CS183C Session 12: Nirav Tolia, Nextdoor
For today’s class session, John Lilly interviewed our special guest, Nirav Tolia, the founder of Epinions and Nextdoor. You can watch the recording here.
Nirav went to the high school on Friday Night Lights — that movie is accurate.
Q: How did Nextdoor get started?
A: I’m running Nextdoor, a private social network for neighborhoods. It’s what we call a pivot; I was working on Fanbase, about two years in, and with some pretty big backers. We realized it wasn’t going to work. Our first impulse was to return money to our investors.
We were funded by Benchmark and Bill Gurley, who I’d worked with since 1999, so we had a high degree of trust. I felt like we’d let them down. You realize you’re putting 110% of your life into something, and it’s not working. At first, you just think it’s the toil of entrepreneurship. But then we realized that some of our core assumptions were wrong.
I had a breakfast meeting with Bill. He brought me a poem, “If…” by Rudyard Kipling. He basically challenged my manhood. We contracted the team from 12 people to 7, and worked on a variety of different ideas in the summer of 2010, and Nextdoor emerged from that.
We feel pretty lucky today that we’re used in 50% of neighborhoods, and have raised $200 million.
We did pitch Greylock for Fanbase, and they said no. That was probably a leading indicator.
Q: I first met you in 1998 or 1999 — you were running RoundZero. What was that like?
A: I didn’t have ambition to work in technology. My parents were doctors, and I always thought I’d be a doctor. The fact that I’ve spent my career running consumer Internet companies is because of Stanford.
I was lucky enough to be an early employee at Yahoo (one of the first 70). I had no training, I didn’t know what I was doing, but watching the success of the company made me think, maybe I can be an entrepreneur. Silicon Valley fuels that belief.
A colleague of mine and I decided to network and meet people who wanted to start companies. We started by having dinner with a couple of our friends, to talk about starting companies. Back then, all the things we have now — incubators, TechCrunch — didn’t exist. There was no infrastructure on how to start an Internet company because they’d just gotten started. After a couple of months of dinners, we had a loose community of entrepreneurs getting together. Within a couple of years, it became the once-a-month-event that entrepreneurs wanted to be at. We were having dinners with many hundreds of people. Even at scale, the format was for everyone to separate into small tables of five — we were trying to replicate that first dinner.
One of the people I met was Naval Ravikant, with whom I started Epinions.
After a couple of years at Yahoo and running Round Zero, I felt that it was time to start a company. It was 1999, and Yahoo stock was near an all-time high, and people thought I was crazy. We started Epinions based on Naval’s idea, which was a brilliant one. We looked at Amazon’s consumer-driven reviews, and thought there should be a platform where experts could write reviews and opinions, and consumers could
We had an article in the New York Times called “Instant Company” about just how fast Epinions was scaling. We were blitzscaling before the term existed. We grew to 120 people, and after the bust, we had to contract the company. It ultimately merged with a New York company called Dealtime, was rebranded as Shopping.com, went public, and was bought by eBay for $600 million. Epinions was founded in April 1999, and went public in 2004. Nextdoor was founded in 2010, and it’s five years later, and we are not planning on going public any time soon.
There are companies that are almost overnight successes. Instagram was almost instantly well-regarded, and almost instantly one of the top apps on the app store.
But sometimes, fast up can mean fast down. Things get popular really, really quickly, but then go down. Think about games on Facebook. It quickly grew into the dominant platform, and just as quickly went away.
Student Q: If we have another bust, what do founders need to think about in contracting their company?
A: I don’t know if there is going to be a bust, but I do think there will be a correction. We’ve made the mistake of confusing the coolness of a product with the ultimate business value of the idea. A lot of these venture funded things are actually cashflow businesses for a single person. They’re good businesses, but they’re not good venture-backed businesses. If you raise money and need to justify a billion-dollar valuation, you start to do unnatural things.
Using venture funding to acquire customers is the wrong thing to do unless there is a network effect.
What founders and CEOs need to think about is that there’s a right way to run a company, and that doesn’t change in boom or bust. A lot of the things you do in a bust are really good to do in a boom as well.
Q: Talk about Nextdoor.
A: While we have social networks that are good for connecting with friends or colleagues, there aren’t social networks that are good for connecting with our neighbors. When you’re at Stanford, Facebook is fine, but then you move. 28% of Americans can’t name one of their neighbors by name. That’s not what it was like when John and I were growing up in small-town Texas.
It was a pretty obvious premise, but completely unproven. Five years later, we have 80,000 neighborhoods, out of 165,000 neighborhoods in America. We also have about 90% of the neighborhoods in the 40 major cities.
Nextdoor was not an overnight success. It was a long, hard slog, but it’s now picking up steam.
The grand vision for Fanbase was to build bottom-up communities to compete with ESPN. That didn’t work. With Nextdoor, we didn’t start with that grand vision; we just wanted to prove out the concept.
In the last few years, Nextdoor has scaled exponentially. In the first year, only 176 neighborhoods adopted Nextdoor. We probably talked with every single one of the people who started a “neighborhood.” You needed to get at least 9 other neighbors to sign up and verify their address. If you didn’t have 10 neighbors, we’d call it a pilot, and it would expire after 14 days. We have about 80–90,000 expired pilots.
At that rate, how many years would it take to get to all 160,000 neighborhoods…that’s a lifetime. We weren’t focused on the rate, we were focused on quality.
The fork we faced was whether we wanted to be like Twitter, where you could just put in an address, and then tune in to any location. The alternative was to be like Facebook — you had to be a part of a university. I had to use my Stanford alumni email address to join stanford.facebook.com. We decided on the latter — each neighborhood would have distinct borders and a distinct name.
When you meet someone at a party, what do you talk about? You talk about who you know. That’s Facebook. Then you talk about what you do. That’s LinkedIn. But think back to when you were a freshman — you ask, “Where are you from?” We felt that notion of identity was very important.
This was hard for several reasons. There is no objective database of neighborhood boundaries. So we had to create a Wikipedia-style system for our members to create accurate neighborhood boundaries. Second, conversations are private to the neighborhood. People talk about reducing friction; we were increasing friction. Finally, we required 10 neighbors.
Student Q: How did you find the people to start neighborhoods?
A: We asked all our friends. We did things that didn’t scale. We asked people to send us photos of their driver’s licenses. We asked for the HOA roster. We went door to door. Since then, we have systematically taken these unscalable manual processes and used technology to make them automated.
In the early days, my co-founder Sarah would call people who were at 5 neighbors and encourage them to invite more people.
We were looking for the signal. Only then did we worry about scalability.
We do the same thing today. We sit people in product groups — a designer, 5–6 engineers, and a product manager— to simulate the early days of Nextdoor, and allow them to do things that aren’t scalable. We recently launched a service to allow you to find a babysitter. Our team walked around Palo Alto looking for babysitters. Only then did we test demand.
Engagement has always been strong; 41% of our members use Nextdoor at least 3 times per week. We didn’t focus on scaling the number of neighborhoods until we had people engaged with the product. It’s what Google calls the toothbrush test — can you create a product that people use every day.
Student Q: How did you know your 176 neighborhoods were representative?
A: We worked hard to bring in diverse set of neighborhoods — by income, ethnicity, age. We found that the use cases and conversations differed — finding parties versus organizing Bridge nights.
The first neighborhood was Menlo Park. Then Portola Valley. So we forced ourselves to go outside SF, so we went to Seattle. Then we wanted to go beyond tech. We ended up in Hamilton, NY, and we’d see postings, “Have you seen my cow?”
Sometimes failure can be a powerful motivational tool…we kept asking, “Is this working? Is this working?” That first summer, I told the company that if we got 100 neighborhoods, I’d bring in a dim sum cart. The engineers said, “We’ll never make it.” We did.
We’re working hard every single day. “The journey of a thousand miles begins with a single step.” We want to be used by every neighbor, in every neighborhood, every day.
JL: Andy Grove said, “Fight like you’re right, listen like you’re wrong.” You have to be relentless about looking for data that proves you wrong.
The first couple of years that you’re growing a tree, you don’t see anything. But when you go to Muir Woods, you see what happens with years of growth.
At Nextdoor, we’ve done cohort analysis that shows that our penetration of those neighborhoods increases over time. 2012, 2013, 2014, 2015 — they look the same.
A lot of businesses are treadmill businesses. Every morning, you get up, you run on the treadmill, and the next day you have to do it all over again. Great businesses have a network effect — each step you run adds to the total.
Student Q: How do you keep people from defecting to other platforms?
A: There really is no substitute; there’s no other way to connect with your neighbors. We don’t have a lot of churn. What we see is that people go in carefully, and then over time, as more of their neighbors join, they do more. We see that our largest neighborhoods are the most engaged.
When I was at Stanford, I’d read the Daily and the Chronicle. That’s where I’d get local information. Nobody reads newspapers any more — they need something like Nextdoor.
Q: How do you think about the product?
A: We’re actually not a product company. We’re a community company. There’s a ton of potential upside in improving the product. We want to get all our neighborhoods to be like our best neighborhoods.
Student Q: What are the bottlenecks to growth?
A: Our bottlenecks are also our opportunity. Because Americans don’t know their neighbors, there’s no way to invite them. But if they did know their neighbors, there would be no reason for Nextdoor to exist.
Facebook doesn’t have a monopoly on the social graph; we all carry our social graph on our phone. There’s no easy way to do that for neighborhoods.
We exacerbated the problem by deciding that all conversations would be private — no virality, no sharing.
We let people look at a map of their neighborhood, and pick the neighbors they want to invite. Then we send those neighbors a postcard. We send millions of postcards…to get people to sign up for an online service.
Q: What about nearby neighborhoods?
A: We allow you to communicate with the nearest 10,000 households. If you’re in Palo Alto, you also want to communicate with people in Menlo Park. This increased the viral vectors, and increased the utility. The best driver of virality is a great product.
26% of our messages are looking for a service provider. The first reply will come in 12–15 minutes, and in the first 24 hours, you’ll get an average of 5 replies.
Student Q: How are you thinking about revenue?
A: In the early days, I was leaning heavily on Reid for advice. He told me, you build user scale, then usage scale, then revenue scale. Growth, engagement, and monetization are hard and different. The only way to solve them is to go sequentially.
Both Twitter and Facebook were 4–6 years in before they began to monetize. 2016 is going to be a big year for us. I have an executive offsite tomorrow to talk about monetization.
Student Q: What can you learn from Craigslist?
A: Marketplaces are driven by liquidity, not the product experience. Craigslist doesn’t have a good mobile app or experience. You go to sell where the buyers are. You go to buy where the sellers are.
We have a slightly different take; Craigslist is anonymous; Nextdoor is tied to your address.
JL: Marketplaces are all about whether the transaction clears or not.
Student Q: What were your management objectives in 2012? 2013? 2014?
A: At any given time, you have five objectives: Growth, Engagement, Monetization, Infrastructure, People. You have to figure out what are the sizes of the slices of the pie at any point in time.
This can change quickly. You can go from spending almost nothing on Infrastructure to spending a LOT on Infrastructure.
We’ve put almost no resources into Monetization or Engagement. We’re about to dial up Monetization quite a bit.
The pie gets bigger — you raise money, hire people, and have more resources. But you still have to slice the pie. In 2016, Growth will be about internationalization. 30% of our engineers are devoted to Infrastructure. As you scale your organization, you need to spend more time thinking about people.
As you get larger, you have a firmer grasp on the near term, which gives you more opportunity to think about the medium- and long-term. I’ve just hired a great HR executive, and she asked me, “What’s the bigger picture? What about 2017? 2018?” If we don’t set that big goal and work backwards, we may not choose the right things to work on in 2016.
As the company gets larger, you want to get as many people involved as possible, but you need to manage it.
The OKR is the What. As the CEO, I articulate Why. The How is what we want everyone in the company to participate in.
When you’re small, you don’t want to waste time on this. But as you get further along, you need to dial this up.
Student Q: What did you learn from Bill Campbell?
A: Bill Campbell is the most incredible mentor/coach/advisor in Silicon Valley. When he was coaching me, he was also coaching Steve Jobs, Eric Schmidt, and Jeff Bezos. He’s like a business father to me.
He ran Intuit for a long time. They had a set of core values, and the first one was, “It’s all about the people.” That’s what he’s exceptional at. He’s done that at his own companies, and he’s helped many other companies do the same thing.
Operationally, Bill taught me, here’s how you run a management team meeting. Here’s how you do a quarterly operating review. Here’s how you manage the board. It was absolutely invaluable. It was like getting management training on the field. We still say things like, “What would coach do?”
JL: He sees the genius of you, and hangs his advice on who you are.
He has three incredible skills. He is an unbelievable leader. He told me, “Your title doesn’t make you a leader; people who follow you make you a leader. He has actually done it. He’s not a professor and consultant. Finally, think about the companies he’s been around — Apple, Google, Amazon.
Student Q: What is the conversion rate on the post cards?
A: It’s not as high as we’d like, but we’ve been able to increase it a lot in 2015, which allows us to send more.
We don’t give every neighborhood the same number of postcards. If it has 2,000 members, we don’t give them many. But if a neighborhood is trying to get to 10, we might give them 100 postcards.
We’ve now built a rewards system, where the number of postcards you get are driven by your past track record of getting people to accept those invitations.
Q: How do you figure out a reasonable CPA?
A: You have to base it on the revenue you can generate in some reasonable period of time. If a postcard delivers a $1 CPA, we have to be able to drive at least that amount of value.
Because, as a management team, we’ve had experiences where we’ve had to lay off people, we’re very careful about spending.
As a CEO, I’m okay with spot spending. But I’m very wary of getting locked into spending money over the long term without any terminal point. Hiring is like that. I like to take risks, but I like to bound those risks.
Office leases are terrible. The last three leases I have signed have been subleases. They cost a little more, but they aren’t long-term. I’m willing to pay more per square foot to not lock myself in. Landlords in SF are looking for a 12-year commitment. People tell them, I can always sublease. But what if they contract? And what if the whole market contracts? Leases are very dangerous for startups.
Student Q: Have you had any demographic surprises among your members?
A: Our team would come to me and say, “Our member is 35–55, and has kids.” I didn’t want them to bound themselves too much. Facebook’s initial age band was 17–21. Today, my 73-year-old mom uses Facebook. You overserve your initial market — people who own homes and have kids. Today, we’re all over the map. Senior citizens, everyone. We even have a neighborhood at Stanford.
Our core value proposition — being able to communicate with your neighborhoods in a secure way — knows no bounds.
Student Q: How do you think about monetization?
A: If you’re not going to ask your members to pay (a la Uber and Amazon), you’re monetizing through them (a la Google and Facebook). In Google, the ads are part of the product; when they remove the ads, trust goes down, because the ads are useful. I have never clicked on an ad on Facebook. I never go to Facebook and think, I hope I see an ad. Facebook has such incredible scale that they can just slip in a few ads for demand generation. Google is about demand fulfillment.
That’s how we’re thinking about Nextdoor. 26% of our use cases are around finding a service provider. That’s demand fulfillment. On the other hand, real estate agents are clamoring to get on Nextdoor to get people to buy or sell their home. That’s demand generation.
50% of our messages are transactional. They’re not posting vacation photos. That’s where we monetize. Google generates its revenues from 30% of its queries.
Student Q: What about the neighbor behavior?
A: If you’re mean to your neighbors, you may see them in real life the next day. Message boards are cesspools of meanness, but we don’t see that on Nextdoor because you might run into people.
Civility and constructive dialogue is the way these neighborhoods get better.
Every large platform has to deal with this. Facebook went through this with bullying. Every large platform has that responsibility.
