CS183C Session 16: Reed Hastings
Today’s special guest was Netflix founder and CEO Reed Hastings, who sat down for a fireside chat with Reid Hoffman.
Q: How did the first Netflix culture deck come about? What led to its creation and publication?
A: About 8 years ago, we were getting tired of new employees joining, and going through this 100 slide deck, and having 1/3 of the employees being shocked at its content. The big driver was that we realized every candidate should get it, and it every candidate would get it, it made sense to make it public.
Putting things in writing makes it more debatable. To some degree, it’s a Bill of Rights for employees — things we aspire to live by. We probably started putting it in writing 5 years before that — it allows everyone in the company to participate and contribute to it.
From the time we were founded in 1997 to when we went public in 2002, all we were focused on was not going bankrupt. In 2000, raising money was as easy as taking a tin can and shaking it, and $50 million would show up. Never seen anything like that until last year! The year after, when you shook your can, someone would steal the can.
We had to lay off 1/3 of the company in 2001 (from 120 people to 80), and eked our way into profitability and survived. After 2002, we realized we were going to survive and thought, it would be awful not to want to work here.
We didn’t set out to build the most productive culture
What we valued — even more than success — was working with really talented people in a productive way. After our layoff, we thought things would grind to a halt, but we actually got more done. We didn’t have to dummy-proof things. So at first we said, let’s do a 1/3 layoff every year. Instead, we decided to continually focus on talent density.
Managers had to decide if they wanted to retain each employee — if they wouldn’t fight to keep the employee, we wanted to give that employee a generous severance package.
When we went public, we had 150 people. People were worried that now that we were public, everything would go to shit — put in a lot of process, stop taking risks. What we’ve done is to promote employee freedom. If you want to operate with very few rules, you need to set context. We added a chapter to our culture deck, Context, not Control.
There is context about the problem, but there’s also context about behavior, which is culture.
Q: Bold new innovation usually requires command-and-control; how do you do this with context?
A: The manager shares a lot of context, but a lot of people make decisions. We have a lot of decision owners. A lot of things happen that I don’t even know about. You guys are going to meet Marissa in a week or two. She went to Yahoo, which was a very brave thing to do. She said she’d review every resume. That’s a symbol of excellence, but it takes a crazy amount of time.
I don’t review resumes, I don’t block hires. We don’t even try to hire perfectly. Interviewing is 6–8 hours. You know much better after 3–6 months. If you want to try something, try it. If it doesn’t work out, we’ll give them a severance package and move on. We don’t look at it as a marriage. There’s no need to.
Q: There are other attempts to build strong cultures — Google is one of them. How do you understand the variances?
A: There are many different cultures. As long as you’re clear about the culture, you’ll get people who are attracted to that culture. Some things like integrity are universal, but other things like dissent can work differently.
Weak cultures are diffuse; people act differently, and don’t understand each other, and it becomes political. Even a Hollywood studio, which is full of politics, is internally consistent, and works for some people.
Culture is what you and your senior team want to be.
Q: Do you have a culture check in the hiring process?
A: It’s not a checklist. It’s easy for people to say, “I love the culture deck.” You ask, what did you disagree with and why, and if they don’t know anything, they’re not a first principle thinker. We’re looking for people who are curious and question everything around them.
Q: How do you define the Netflix mission and how is it embodied in your strategy?
A: I’ve always believed a strong mission statement is important, and we’ve never succeeded at it. We’re an emotional product; we cater to things that are important but not necessary. When you’re selling milk or penicillin, you’re selling something people need. We sell what people want.
Facebook talks about a more open and connected world, and you hear it over and over. We don’t have that. Once, we said, “We connect people with movies they love.” That was long, and awkward, and didn’t really stick.
Q: What would you tell your younger self to do differently?
A: We were growing nicely in 2003 and 2004, and Blockbuster, which was about 20 times our size, finally attacked us. We thought we were very clever, and came up with ways to counterattack. Remember, we all had the same DVDs. They were discounting heavily and were half our price, so we were losing share. We tried a couple of big things.
In 2005, we launched Netflix Friends, our own social network. We added ad sales — Yahoo-type banners above the choosing interface. We added used DVD sales to consumers — $4/DVD. We added Red Envelope entertainment to buy films from festivals and send out the DVDs. Each effort had 12–15 people, and employees loved them. We were doing something!
We went through battles in 2006 and 2007, and in the end we won. None of those four efforts made any contribution to our victory. One by one, we closed them down.
When attacked, we should have retreated to do our core better, rather than broadening the surface of attack. When people ask, “Are you going to get into news and sports,” we say, “Absolutely not!”
In hindsight, if we just went from 98% perfect to 99% perfect, we would have beaten Blockbuster sooner than we did.
Q: What about Quixter?
A: It’s 2008 and 2009, we add streaming. In 2010, we launch in Canada as a trial of international and streaming only. We expected X, and we got 10X the signups.
Q: Wasn’t getting ridding DVDs always part of the plan?
A: We held on to DVDs because they were a differentiator from Hulu, but we realized it was a dying business. The mistake we made was pricing. We priced DVDs at $8, and streaming at $8. So the bundle was $16/month, and our previous price was $10/month. If we had just grandfathered the pricing, it would have been fine.
People hated Quixter so much that we “killed” the product, and people forgave us.
Q: Was the letter to the public you wrote a personal thing or a PR strategy?
A: Everyone who was a Stanford and Oxford graduate liked it. It was a boneheaded move. It just stirred things up. It was just too intellectual.
Q: So was streaming part of your original plan?
A: In 1987, I took Tennenbaum’s networking class. One of the problems is calculating the bandwidth of a station wagon full of backup tapes driving across the country. It’s a high-bandwidth, high-latency network. When I saw DVDs, I thought, “I bet we can mail these.” We always saw DVDs by mail as a high-latency network, and that we would use technology to reduce the latency over time.
Q: Anything you got wrong in your plan?
A: When we first started raising money in 1997, we thought we’d be mostly streaming in 5 years. In 2002, we had no streaming. So we thought by 2007, it will be half our business. In 2007, we were still nowhere. So we made the same prediction. And this time we were wrong the other way — by 2012, streaming was 60% of our business.
Q: What lessons did you learn from financing your startups?
A: When some idea is shaking you so hard, you’re willing to go into poverty to make it a reality, that’s when you become an entrepreneur.
My first company was funded 20 people, $20K apiece. Friends and family. On that money, we were able to launch a product, get some initial traction, and got our first venture round from Merrill Pickard and Mayfield. That was very helpful. A lot of third parties will screw with you because they don’t care what you think, but they care what your board members think, and that is a subtle value-add.
With Netflix, I had done well, so I self-funded the first couple of rounds. In hindsight, I wish we had done more debt and less equity. Reid: The prediction from success fallacy!
Q: How do you think about those first hires?
A: It’s people you want to work with, but you have to be willing to change. I owe it to my team to do the right thing for the company. For my first company, I had never been a manager. I couldn’t possibly fire someone — it felt cruel and inhuman, like breaking up with someone. It took my three years to come to the view that the rest of the company was depending on me to do the right thing, and that it wasn’t about me being selfish, but me protecting the company. Once it was “I think it’s best for the company,” I could get there.
One part of compassion is a severance package. Our minimum is 4 months, and it goes up from there. If we realize that we’ve made a mistake after 2 weeks, it’s still 4 months. We’ve never had an employee lawsuit.
Managers are nice people; they don’t like doing mean things. The severance package is a bribe to the manager to do the right thing, because we make it easier to fire someone.
Q: When did you start this policy?
A: It was three months from the very beginning. It’s real money, but it’s worth it. It doesn’t really cost you any more, because if you don’t pay severance, you’ll still spend three months managing the person out.
Q: Is there anything you added in when you got to larger scales?
A: We believe in people judgement. Some companies try to use data and statistics. Use data to pick stocks, not your spouse. We also do a lot of blind references. I am stunned by how many people don’t bother to call.
Q: Talk about the recommendation algorithm contest.
A: It’s a little disappointing. The first one we did was epic, it was transformative for Netflix in the machine learning community, but we’ve not found another problem that was amenable to that approach.
Q: Did you end up getting a better algorithm, or better goodwill?
A: We got a better algorithm, but we got more value out of the goodwill.
Q: Talk about original content. It’s not a very Silicon Valley thing to do. But you pull it off! What led to the strategy? What were the lessons from it?
A: Necessity is the mother of invention. Every cable network started on other people’s content, got some scale, and then went into doing their own original content. The path for Netflix was well understood.
We tried it early in 2005 on a DVD basis. We had only 5 million subscribers. We bought this cool Maggie Gyllenhaal movie. It cost 10 times as much to buy the movie, and still bought the same 20,000 DVDs, so we realized we didn’t have the scale and got out.
In 2010, we realized that we were spending so much on TV shows that my partner Ted Sarandos said, now is the time. It was $100 million, 1/4th of our content budget, but he picked great shows. Now he has a team out of the studios, and the freedom we provide showrunners has been very appealing. Our culture supports creative people, which includes showrunners as well as engineers. You don’t need 15 hours in the office; you get your best thoughts when you’re recovering from a mountain bike ride.
Q: Why did you do TV instead of movies?
A: We just had our first big movie, “Beasts of No Nation.” The economics of TV are better. It costs much less per content hour. “Narcos” is an American company filming in Mexico with Brazilian talent, and it’s huge in Germany.
Q: Are there any strategies that you pioneered that we haven’t touched on yet?
A: Once you go public, you’ll have a lot of issues around stock volatility. If it goes down, people feel cheated. If you get to that phase, look at how Netflix does stock options. It’s basically price averaging.
Q: How do you define the role of the CEO?
A: That varies by stage of company. In the first couple of years, you do everything. Customers, investors, dishes. You have so many disadvantages as a little nothing company, you have to make up for it with talent and hard work.
As you get to 50–100 people, you have to evolve your management style to be more strategic.
When you get to real scale, most of what I do is vision about what’s important. We should be global, but I’m not picking markets. We should be spending 10% of revenues on marketing, but I’m not picking campaigns. Vision, focus, inspiration, culture. But you can’t do much of the work — if you try to, you’ll burn out and get everyone else upset.
At my first company, I was 33, and we had 50 people. I was still coding at night, and trying to be CEO by day. I wasn’t taking showers. Finally, someone said, “Take a freaking shower. And when you have bugs in your codes, it takes forever to fix things because you’re not around.” I held on too long.
I felt like investing in me was selfish. I thought, “I should be working.” I was invited to join YPO, but I thought, “I can’t take a day off.” I was too busy chopping wood to sharpen the axe. I should have spent more time with other entrepreneurs. I should have done yoga or meditation. I didn’t understand that by making myself better, I was helping the company, even if I was away from work.
Q: Have any things you’ve encountered caused you to question yourself?
A: I’m like Hank Paulson — you can never know anything for sure. Every 18 months, we do an exercise, “What would you do at Netflix if you were CEO.” It’s a way for me to gather input — we call it “farming for dissent.” Yet you also have to be executive very firmly.
Q: How do you compose the board of directors? What would you do differently now?
A: When you’re interviewing VCs, pick on the people not the valuation. The rapport, connection, and trust you have with the people in your board room is important. It’s like a marriage, and it’s very hard to get divorced. You’d be amazed how many CEOs are not honest with their board members. That’s poison. You should lay everything on the table and if they’re good board members, they won’t freak out.
When you’re smaller, your board members know a lot more than you because they can pattern match. The larger you get, the level of depth of the board member is not nearly as strong as the CEO and leadership team. The board is more of a safety net.
Q: How do you know if you’re a capable founder or not?
A: I didn’t know. I just cared about the product. I was a reluctant CEO. A couple of times, I screwed up badly enough that I asked the board if we should bring in a CEO. The whole first company, 1991–1997, I was miserable. As soon as I learned something, the company got bigger.
When I started Netflix, that was 15 people — in comparison to running a 700 person company, it was so easy! It’s like having a second kid — you’re much less freaked out. I’ve enjoyed Netflix tremendously, because I could keep up.
Q: What are the key characteristics that let you know when to scale?
A: There are scale economies in companies like Netflix. Those merit investing forward to get to scale. Call that normal business.
Then there are rare businesses like LinkedIn and Facebook where there are network effects. The power of getting to scale is so great, and they are such extreme winner-take-all markets, that it’s worth doing anything to scale. You do crazy things because you have to grow 300% to maximize the opportunity. You do things that are sloppy.
Q: When do you think it’s necessary to move fast because of competition? For example, how does this apply to globalization?
A: A company can get started in Germany, and we can knock them out in 12 months. If you’re in a network effect business, you get more of “first is forever.” eBay in Germany, or Yahoo Japan. How quickly did Amazon wipe out CDNow in 1999? It didn’t matter that CDNow had gotten started earlier.
Q: Were there any key strategies that came with scale? Corporate development?
A: Not in our market. In 17 years, we’ve never done an acquisition.
Q: What do you think makes Silicon Valley so unique?
A: There have been a lot of businesses like Amazon that have scale economics, that succeed outside Silicon Valley. But when it comes to network effect businesses, it’s Silicon Valley.
(Reid: And China in that market)
Q: Why is that?
A: There are historical factors, like Detroit for cars, Manhattan and London for financial services. It’s almost unimportant how it got started.
Silicon Valley is better for entrepreneurs now than 20 years ago, and better 20 years ago than 40 years ago.
But if you have a little idea, move to Austin. Here, you’re competing with all these giants for talent and resources.
Q: Silicon Valley people think that they’re great managers, but it’s really about the network effects and structure of their business.
A: I’m not sure Silicon Valley people think of themselves as managers. Maybe Tim Cook. There are many ways to be a great manager. Be yourself, don’t try to be someone else. Don’t read about Steve Jobs and wear black turtlenecks. If you look at successful companies, the variety of leaders is quite wide.
Student Q: How did you evolve the internal management to move from DVD to streaming?
A: Because the first streaming product was integrated into the DVD offering, it only got difficult when we had two different offerings. In 2010 or 2011, out of 30 VPs, we had 5 people who were still focused on DVDs. We had to kick them out of the executive team meetings because we needed to eat, sleep, and breath streaming.
Student Q: How have regulatory strategies been important?
A: We’ve had to fight for Net Neutrality. The cable companies wanted to charge us to reach their customers, and then not charge themselves.
Any company that gets large has to deal with antitrust.
Student Q: Why don’t you police users sharing their account with other users?
A: When a wife shares her password with her husband, it’s probably okay. If a dad shares with his kids, it’s probably okay. If the kid is at college, it’s kind of okay. The only thing that’s not okay is the dorm room. If the entire dorm shares the account, that’s not okay. But we have an effective limit on sharing — only 1–2 concurrent screens, and you have to pay more for more screens.
Student Q: You talk about Netflix as a sports team. How does that work?
A: The old metaphor was a family. The better metaphor is a professional sports team. We pay people well, we want to win, and if you have one bad game, we don’t fire you. But you’re fighting for your spot on the team every year.
Q: Comments on “The Alliance?”
A: You’re signing up to get something done, which is like a player’s contract. The next evolution of “The Alliance” is introducing contracts into Silicon Valley. It works in Hollywood.
Student Q: What do you differently at Netflix than Pure?
A: At Pure, we were engineers, and whenever something went wrong, we tried to put a process in place to prevent a recurrence. It’s a very semiconductor-yield approach. And when we had to shift from C++ to Java, our people were good at following the rules, but not at adapting to a new market. In the long term, flexibility is what’s chosen for. Humans are not well-adapted to any particular climate, but we’re flexible enough to dominate the planet.
Student Q: Best thing for investing in yourself?
A: Reading. A combination of Western CIV and management books. “Beyond Entrepreneurship” by Jim Collins is a great book. You could read it every few years.
Student Q: How do you evaluate content? Can you apply machine learning?
A: We don’t. What shows up when you open up Netflix is a ranking problem. We can apply that in marketing as well. But picking shows is like picking employees — it’s a human judgement task. Other people have tried to do script analysis and none of it has worked.
Student Q: Is the focus of Netflix one show for a big audience, or smaller shows for smaller audiences?
A: We’re agnostic. Our focus on intensity. We want you to love what you watch. We do have to match production cost to the size of the audience.
Student Q: How do you ensure that the shows are good when you don’t interfere with the creative direction?
A: If we like the team, we back them and let them run with it. Relative to other studios, we’ve won a reputation for creative freedom.
Student Q: How do you build your board and advisors?
A: The ecosystem is much better for that now than 20 years ago. Reid Hoffman or Sam Altman could give you a good answer.
