Work on Startups Early In Life, Get Technical If You’re Not. A Chat With Garry Tan—Founder, VC and Former YC Partner
Garry Tan has been a trusted advisor to hundreds of founders that have come through Y Combinator, including both authors of this piece. When Garry dedicated time to meet with YC founders, there was always a scramble to get into the queue.
He possesses a unique mix of design, technical and startup savvy that could have been acquired in only one way: by co-founding, building and selling Posterous, a major startup, himself. Garry has helped rear hundreds of seminal startups through his experience as a partner at Y Combinator.
Garry now pilots Initialized Capital, an early stage VC firm with $160 million under management.
For those who don’t know Garry’s background, we’ll give you the very short version here:
- Stanford graduate, computer systems engineering
- Microsoft for two years and was employee No. 10 at Palantir
- Co-founded Posterous, a nimble blogging platform, in 2008 (later acquired by Twitter)
- YC Partner for nearly five years, now running Initialized Capital full-time, with Alexis Ohanian, the co-founder of Reddit
Having experienced so much in such an intense manner through building his own venture, and helping build other startups, as well as Y Combinator itself, Garry was one of the first people we wanted to speak with when we put together the concept for FC Live discussions. And, as usual, Garry didn’t disappoint, as he dispensed fantastic tips and wisdom throughout the conversation.
Some of the core takeaways from our chat with Garry—if you have time to read nothing else, read these:
- Founders can often be surrounded by the ‘fog of war’ when working hard on a rapidly growing startup, and they often don’t realize why it is that the company is growing so quickly. It can be hard to discern which features or differentiating functionality has brought people in, so it can be hard to determine where founders’ and engineers’ time should be spent.
- At one point, Posterous was challenging Tumblr, and had been growing at 30% to 50% a month for 2.5 years, becoming a top 250 website, while Tumblr’s growth had stalled. It was during this time when Tumblr rebuilt its backend to be faster, and to be more available and pliant when the company scaled further. Posterous failed to do this, however, something that Tan feels held the company back, not getting its response times back under 200 milliseconds.
- Tan sees the next few years as ones that will continue to be good for tech M&A, due to the large balance sheets that currently exist in corporate America.
- The best hardware startups are often those that take rather standard, commodity-like components, and build around them with extraordinary software. This is something FundersClub spotlighted in a separate piece.
- As a VC with a large portfolio of early-stage startups, Tan is constantly meeting with his founders. Often, he catches founders while they’re cresting—just raised a Series B, they’re being acquired, experience incredibly growth—and nearly as often he catches founders in a trough—they’re running out of money, they’ve lost a co-founder, the business has stalled. This dynamic of riding an emotional roller coaster with people he cares about on a daily basis can make for intense days.
- If he could go back and talk to himself, as a young college graduate who was doing contract work, or even earlier than that, he would say: “Go work on a startup, a product, NOW!”
- For the classic conundrum of a non-technical founder with a great idea, Tan suggests in lieu of searching for months or years for the right technical partner, it’s often better to just dive right in, learn how to code, and build the first iteration of the product. He recommends using onemonthrails.com as a learning platform. Non-technical founders needn’t morph into CTOs, he says, but by becoming somewhat technical, they can better attract and evaluate the right technical talent, and they’ll be better equipped to work with them.
Below, readers will find an edited version of a live conversation between Garry Tan and FundersClub CEO, Alex Mittal. While this conversation is edited for clarity and, in some cases, brevity, it should be noted that this is live exchange with questions from an international audience arriving in real-time. Many of the questions come directly from viewers.
Watch the full conversation here.
Mittal’s words in bold:
What lessons did you learn from your Posterous adventures? In retrospect, would you have done anything differently?
Oh my God, so many things.
Let's see. We started Posterous not really knowing what it would be. Like every startup, there’s some reason at a particular time, a market and an idea collide. It’s like catching lightning in a bottle.
The hard part about it is once you catch it, you don't really know why it came together like that until many years later. That’s the case for many YC startups, Initialized Capital portfolio companies, and definitely for Posterous itself.
Posterous went live in 2008. Most people would ask, "Why would you start a blog platform in 2008? You probably should have done that back in 2002.” In retrospect, 2008 was actually kind of the perfect time to found Posterous because smart phones were brand new.
Not only were smart phones new – phone camera quality was improving, but it was still fairly difficult to transfer photos off a smartphone and get them online. You'd have to connect your phone to your desktop and then download the photos and then sync them to your computer and then log into your blogging software to post. It’s a pretty similar process to posting photos to Flickr.
Thankfully for us, smartphone and iPhone apps were pretty new. There were no really great apps yet. You could just click email and email worked—and it was easy to send, which we used for posts, of course. That was our reason why in 2008 something like that could happen. At the time we had no idea.
The other thing about that time was it was the Cambrian Period of social networks. Facebook was still relatively new; it certainly had less than a hundred million users. Then Twitter was brand new, and Blogspot came out.
To follow-up on that, you mentioned catching lightning in a bottle and oftentimes not being able to trace why exactly things were clicking.
Yes, things grow but you don't really know why.
From a founders’ point of view – and there are lots of founders in the audience – and from your learnings with Posterous, how should founders evaluate early ideas? Especially given that, in your experience, it’s not always apparent why things are clicking?
Well, we were all prolific bloggers, but we were all using different tools. I was using X-A-N-G-A, Xanga, the self-hosted blog. So, at first, we just made Posterous for ourselves.
Sounds like you kind of had a unique insight based on using Xanga.
Yes, exactly. And we'd seen posts by email. Posts by email was something that a lot of people had, but nobody made it the cornerstone of their blogging service. You would have to remember to send the text for your posts to “firstname.lastname@example.org,” and users would never remember that, so then they’d never that feature.
Awesome. So why did you guys decide to sell? Posterous was acquired and sold to Twitter. Was that a difficult decision, an easy decision, or something in between?
By that time I had become a designer in residence at Y Combinator. Again, part of the problem with catching lightning in a bottle – you don't know how you caught it until much later.
Sometimes when growth does stop – and we've seen this with consumer startups over and over again – you just don't know why it stopped. You’re asking yourself, ‘Do I keep making new features? Do I take a step back and just work on the engineering of it? Are my response times broken?’
Again, it took years of looking back on that experience to figure out what happened. Posterous grew 30 – 50% month-over-month for two and a half years. It became a top 250 site, but then the growth stopped.
It was really, really great. We were catching up to Tumblr at the time. Tumblr was basically flat-lining in user growth and we were growing throughout that period. But at some point, it flipped back the other way. Posterous stopped growing and Tumblr started growing.
If you look back on the timeline, Tumblr actually stopped growing because they basically went back and rewrote their back end to make it performant and scalable for a much larger user base, and they stopped shipping new features.
At Posterous, we thought that the reason we kept growing, even when Tumblr faltlines, was because we were constantly asking our customers, "What do you want? What do you want?" And then we just quickly gave it to them.
That was the exact moment when we should have stopped and said, "No more new features. Let's go back and lower our response times under 200 milliseconds to benefit our best users: people who write stuff." But it's not something we did.
It feels pretty clear to me that had we done it a little bit differently, we'd be in a different situation, but you're always fighting the last war at your startup.
At the end of the day if you're not growing you need to ensure that everybody who has put time and money into your startup gets taken care of—if you can do it. You have investors and employees to take care of, you have users who you care about and you need to find the best possible home for them. Selling was the right thing for Posterous at the time.
Okay – new question: There's been a lot of attention on some major acquisitions like LinkedIn, Cruise, Dollar Shave Clubm, and Jet.com. How do you view the mergers and acquisitions landscape for venture-backed startups during the next 12 months?
2014 and 2015 were a very unusual time for later stage funding and people probably got out over their skis, so to speak. They did deals for companies that had growth but they couldn't get back to unit economic profitable and now we're sitting with this list of a few hundred unicorns. A good deal of them are very fundamental, real businesses that will continue to grow, but some of them maybe are not. I think the later stage funding crowd is a bit scared, and wondering, ‘what's going to happen here?’
But it’s also kind of an unprecedented time with corporations out there having an insane amount of capital just sitting on their balance sheets. What can they do with it? They can do share buybacks, but that won’t actually generate new economic activity like putting money towards R&D for new products and services.
Do you think this dynamic, with companies sitting on a lot of cash, will actually result in more acquisitions?
Yes, I think we're going to see a lot more mergers and acquisitions than we traditionally would, and I’ll think we’ll see startup board members from the large growth stage firms being increasingly more open to earlier exits, because that's the right thing to do for many of them.
Another audience question: What are some red flags you look for when you first meet a startup founder?
For co-founders, it’s a bad sign when they talk over each other. As you know – and you clearly have a great co-founder relationship with Boris – if you put the two of you in a room you can kind of tell there's a good rapport and that you work together well.
Other founders, especially when they're new, they've never worked together before or maybe they've never started a company before, or they frankly just don't know each other well, will often talk over each other. It's like, "I'll take this question." They're physically motioning with their bodies to say, "I'm going to take that one."
Sometimes there's rolling of the eyes toward the other co-founder, which, in an investor meeting, is probably something you shouldn’t do. Being a co-founder in a startup is like getting married.
Yes, you spend more time with your co-founder than your spouse—at least awake time.
Yeah. I remember Alexis Ohanian coming in and speaking to our YC batch and saying exactly that. It's marriage without the sex.
Actually just for listeners out there we happened to have published a guide on co-founder fit at FundersClub.
I read it, it's good.
Garry's endorsing it so you should check it out. It's built on the lessons we’ve learned, as entrepreneurs, but also on seeing other startups go through the same experiences.
It's the number one thing that destroys companies at YC, and I think this applies broadly, as well. It's hard to find people you can work with.
Another audience question: I have a project I've been working on in my spare time, and I feel positively about the direction of it. How do I know how much money I should raise to get it off the ground?
Sometimes what you can do is work backwards. It depends on the product, it depends on the market, it depends on the team you need. Is there technical risk? Who do you need to hire? Do you have the co-founding team you need?
Given that, how much money do each of you need to actually get this going? Some people can save 20 or 30 thousand dollars and live on it. Some people can't. Some people need to hire two or three more people to actually even bring a reasonable version of a product out to the market.
I would work backwards from what you think you need to get to. If you're pre-seed, then ask yourself what would get you a $.5M or $1M seed round? If this is your seed round, what would get you to series A? Then kind of make a budget around accomplishing those goals to get your to the next step. So many things are dependent on one another, but working backwards is usually a good way to figure out how much capital you need.
Start with where you want to get to and figure out what you need to get there.
Right, and the order is usually:
- Get product to market
- build the team
- Get some sort of traction number to prove that your business model is working
As a quick follow up to that, our previous guest, Megan Quinn, commented that she doesn't think that VC money or angel investor money is the best thing for all companies. What are your thoughts on that?
That's totally true. There are really great businesses that don’t need venture funding and that, frankly, don’t meet the criteria for venture capital. Then there are companies that need hyper growth, and it's going to be a foot race, and five other companies are going after the same market. If there's hyper growth and capital is a big differentiator, then you should raise VC money. If not, and you have another way to finance your business, then you should do it the other way.
Another audience question: What are the most repetitive parts of your day to day? What sort of ‘in the weeds’ type of activities are part of being a partner at a VC firm?
Every day is so different. You're generally meeting new people a lot of the time.
But that's not ‘in the weeds’. Let's see. What is ‘in the weeds’?
Something I didn’t consider early on is that, as a VC, you’re often meeting with your portfolio company founders, who are at inflexion points in their life and in their companies. These are major life moments for them, and not always good ones.
We have 180 portfolio companies—that’s a lot of founders. Basically, you're going to meet with founders whom you've known for years, whom you really care about, and it's going to be either the best day or the worst day of their lives. You meet people whom you care about, whom you have worked with for a long time, who are in really extreme sorts of situations. Their companies are skyrocketing or they’re failing in many cases.
It's like the roller coaster of founding a company. You're either catching them at the top at the bottom.
Yes, and the hard part is, if you're an empathetic person, that really affects you. Some days are very hard.
Another audience question: How do you see the VC landscape changing and what are you doing to prepare for that, or how have you, perhaps, already prepared for that?
I think the craziest thing we're seeing, through YC, is that there so many industries that can be venture funded—things like biotech med devices. All sorts of products that have a software component. Cruise was a ‘hard tech’ company manufacturing self driving cars, but it also was deeply rooted in really, really great software. The key differentiator was really, really great software. (This exact topic the focus of a thought piece at FundersClub.)
Most VCs out there are very focused on software margins and it's like how do you make that model work, now that software really is eating the world? It's eating these low margin businesses and turning them into higher margin businesses. It's taking fragmented industries and turning them into winner take all markets. That's new. That's different.
Another audience question: What's your vision for the firm, and what makes you better than other firms? If a founder happens to have optionality, meaning they’re picking their investors from multiple willing partners, how do you think about distinguishing yourselves?
Here's where we're going to sound like everyone else. To be totally frank and honest, everyone says they’re founder friendly. Many VCs are operators now, and they say ‘we're just going to try to be your partner.’ Everyone says that.
This is the part that I struggle with. Look, I was a YC partner. A large part of that role iswe're helping founders avoid the ten thousand landmines that kill companies.
As Paul Graham says, “If you don't die, you will succeed.”
We genuinely love to do this. It's not about the money. It's not about power. It's not about the name or being famous or anything like that. It's just purely about living in this freaking crazy place where people from around the world come to try to make and build something that’s never existed before. That's what we're here to do.
I guess for me, I'm like chasing the dragon–companies like Instacart. You can meet really amazing folks who look at the world in a different way, hire thousands of employees, and actually build this thing that millions of people use.
This is another question from the audience: What advice might you offer an entrepreneur who needs a technical co-founder?
I'm assuming this means a business person who doesn't have the technical skills to launch the company they want to launch. They need somebody technical.
It's a common problem. Normally what I tell people, both my friends and people we meet everywhere who have a great idea, have a great market in mind, have domain expertise, is that the best co-founders are people you already know.
College roommates, people you knew from school, people you grew up with, people you’ve worked with in the past. That's the best place to start. Beyond that, it's always great to kind of be able to code it yourself.
The key thing is everyone wants to help a rocket ship. They want to latch onto the rocket ship that's blasting off, but nobody wants to be a part of a thing that's going to fail without them.
If you can get a first version out, and you hook up Stripe to collect payments, that's really, really magical. It will give you the advantage of being able to A: evaluate really good technical talent, B: really attract them, and then C: be able to work with them really, really, quite well. All of that's highly valuable.
I'm not saying be the CTO, but I am saying to gain some technical experience so that you can hire, manage, and attract the right people.
Another audience question: 180 companies in your portfolio? Wow, that's a lot of companies. How do you handle competitive conflicts when evaluating new opportunities?
I think you just have to have a bit of a Chinese Wall about things. You know what would be proprietary for a given company. Trust takes years and years to build up and it takes one text message to ruin. I think all VCs kind of have to understand the boundaries, seed investors especially.
It's a little bit of a grey area. Some firms absolutely do not fund competitive companies; others maintain the Chinese Wall. Sometimes, Alexis will talk to one company and I'll talk to the other one. And we just won’t discuss that space. There are ways to do it. It's always tricky.
Another question from the audience: If you had to start again from scratch with everything you now know, what would you do?
I wish I worked on startups way earlier. Straight out of college or even before that. I loved programming in college, so for I was thrilled to be making $100 or $200 per hour to do contract programming.
I’d like to go back in time and shake myself and say, "Hey, make a product.”
When you sell your own time, you only have 80 hours per week you have to sell. But if you make a product, that product can sell over and over again. That's zero marginal cost. That's so much more powerful. A product can touch so many more people. You can make something a lot more valuable if you make products. That's the one thing I wish I could change—I would have made products over and over again way earlier. I wouldn't have spent so much time consulting.
Thanks so much for joining us, Garry, it was great.