mmhmm is Phil Libin’s fifth startup and he learned a lot from the first four. One of those startups was Evernote, which he co-founded and led as CEO for 8 years. mmhmm started from a joke of an idea in May 2020 but quickly evolved into a real business with the launch of their closed beta in July, just two months later. Last week, exactly a year after their beta launch, they announced that they had raised $100m, led by SoftBank’s Vision Fund, with participation from the Silicon Valley force, Sequoia Capital.
mmhmm is a video communication company that helps people make live and recorded video presentations more interesting and engaging to watch. You can choose custom backgrounds, put the slides behind yourself like a news anchor, or use a bunch of other unique and innovative features.
In our conversation, Phil talks about:
My questions are in bold; this interview has been lightly edited for clarity.
Let’s start at the beginning. Can you give me a quick overview of how mmhmm started?
Phil Libin: We started out just as a joke, a year ago. May 27th was the one year anniversary of founding the company. A couple of months into the pandemic, we were just bored and noticed that our new video lives were kind of tedious and pretty ineffective. We were just goofing around trying to make something that would make people smile a little bit and liven up our video meetings. Everyone we showed it to internally wanted to use it and then before we knew it, it was a real product. It went quickly. But yeah, we started by just trying to make the best of a bad situation. Within a few months as it started to look like maybe this pandemic actually would end and the world would come out of it, the product transformed into something that was much more positive. Once we are free to move around the world again and do things in person again, what do we still want to be doing on video? What did we learn about what's actually better on video than in person? And how do we lean into those new superpowers?
Wow, so you founded it May 27th and then launched on Product Hunt on July 7th. You built the product quickly. Did you start with a private beta?
PL: It was 42 days between the founding and when we released our private beta on July 7th. This year, we are celebrating that with something called mmhmm summer, which is a 42 day online festival where we are releasing a bunch of new stuff, announcements, along with new videos for our community and other content. It’s sort of a 42 day sprint for us, then maybe I’ll relax for a couple of weeks and be ready to roll again in August.
Tell me more about the launch. How were you gauging the success of the launch?
We actually didn't want publicity in July because we wanted to do a limited closed invite-only beta so that we could iterate quickly on the product. I had come up with a four year plan for our investors and it was just one word per year. Year one was, “get started.” Put something out into the world. We didn't want fomo or marketing. A closed beta would allow us to invite a statistically significant number of new people every day so that we could test whether the previous day's work had made things better or worse. If we’d let in everyone all at once we won't be able to do that cohort analysis and see what was or wasn’t working. But we also didn’t want a lot of people waiting around and being disappointed, so we tried to keep it quiet. We had considered making it confidential to ensure people didn’t talk about it, but then we thought, “well, we're a communication app, so it's kind of a dick move to ask people to use it to communicate but not allow them to talk about it.”
We decided to do one story and I've been friends with Casey Newton for a while so I reached out to him and he wrote a piece for the Verge. We didn’t try to get on Product Hunt, it was hunted indepently. Trying to keep it quiet failed. We had a hundred thousand sign ups within a few weeks and we were only letting in a couple of hundred people at a time so the list got pretty long. There was definitely a real pent up demand for something to make you smile a little bit on video, which we're happy to do.
We ran the private beta from July until a few months ago when it came out on Mac. We just came out of beta on Windows and we're about to introduce iOS. So it’s been constant product development like we're all used to.
You mentioned you were doing cohort analysis to understand how each new set of users was using the product. Were you primarily looking at retention?
PL: I have a whole quasi religion about metrics and there’s a framework that I use called State Machine. This is how we ran Evernote and I've applied it to every other product or company that I've worked on or invested in since. There’s a lot of different things to optimize for, but in the long term what we care about most is the growth of high value users. In this context high value means we are providing high value to them, not vice versa. Who are the people for whom we are having a meaningful impact on? Whose lives or jobs are meaningfully better off because we exist as a product? We want to measure that and we want that number to be growing every month. But on the way to that, there's a bunch of other steps in the middle. We've been pretty intentional about how we think about how we measure value.
I love how you’ve reversed the term “high value customer.” When looking at your initial product, were there things that you thought would deliver value to the customers that later turned out not to matter as much?
PL: We've stayed surprisingly on track, but it's only been a year so maybe it's not that impressive. We’ve managed for the past year to do the things that we said we were going to do for the most part and stay on the path. Let's see if that keeps up for five years. Our design philosophy is this word that we stole from one of our angel investors, Biz Stone. Everything we build has to be fun and useful together. We call it ‘fuseful.’ They also used this in the early days of Twitter. Their test would be: is this feature ‘fuseful?’ Meaning, is it useful because it has to do something useful if we hope to change the world. But also, you need fun because the fun is what builds muscle memory. So our criteria is that everything has to be ‘fuseful’, fun and useful. It was pretty easy to be ‘fuseful’ in the early days because everyone was locked up, no one was having any fun. A little bit of levity went a long way and video meetings tended to kind of suck because there hadn't really been massive innovation in that for a long time. It wasn't that high of a bar to add useful things, like being able to present over your shoulder. We actually just announced a bunch of new things today as part of mmhmm summer. One cool mind bending feature is called teleport and it comes back to ‘fuseful.’
I’d like to go back to the four year plan in your investor deck. You said the first year was to get started. Can you tell me more about that, what you are focused on now, and the next three years?
PL: Yes, the first year, we were just focused on getting started. Here’s the four year plan from our investor deck:
Our first year just ended a month ago on May 27th. ‘Start’ means we wanted to put something out into the world, make a company, and raise some funding. Basically have something rather than nothing. We did pretty well at this last year. Now we are all focused on product, which means product market fit. How do we make sure that the thing that we put out there actually solves a real problem for some real people? We're continuously improving it to make sure it's actually good, meaning it actually solves some important problem and it makes some people better off because it exists. Next year, we’ll focus on getting it into the hands of people, the scale and reach. Finally in the magic sci-fi future of year four, now let's tighten up the screws and make sure that it is a self-sustaining, profitable business. We only have enough mental bandwidth to focus on one word per year and these are our words.
Right now we are focusing on making a good product and part of that is learning how to explain the product and how it works. There's this process of mutual attraction where people for whom we have the potential to solve important problems are naturally attracted to it and so on. We haven't really spent anything on marketing or sales, but we have put out videos which are a combination of me demoing new features and stories from the community on how they are using it. These videos help us understand what people find the most interesting about mmhmm. It helps us understand for whom we provide the most value. Is it teachers? Is it students or salespeople? Investors? Is it clowns? Literally, one of my favorite examples is this company of professional clowns that use mmhmm and love it. We just put up a video with them a few weeks ago. Anyways, the goal is in twelve months we have enough confidence and enough data that we have a high quality product to then focus on ‘reach.’ By then we had better know how we are actually going to scale. How we are going to attract customers and how much it is going to cost.
I like the focus on product market fit and taking the time to understand how to position your product before you try and scale and spend a lot of money on marketing or sales.
PL: Yeah, so when I left Evernote, I kind of semi-retired and I became a VC for a couple of years. A friend of mine and one of my former partners at General Catalyst is Hemant Taneja. A brilliant guy and very successful. Recently, he wrote a really good book, called Unscaled. The main idea is that in the past, in the pre-internet economy, companies had to get big and to scale before they could have an impact. Because in order to make an impact, you needed the economies of scale. You needed to be big enough to have efficient manufacturing, distribution and marketing. So it used to be the first question a company asked was, how do we get big? Now that's flipped. Hemant's main thesis is that now you first get to impact. First you make a real difference for somebody then you figure out how to scale that impact, as opposed to the other way around. That's a pretty profound shift in the world and it's really inspired me to think about our strategy this way.
So, one step at a time. First, we make a good product because we can. We don't need scale to make a good product, in fact, it's a distraction if you focus on scale prematurely. First make a good product. What do we mean by a good product? We mean that there is a group of people in the world who are materially better off because our product exists. How do we measure that? With the State Machine framework I mentioned earlier. Once we are pretty sure that we've made a good product for some group of people, hopefully a growing group of people, then we focus on how we get it to as many people as possible. After that, how do we get as much profit as possible? But again, that’s a distant sci-fi year four.
A lot of entrepreneurs struggle in that first year as they try to find product market fit and the buzz from the launch has worn off. How did you manage to stay motivated and focused after the excitement of the initial launch?
PL: I think the most important job of a CEO is to isolate the rest of the company from fluctuations of the hype cycle because the hype cycle will destroy a company. It'll shake it apart. In tech the hype cycles tend to be pretty intense. At mmhmm we are very much in the Venn diagram of two hype cycles. There’s a general hype cycle around video, which is going to be way up and down over the next few years. I have conviction that there's going to be far more video five years from now than there is today. 100% conviction, but in the interim, it's going to be up and down. There is also a hype cycle around early and mid stage startup investment. It’s super volatile, now more than ever, because of potential changes in the tax laws, interest rates, and inflation. So you've got these two very volatile areas, video and startup investment, and we are sitting right in the bull's eye of that. This means that my most important job is to isolate the team so that we don't float based on the ups and downs of the current. Make sure we have enough mass and momentum to go through it, meaning we don't change what we do based on the hype cycle. And that takes capital, which is why we have to raise some capital to do this. It also takes understanding of where you're trying to go and knowing where you're going is not based on the hype cycle. You have to have a long term conviction about that. You may be wrong. The conviction could turn out to be wrong, but you're not going to know that based on day to day fluctuations of excitement or month to month. So have a clear direction of where you are going and then make sure the ship has enough momentum so it doesn’t matter what the waves are doing, you’re still going relatively straight.
That’s my main job. Yes, I make a lot of demo videos. I do a lot of jazz hands. But that’s not actually my job. My job is to gather really brilliant people and then make sure that they've got a few years of not being buffeted by the hype cycle so they can actually do something great. That's how I try to advise people when they make the mistake of asking for my advice on stuff like this.
I really like what you said about making sure you don’t need to rely on the hype cycle to get to where you are going.
PL: Exactly. If you're driving your internal energy from the hype cycle, that is a very transient and fragile source to derive energy from. If what makes you get up in the morning is because you happen to be popular on Product Hunt or something that day or that week, that is a really shitty thing to motivate you because that is not going to last. That is a very short term sugar high and it just can't be where you derive your internal reservoir of energy for yourself or for the team. It has to be something else.
Let’s go back to something you said earlier. You mentioned that when you first launched the beta it was easy to make something that people found fun because everyone was locked down, but now as the lockdown has loosened and people are venturing back into the world, video call behavior has changed. How has this shift informed your product strategy and finding product market fit?
PL: We always knew that people would get off of video calls at some point. The early product functionality was very much focused on video calls, but this was never the interesting part. It’s the tip of the iceberg. The really interesting part is how do you manage a distributed live work culture and situation? How do you shift from everything being synchronous in real time to being asynchronous? How do you live in this kind of hybrid reality, which is very much what we're moving into. The exciting shift really isn't so much about people getting off video, it's about people transitioning from being in survival mode. People are going from, ‘how do I survive this pandemic to how do I thrive in this new world?’
I think the amount of innovation that you can come up with when you're in survival mode is pretty constrained. It's very sharp and there's a lot of innovation that comes from survival circumstances, but it's a very specific type. It isn't generally the big picture stuff that really transforms society. That stuff comes when you can breathe easier and you can say, ‘OK, I'm no longer trying to survive, what lessons did I just learn that will now help me thrive, rather than going back to the way things were before?’
There's been a really stupid debate recently with people asking if all the meetings that moved online during the pandemic are now going to go back to being in person. I don't care. The best thing you can do for meetings is not decide whether to have it online or in person. The best thing you do for meetings is to cancel them. Get rid of as many of them as possible. It’s not like we were sitting around in 2018 saying, ‘oh man, I love meetings in person...more in-person meetings please.’ Taking something that sucked in person and then being forced with almost no planning to do it on video made it worse, but the answer isn't to go back to the way it used to be. The answer is to get rid of the meetings or at least 80% of them. We are starting to do this at my company using some of this technology. We are thinking through how do you shift from synchronous to asynchronous life? How do you live in a fundamentally hybrid world? I show off a lot of our upcoming features that change how people can present, consume and interact in the first keynote on our website. It shows a bunch of these concepts that show where we think the product is headed.
We think of our product as giving you ‘communication superpower.’ We have one big commandment, which we call the ‘Distributed Commandment.’ It says: ‘do in person what's best in person. Do on video what's best on video.’ Those two things aren't a trade off. Video is not a poor substitute to being in person, it's a different kind of thing. It had to be a poor substitute for being in person when that was all we had during COVID-19. But now it isn't the trade off anymore. Now if you want to do something in person, do it in person. The stuff that you do on video you should do because it's fundamentally better, foundationally better. There is a lot of stuff in the realm of interaction, information exchange, and distributed team collaboration that we are leaning into on the product.
You mentioned that you apply the metrics framework you developed at Evernote to your new companies, what other learning did you take from building Evernote that you applied at mmhmm?
PL: First, I didn't build Evernote, a team of brilliant people built Evernote. I was just lucky enough to be a part of it. I think my LinkedIn profile says my job was, ‘assembling a brilliant team of people and making sure they had enough coffee to change the world.’ I really think that and now I don't even need to get the coffee because it's all distributed. But that was my main learning from Evernote. It’s not about me. It’s about the amazing group of people, which includes the people that are building the product, but also the beta testers and the customers giving us feedback. At best, I'm like an editor. I'm encouraging people to interact and pulling some ideas together. I'm doing some jazz hand demo videos, but fundamentally, it's the success of the team and the community.
The other thing that we intentionally did at Evernote was build for ourselves. I said I’m never going to do that again, but here I am doing it again. Evernote was my third startup, this is my fifth. My first two startups I took the advice to really imagine what the customer wants. My first startup was an early e-commerce company, so we had big retailers as customers and built a lot of what beame e-commerce on the Internet. My second startup was a security company, so I dealt with a lot of governments and big banks. For years I would wake up thinking, ‘what does the customer want?’ By the time we were ready to start Evernote we were exhausted of that approach. I don't care what the customer wants, I want to build something that I want. When you build something for yourself as the customer you can iterate so much faster. Every 20 minutes you know if the product is getting better or worse as long as we're honest with yourself. In the early days of Evernote we were building the thing that we wished existed in the world because we wanted to use it. That's a cheat code, like finding a cheat code in a video game and everything is just easier and faster and I think that was a big part of the success of Evernote. But it also has side effects that are unpleasant.
After Evernote, I said to myself, ‘there's a lot more people in the world that are much more worthy to have their problem solved than me.’ Yes, building things for myself is fun and fast, but it's not necessarily the best thing to do. At my fourth startup, All Turtles, which is a product studio, almost all of our work, isn't for us. We're doing things in healthcare, eldercare, and in preventing harassment and discrimination at work. They are products that we think other people need. It's harder to build something that isn't just for yourself, but it's worthwhile.
At mmhmm, since it started as a joke, we fell back into this mode of building for ourselves. It reminded me of Evernote and how fast this goes when you really are the target customer. We're enjoying it, but hopefully we now understand the pitfalls of this approach and we can be wiser.
What are those pitfalls?
PL: The obvious one is that if you are literally building the product for yourself then at some point you completely lose touch with new users. For example, no one who worked at Evernote was a newbie, we were all power users. At some point we were only building power user features and we forgot about the first launch experience and new user tutorials. That’s a very trivial example. The more important one is the typical Silicon Valley blinders...problems for 20 to 30 year old tech workers are a pretty constrained and parochial set. If you're trying to build products for everyone in the world, or at least a large segment of the world, then it’s important to make sure the people building it look like a broad representation of the world. The more you focus on building for yourself, the harder that becomes to do, unless you're very intentional about it.
I totally agree with that. Being a young white guy in Silicon Valley I appreciate that there is not diverse enough representation at Silicon Valley companies, and if I just focus on building for myself, I have the privilege that there will be a decent market for the solutions to my problems, so I could not focus on just that without thinking about its impact.
PL: And more specifically, people who look like me who have the money to invest in the product. You can go a long way just living off of the investment.
Right, it’s important to make sure you bring in diverse viewpoints when building your product so that you can appeal to a much broader audience and have a bigger impact.
PL: Yeah, exactly. We are very clear. We are trying to optimize for impact. We're not trying to optimize for money and we're not trying to optimize for a number of users, we're trying to optimize for impact. You can have a very deep impact on a small number of users and that may be better than having a very shallow impact on a billion users. Of course, it'd be better to have a very deep impact on a billion users.
Last question, you mentioned that you retired from being an operator and were a venture capitalist for a while. Why’d you jump back in and start a new company?
PL: I don't think I was very good as a VC. It didn't hold my interest and things that don't hold my interest are a good indication to me that I'm not that good at them. If I was really good at it I probably wouldn’t be bored by it. I didn't have the skills that I think you need to be a good VC, like Hemant or Roelof, who led our Seed and Round A from Sequoia. They have a very particular set of skills that they've acquired Liam Neeson style through a long life of investing. I don’t have those skills. I acquired a different set of random skills so I wanted to focus on the thing that I’m good at.
Where can people go to learn more about you and your business?
PL: The first keynote is a great starting place. I think that really explains what we’re trying to do.
I hope you enjoyed my interview with Phil and learned a few things. If you did, give it some love on Twitter with a RT or Like. This helps more people discover my newsletter 😊 -Tyler