Evernote CEO Phil Libin on Turning Loyal Users Into Paying Customers
Chances are that if you’re an Evernote user, if given the chance to sit down with its founder and CEO Phil Libin, you’d want to take the opportunity to do a deep dive on the company and its product. That’s just how Evernote users are.
And that’s exactly what I did last month. We met while he was on a visit to New York, and ended up talking for more than 90 minutes about a wide range of things. Among them: Evernote’s intriguing move into selling physical products, Libin’s vision for building Evernote as a “Nike for the Mind” and who Evernote users are, demographically speaking. I published Part One of the conversation earlier this month. Below is the second and final part.
In this portion we talked a bit about what has been described as a “cult-like” feeling among Evernote users and where it springs from. Libin also talked about what he’s learned from a security breach that hit the company earlier this year, why he thinks “a good problem to have” is actually worse than a bad problem and future plans for a new data center and an initial public offering.
AllThingsD: This is the year during which being an Evernote user was compared by a business magazine to being in a cult. It was a tongue-in-cheek reference, but at the same time, as a user of the product, I understand the feeling. There’s this sense in using Evernote that you never quite master it. I know it can make me productive, and yet there’s no way I’m using it to its fullest potential. But I never feel like I get there. I guess I’m not the only one who feels this way?
Libin: No you’re not. And part of that feeling is on us in that we need to make the product better. But part of it is also the point. I had this interesting epiphany over the years, and it became even clearer as we got into Evernote Business. … The idea with Evernote is that there is this gap. What you want changes months to month, and then there’s what Evernote gives you. And what Evernote gives you is like 90 or 95 percent, but then you have to put in that last five percent to make it do what you want. A little bit of that friction is useful, but it makes the whole thing more flexible. You start a new project or you change jobs or you change devices, you can bring everything in Evernote along with you. Whereas if it was this perfect tailored fit for where you are in life, it would be really brittle. Also that little bit of work tends to make people more loyal. They’ve invested some effort and work into it, and so that feeling that it’s never quite perfect and that you’re always tinkering with it is sort of a desirable feeling. … That little bit of friction is probably costing us about half, if not two thirds, of our potential user base. We spend a lot of time thinking about the right balance of friction.
So about how many users do you have now? And how readily are you making money off the users you have?
We have about 80 million users globally, and that’s a combined number for free and paid. Our conversion rate increases linearly with cohort age, so the longer you use Evernote, the more likely you are to convert to premium. The idea is that we want you to use Evernote forever. Once you’re using it, we want you to keep using it, and it’s more important that you stay than you pay us. We want the engagement. The longer you use it, the higher the perceived value gets. And the higher the perceived value, the more willing you’re willing to pay. It’s up to us to make something that you want to pay for. The percentage of people who pay in the first month is like one half of one percent. But if they use it for a year, that goes up to seven percent. In the second year, it goes up to 11 percent. Our oldest cohort, the people who have been with us five years or so, it goes up to 25 percent.
And you make most of your money from premium users?
We have one business model, which is to sell direct. And then up until about 11 months ago, we had only one revenue stream, the freemium model where you convert from free to premium. Now we have Evernote Business, which companies pay for, and we’ve seen great growth in that in less than a year. And then we launched the Marketplace with physical products, and it has barely been two months and it’s already exceeded our expectations a lot.
So where are you in the IPO process? Will you consider going public in 2014? You’ve raised so much money that I’ve lost track, which is a good problem to have …
It’s funny you say that because in all seriousness, if I ever write a book about the whole Evernote experience, the title would be “A Good Problem to Have.” The reason is that good problems actually suck worse than bad problems. They are harder to deal with and more stressful. For example we once thought it would be a good problem to have that the company would be growing so fast that we would outgrow our data center. Since then we’ve had to deal with it. But to get to your original question, we’ve raisef about $250 million in five rounds and there will be more. I think we might start thinking about an IPO in two to three years. But no, we’re not ready to go public in 2014. I want Evernote to be a company that deserves to be a great public company. And we’re not at this point. Not yet.
So let’s talk about your other good problem to have. What is the state of your infrastructure? You said you grew out of one data center. What are you doing about that?
We are now running in one big data center in California. We have a backup data center that is geographically far away for disaster recovery, and we’re setting that up now. We have two data centers in China for the Chinese version of the service. We lease dedicated space and we build our own servers. This is probably the last time we will do that. Our next move in the U.S. will probably be to build our own data center. It will come down to the headcount issue. How many full time people are you going to pay to be there. Our next big move in the U.S. will be maybe in 18 months, and it will be our own physical building. We’re starting to figure it out now.
What have you learned since your security breach?
A ton. The short answer is that we had a plan that we were going to follow. So when it happened, there weren’t that many decisions to make. I’m glad that we made that plan because the decisions were very painful. The time that you want to be making these decisions is not when you’re reacting to the situation. We knew ahead of time that if there was any chance that passwords were compromised, we would do a full reset. We knew it would be expensive, because all Evernote usage stops right there until the user resets. We did it because the real potential for damage to the company was to our reputation. As it turned out there was no data that was accessed. They got encrypted passwords, which are salted and hashed. We’re still cooperating with an investigation. We think we know who it was that did it. It wasn’t state-sponsored, and it wasn’t targeted at any Evernote data. It was broadly aimed at trying to get people’s credentials that could be used to transfer money. It was a textbook identity-theft ring.
Did you lose any users?
We lost a lot of users. I would estimate in terms of forward progress, maybe three or four months. It took us about that long to get to where we were before. It also cost us revenue, because we had a lot of revenue-generating features that we had intended to roll out, and we had to put them on hold because all we did was deal with the security incident. It set back everything. Two-factor authentication was a feature I was going to announce at Le Web, so two-factor got moved up by nine months. But a lot of other stuff got pushed back. So it cost us a lot in terms of forward progress. But we’re a stronger company because of it. We have a strong in-house security team that we hired from Zynga. And we’ve been working with IPSec and Mandiant on the initial response and audit. So we’re spending a lot more money on security than we did before. We’ve put in new processes around security, things that as a startup you don’t do. We got more mature a lot faster than I think we otherwise would have.