Are you planning to do the communications differently with your new company?
A: My new company has a lot of different kinds of people. It has designers…I’d never worked with designers before. I have incredible respect for them, but they’re challenging for me. So you have designers who don’t like to write things up, but we’re using a similar system so that we’re all speaking the same language and doing the same things.
When you get bigger, I brought in a new CFO and a new Head of Sales, and I told them to go read the reports. Today, it’s “Go read Slack.”
Q: How do you know when to launch new products?
A: Our vision was to build the server product. We thought, nobody is going to deinstall their OS and run our hypervisor. Linux was coming out, and we thought, what if we offered a tool to run Windows? Mendel had the idea of using Windows device drivers instead of building our own. We got lucky — Linux got hot.
There were two companies that tried to go into the server market with a container product, but both failed.
We got the desktop business to $100 million; we thought we might have to give it away because…Linux users. Michael Dell invested in us, and he was quoted in CNN saying he was investing in a lot of Linux companies to level the playing field against Windows, even though he was going to lose money on VMware. You know how much he’s going to pay for VMware now?
We were cash-flow positive from Day 1, but we ran the company as break-even.
We went to bring out the server product ESX Server…we launched it and nothing happened. No one was buying it.
We also just took our desktop software and relabeled it as GSX Server, and everybody wanted that.
We created a preferred hardware program and went to HP, Compaq, Dell, and IBM…once they were on board, we relaunched ESX Server and this time it worked.
Q: For the first 11 years, VMware was the fastest-growing enterprise software company in history. Part of that was the channel. Talk about that.
A: We were basically channel-neutral — we had every channel. But our focus was really on the hardware vendors. I took the attitude that we’d be Switzerland, which worked well for us. I had an algorithm: What’s good for customers is good for the partners, and good for us. What’s good for the partners is good for us.
If you take us into an account, we promise not to tell anyone else about the account. If you develop something proprietary, we won’t share it with anyone else. We called it the rules of engagement, and we told the salespeople we’d fire them if they broke those rules. I only had to fire one salesperson.
A lot of our customers were running Microsoft OSes, and Microsoft wouldn’t support running Microsoft in our VMs. So the hardware partners were really important because they had support contracts with Microsoft, so we had them deliver the frontline support to people running Microsoft on our VMs on their hardware.
If a customer bought virtualization, they bought a ton of hardware.
Jerry: When you sold $1 of VMware, you also sold $8–10 of hardware and services.
Q: When EMC bought the company, you doubled down on that.
A: And it got me fired! But it was the right thing to do for the customers.
Q: Startups worry about hiring Sales — both too early and too late. When do you pour on the gas?
A: This was pretty new territory for us. We didn’t even know what salespeople looked like. (One of our first great salespeople was a woman who was also a standup comic.) We got a couple of different types of salespeople and tried one of each. And we had to figure out if they were bad people or if we had a bad market.
We started seeing traction in some places. I started getting intros to every VP of Sales I could. They all had different advice.
I and Tom Jurewicz and Jeff Byrne came up with this model for sales — which they still teach at Stanford Business School — we need someone on the phone to drum up business, we need a Sales Engineer to go in and figure out hardware and compatibility, and we need a direct sales person to close the deal. All three had the same number, and it worked like a charm. Another thing that worked like a charm was to give the same commission on channel sales so that they were compensation-neutral. The smart sales guys went out and educated the channel, and were making millions, and I was happy to pay them because it meant that VMware was making more millions.
I think it pays to be generous in Sales. But you can’t force Sales — if you’re not getting traction, hiring more salespeople doesn’t help.
Q: How did you think about compensation?
A: It’s different today. You can make more money at Google and Facebook today, or at these unicorns (some of which are actually real) than at a startup. So it costs a lot more. I started paying people a bunch of money because I thought it wasn’t fair otherwise.
Q: How did you keep the engineering teams innovative, even at scale?
A: We had exceptional people who pushed each other. You want to have people who are self-driven. Even the person sitting at the reception desk, that should be a driven person. Every person should like to set the bar high for themselves. People told me they really enjoyed VMware because they did more than they thought they were capable of.
In engineering, I told my VPs, give me your schedule, but always leave some slack in the schedule so that they have time to play around. We also had poster sessions — little conferences within VMware where people could present new technologies and if they were picked, go and build it.
When the Mac went x86, I went to Steve Jobs to partner, and he would have none of it. So I told everyone, “We can’t build this.” But I also let them know that they should use their own initiative, and I saw all these people with Macs.
Q: At VMware, we didn’t have a fear of failure, thanks to that slack in the schedule.
A: When things failed, we highlighted our mistakes, but we never highlighted someone screwing up.
Q: How did you think about diversity in hiring?
A: We had this attitude, hire everybody you can out of Stanford…and MIT and Brown and a few other schools. Back in 1999, we had an intern program to recruit juniors to come and spend the summer. If we liked them, we made them an offer, and got a lot of good people that way.
We did our best to hire different kinds of people. We did pretty well with women.
Student Q: How are you doing hiring for the new company?
A: We only have 36 people (including 5 former VMware senior engineers), so many personal network. Back at VMware, I had trouble hiring a VP Sales. This time, it’s design — I had to read about 10 books on design and hang out in a lot of Mission coffee shops. I finally find a great head of design who had been at Walmart Labs.
Someone sent me a designer resume, and I was looking at it on LinkedIn, and on the side I saw this photo and I though wow, she has so much energy. So I sent her to our recruiter, and we hired her. It turns out her husband is in advertising, and he took that photo!
We’re designing complex interfaces with lots of roles and lots of communications…to make that easy to use and rapidly iterate, we’ve had to build a lot of stuff.
There’s the product person, the user researcher, the designers who are product people, designers who can build the UI, Front end people who are full stack, Front end people who are only front end, Designers who do front end…you have to bring all these people together and it’s not trivial. My husband is teaching a course about it this Winter.
Student Q: Did you aim to be a core enterprise technology platform?
A: It was not all premeditated. We did believe that this should run on every piece of hardware. We were sure this was a better way to run systems, and we thought of it as a platform from Day 1. It needed to be a platform to maximize its value to people. But the beauty of VMware was that it was useful on a single server. So some sysadmin could use it on their own, and then it would spread to different pockets. We didn’t want to go to the CEO or CIO right away. They have a purchasing person, and they’re compensated on how big a discount they can get. We wanted to avoid that person. We’d wait until VMware was all over the organization, and then we’d offer them an enterprise license agreement.
Student Q: What are your views on containerization and the “Docker Revolution”?
A: Jerry is the world expert on Docker, so I’ll let him cover that. We couldn’t do a container at that time. Nobody believed in virtualization. We had to build a P2V (physical to virtual) tool. We had another tool that could measure the utilization of your servers and point out good candidates for virtualization.
Now that everyone appreciates the value of virtualization, you can do things like Docker.
Jerry: Both Diane and I are very bullish on the container ecosystem. Docker would not exist if not for VMware.
Student Q: VMware liked to stay under the radar and underpromise and overdeliver; other companies take a very different approach of staying over the radar. How do you think about that?
A: There are a lot of advantages to underpromise and overdeliver — you set your pace, you build trust in your brand and product (you don’t try to deliver until you’re ready). We needed to be a trustworthy company. It was a lot less stressful, and a nicer way to operate. If our engineers found a problem, we could fix it rather than looking for shortcuts.
If you make a lot of noise, you can be super-successful as well. Someone once told me, “Diane, if you took you and Marc Benioff and put yourselves together, you’d really have something!”
Q: You made VMworld an industry event, rather than a VMware event.
A: I really believed our stuff belonged on every machine, and to be valuable, you had to be very open. VMworld was about virtualization, not VMware. We had the virtualization solution — nobody could compete — so it was a pretty safe thing to do.
Over and over again, we did the generous thing, and it always paid off — whether with our own employees or with our partners.
Q: A lot of hypergrowth companies try to buy other companies to reach critical mass; VMware never did that. When does M&A make sense?
A: VMware was a little unique; we were creating a new industry, so there weren’t any companies to buy. Later on, when we did buy companies, I don’t think we did a good job of it.
If you’re a startup, and you’ve never done an acquisition, you’re probably going to mess up. 95% of the time, they don’t work out.
It’s a big distraction, because they’re very time-consuming, and the smaller you are, the bigger the distraction.
Q: When did you know it was time to upgrade the VP Sales?
A: It was always something different. Our first sales guy — who ended up staying at the company — came to me when we were doing $10 million in revenue and told me, “I don’t think we can get to $20 million.” I knew it was time to get a new VP Sales. The next one, I caught lying. I had a VP Marketing who I walked out the door; Intel sent me an email he had sent them where he sent all our confidential information trying to get them to buy the company. It’s amazing the kinds of things that go on.
If someone is not being treated like a leader, if people aren’t listening, it’s a good sign you don’t have a leader.
Our first VP Engineering, people were frustrated that he wouldn’t make a decision. It’s true, you never say, “I wish I had waited longer to fire him!”
Unless they sent all your IP to another company, it’s not personal. I’ve remained friends with a lot of people I’ve fired.
Student Q: What have you consciously done differently with the new company?
A: It’s such a different climate now. It’s so hard to hire, and the traffic between here and San Francisco is so intense, we made the decision to get two offices from Day 1: Los Altos, and San Francisco. Two days a week we’re in Los Altos, two days a week we’re in San Francisco. Now that we know each other well, we do one day each.
Student Q: Now that you’re on the board at Google, what do you wish you had a VMware? And vice versa — what did you do at VMware that you wish they did at Google?
A: I think Google was bolder. They knew not to sell, and I wish VMware had that…though I didn’t want to sell. Incredible quality, incredible people.
Student Q: Best piece of advice you’ve ever received? Worst?
A: Best: You can never overcommunicate. Worst: I was at a startup that did streaming video. I had a disagreement with the founders and I left. One of the investors told me, “You shouldn’t leave. This is the biggest deal you’ll ever be involved with.”
Student Q: What steps did you take before you founded VMware?
A: VMware was my third startup. My husband Mendel had invented this idea, and he and his grad students were building the technology. The first thing that happened was that they put out a paper, and Bill G and Nathan M wanted to read the paper. So I told Mendel, you should get a patent on this. As it went along, I said, we should take this to the market, it could be really neat. About that time, I found out I was pregnant with our second kid. I thought, I’ll help get this off the ground, and then we’ll bring in a CEO.
Because one of my previous startups had been a little dysfunctional, I had us write the mission and vision, and had everyone sign it. I had my baby, and we were in this office where you could open the windows, and I’d bring my baby in because she could have fresh air.
Q: What’s hard about scaling?
A: The thing I remember as the hardest was space. Having enough space for everybody, anticipating it. It was during the bubble, and rents went up to $10/sf/month and they were asking for 10-year leases. That’s what put Donna Dubinsky out of business; that’s why she had to sell Palm.
Jerry: We went through a lot of office defrag. I opened the box at my desk, and it was my officemate from two moves ago, and he had already left the company.
Q: What advice do you think would be helpful to the students?
A: You can be an entrepreneur anywhere. You don’t have to start a company to be an entrepreneur. Everybody wants to start a company, which is a problem, because then there’s nobody to hire!
I was an entrepreneur before I started a company. We didn’t have great companies back then; they wouldn’t do new things. They wouldn’t jump off the elephant; they’d ride it right over the cliff.
If you have an idea and want to build something, figure out the best environment in which to do it — it might be a startup, but it might be a company. I did a lot of things I really enjoyed at big companies.
