Would you refuse to launch with venture capital?

I first discovered today’s guest because he was doing something that just seemed crazy to me.

At the time his company did online backup, so it seemed natural that he would use one of the cloud storage systems that are available, like Amazon.

Instead of doing that, he created his own freaking hardware.

I remember thinking, “Isn’t this going to fail? What happens if he loses his data?”

Today it feels like all those questions that I had are gone. He’s proven.

Gleb Budman is the Co-Founder of Backblaze which offers the lowest cost cloud storage on the planet.

Gleb Budman

Gleb Budman


Gleb Budman is the Co-Founder of Backblaze which offers the lowest cost cloud storage on the planet.


Full Interview Transcript

Andrew: Hey, everyone. My name is Andrew Warner. I’m the founder of Mixergy. By now you know Mixergy is the place where I interview entrepreneurs about how they built their businesses and I’ve got an audience of real business people, real entrepreneurs who listen to these interviews and often they code up their companies while they’re listening to these interviews. Other times they commute or run.

But the goal is to give you stories that will just be embedded in your brain, make you smarter so that when you’re ready to need the ideas from these entrepreneurs, they’re just there in your head and you can access them as naturally as you can your home address without having to write it down. You’ll just know it.

Joining me today is Gleb Budman. He is the cofounder of Backblaze. I first discovered this guy because just something that seemed a little crazy to me. Backblaze at the time did online backup. It seemed like a natural that if you’re doing online backup, they use one of the cloud storage systems that are available, like Amazon. This guy, instead of doing that, was creating his own freaking hardware. The hardware exists. He decided he’s going to create his own freaking hardware.

I’ve got to be a geek on some degree because I would just eat up his freaking blog posts about how he’s creating his own hardware. I would read comments on Hacker News about this. I would look at screenshots like I was looking at porn to see how was this guy creating hardware to store data, people’s backups in the cloud, his own person cloud he created. Isn’t this going to fail? What happens if he loses his data?

Anyway, that’s the stuff I was reading about. I was so fascinated about the company. Today it feels like all those questions that I had are gone. The company is doing really well. It’s all proven. They’ve proven their model. I invited him here to talk about how he built up his business, why he decided to go that direction, about whether he was secretly a little freaked out when he was putting everyone’s data in their own customized hardware backup system.

Anyway, Gleb Budman is the cofounder of Backblaze. They offer the lowest-cost cloud storage on the planet. That means that they’ve expanded beyond backup. This interview is sponsored by Toptal. If you need a developer, I really urge you to go check out Toptal.com. But later on I’ll tell you about them. First I’ve got to welcome Gleb.

Gleb, what did you think of the way I did my intro?

Gleb: I thought it was great. I love that you followed us for as long as you have and know some of the backstory. A lot of times people start off and they just see where we stand today. So, having you have that background is awesome.

Andrew: When you were creating these old pieces of hardware, did you think, “What am I doing?” at any point or did it just click for you mentally and so you went with it?

Gleb: No. It was definitely an evolutionary process. Just to be totally clear, I am one of five people that started the company together. So, I take only–

Andrew: Believe me. I’m going to bring up photos that one of your cofounders put up of your parties. I know about your cofounders. I’ve seen what you guys are up to.

Gleb: Okay.

Andrew: But did you secretly–I want to get into how you built up the business, but before I do, I just had to ask you–did you secretly think, “What am I doing? I’m going to lose people’s data. I’m going off half-cocked in this whole other direction that everybody seems to be saying is wrong.” Did you ever panic like that?

Gleb: Yeah. There’s the analogy of being a duck, where everything seems calm above ground and you’re just running as fast as you can underneath the water. I felt like that at various points in time. We had launched and we were getting customers in and on the website, it was like, “Hey, come and sign up,” and inside the company, we were like, “Are we going to have enough space? Are we going to have enough cache? Are we going to have enough drives? Is this going to stay up?”

There were various points in time early on where we were iterating through on the design and we didn’t feel like it was solid. So, we had something where we called the octopus pod because we drilled all these holes in the bottom of our storage server and had all these dangling cords coming out the bottom so we could be analyzing the storage server while it was accepting data. So, early, early on, I was definitely pretty nervous about it.

Andrew: All right. And this whole crazy thing started because a friend of yours lost their data and your response to your friend was what?

Gleb: So, a friend of ours last her data. Our response was just, “Great. We’ll get you setup. We’ll get you a new laptop and get you reinstalled. Where is your backup?” It seemed like, “Everybody backs up their data.” She starts pounding the table and saying, “Look, what I don’t need now is a lecture. What I need is for you to get me my data back.” It was like, “You don’t have a backup? We can’t help you.”

So, it kind of started from there. We started asking all of our friends and family what they did for backup. The answer was usually, “Umm… Well… I’ve been meaning to do something. I’m going to get around to it this weekend.”

Andrew: I get that. There were all kinds of solutions for doing it. Why didn’t they use any of the solutions that existed?

Gleb: So, when we started, our CTO was always religious about backups. So, the way he did backups was he burned his data every week onto a CD and then he would burn a copy onto a DVD and he would check a file on each. He would stick both together, put them in an envelope and mail them to his brother in a different state. Then he would do the same thing and put a copy in his own closet. He would do that every single week. He was just religious about backups. But obviously this was like–no one in the world is ever going to do it.

When we talked to a lot of people about why they don’t do backups, the number one response we got, “I meant to do it. I was trying something and it didn’t quite work. It was too hard and I gave up.” So, we just kept hearing that over and over again. It was just like people would just get stuck. Price came up, but nearly as frequently as the difficulty and complexity. When we dug into that–with technology, a lot of times the answer is, “It was too hard,” and you really have to peel the onion back to understand what that means.

When we dug into it, the part that we heard for most people was the hardest was every single solitary solution had always said, “Hey, what do you want to backup and then we’ll take care of that for you.” To a tech geek, that makes sense. But to normal people, it makes them go, “You’re asking me a question I’m not exactly sure how to answer. I don’t know where everything is.”

Andrew: They don’t even know where in the file structure their own personal files are.

Gleb: No. It’s like good luck trying to find your outlook .pst file.

Andrew: You know, the other thing you did that I thought was interesting, you did focus groups. What startup guy does focus groups? I thought that’s we laugh at big companies for doing. Why did you decide to do focus groups?

Gleb: So, in ’99, I was starting another company with some of these same folks. We were trying to figure out–we were on the raggedy weird edge of what we were going to do. We were trying to figure out whether anybody would use the stuff that we were building. At that point, I put out some ads to get some people to come into the office and eat pizza and talk to use about how they interacted with some of the things we were doing at the time.

It was immensely helpful, in part because one of the things I found is it doesn’t work great when the product manager or the CEO goes out and talks to customers and then comes back and says, down from the mount, “Here’s what I have heard.” It works so much better when the engineers building the stuff can actually hear it directly from the mouth of the people you’re trying to get to use it.

So, by bringing people into the office to have the focus groups, it allowed the engineers to actually here what people are saying, being unfiltered from how we would present it.

Andrew: They would build and then regular people would come in over pizza, check out the software and give feedback directly to the engineers?

Gleb: They would. Some of it was even earlier. The way our design process at Backblaze works is we usually sketch UIs on paper first, literally by hand and then we have one of our cofounders is a graphic designer and he will create mockups of those. We put them up on a wall and then we walk through them that way before any code gets written. That way you can kind of take a step back and see what the user experience is going to be as opposed to what is the code that is required to build it.

So, the first thing that we would do is we would sketch up some of these things bring people in and just talk to them. The first focus group was actually just saying, “What do you do to protect your data? Do you do anything?”

And just even learning what people’s processes were. We heard, “Well, I back up stuff to an external hard drive,” “When I go on vacation, I copy my documents folder. I guess the last time I did that was a year and a half ago.” That was a common one. We had people say, “I just take my most important files and mail them to myself in Gmail.” We had someone who was just like, “I know I should, but really my backup method is praying.”

Andrew: Well, what about this, though, Gleb–I interviewed David Friend. I know that he had a company before you doing online backups. There are other people who had online backup companies. Why couldn’t you just say, “There’s this online backup company that exists”? You can pick any number of them. Go and use them.” Why create your own one when you see this problem?

Gleb: So, it’s probably partially naiveté, partially egotism, partially some belief that maybe we could do it better.

Andrew: What could you do better when you were looking at the problem?

Gleb: When we first started, what’s funny is we didn’t come at it from the perspective of, “Let’s build an online backup service.” That was not the original plan. What happened was we were all still employed. We were all at a company when our friend lost data. When we had that discussion and we started talking to friends and family about it and realized nobody is backing up data.

When we realized that was a problem, five of us decided, “We’re going to quit our jobs and commit to each other for a year without salary and we would start this company to try to solve backup.” It wasn’t a, “We’re going to start an online backup service.” So, when we first started, we were sitting in this one-bedroom apartment with one of the founders, going, “Maybe what we should do is write better software to back up to external hard drives.”

This was before Time Machine existed. There really wasn’t good software. So, most people would just manually drag and drop files. We thought, “Well, maybe what well do is we’ll create a service that will take all your files and send them to all the other services out there. It will put your spreadsheets into Google Docs. It will put your photos on Flickr. It will do all this stuff.” So, we thought that might be an option.

One of the options we contemplated for a fairly long time was peer to peer backup. So, I would backup to me and you would backup to me and that would be the software we’d make. So, we didn’t start at it and say, “We’re going to do an online backup service.” We said, “Here’s the problem. There are many possible solutions. Which one do we want to go down?”

After digging in a little bit on each one of them, we decided that online backup was the best path. One we decided that, we started looking around. Certainly we found Carbonite out there. We found Mozy out there. We looked at both of those and we’re like, “These are good services. They already are out there. Maybe people should just do them.”

One of the things that we found with both of those services is that this question, this very key question about what do you want to backup, we realized that everybody was getting hung up on this one place. Carbonite went that same direction. They said, “Hey, what do you want to backup and we’ll back that up for you?” Mozy went the same way and every other company did as well.

So, what we wanted to do was flip it upside down and say, “We are going to back up everything on your computer. We’re going to backup all of your external hard drives. We’re not going to ask you so that you don’t have to worry about it. And then what we’ll do is we’ll exclude your OS and we’ll exclude your apps. But that way you don’t have to think about where everything is.” So, that was kind of like the key early thing where we felt we had this aha moment.

Andrew: But Gleb, couldn’t that be a feature that the others copied?

Gleb: So, we didn’t think it was that easy for two reasons. One is that just architecturally, it reverses the model. So, it changes the UI. It changes the flow through the system. It changes the overall model that you go with.

Andrew: What do you mean by the model that you go with?

Gleb: So, things like, for example, Mozy went with a 2 gigabyte free approach. People love free. Free is great. But fundamentally, if you’re backing up everything and excluding some things and you’re not giving people a choice of what gets backed up, you can’t do a 2 gig free model. It doesn’t work because it’s a different experience.

Andrew: I see.

Gleb: So, as we started looking at it, we said, “There are a lot of pieces this is going to touch.” Part of it was that. Another part was we realized we’re probably going to be storing more data than these guys. That one early on, we weren’t sure whether that was going to be a good thing or a bad thing. But we figured it was going to be a challenge.

Andrew: What about this other thing that you mentioned earlier, peer to peer backups? It seems like we all have tons of space on our computers. I know I do. My wife does on hers. Shouldn’t I just backup my stuff to her computer and her to me? You don’t have to pay for an online service. That never took off. Why doesn’t that work?

Gleb: It is a technologically beautiful, elegant solution. It’s something that our CTO started coding it up because he was like, “This is awesome.” And the technology could be made so good. And the problem was that as we looked into it, it’s just too complicated. No matter how easy you make the technology, as it is, backup is not something that people go, “Oh my god, I’m so excited. I’m going to do backup today. This is going to be fun.”

So, as it is, people are like, “Okay, I guess I’m going to do backup.” So, now what you’ve got to do with peer to peer is I’ve got to decide backup is important. Then I’ve got to find a solution and I have to understand what peer to peer even means.

Then I have to find you and I have to say, “Hey, Andrew, I found this thing called Backblaze. You should use it too and you should do backup and we should do it at the same time. Now let me explain to you how it’s going to work.” Both of us have to do it. Both of us have to have space. Both of us have to trust each other and trust the system. Both of us have to have our computers on at the same time so the data goes up and down.

Andrew: And in a world of laptops, especially Apple laptops, they’re the big issues, as soon as you shut your lid, that’s it. My friend can’t backup. I can’t backup on her computer. What I’m curious about is the thought process that allowed to eliminate all these different options. How did you know, for example, that peer to peer would have all these problems? Was it the focus groups? What did you do?

Gleb: One of the guiding lights that we have always had was ease of use. Even before this company, this team, we’ve worked together for about 15 years. Ease of use is something we just sort of share as DNA. So, whenever we look at it, whenever we’re digging into it, we say, “Is this going to be easy or hard?” So, the focus groups help. But the focus groups were these limited things. We did it a few times with a handful of people. It wasn’t this grandiose thing.

But what we did all the time was go out and just talk to people. Literally I remember I was at a friend’s wedding in Tahoe and I was on a hike with a group of people. I was just like, “What do you do for backup?” And having this conversation with these people. I didn’t say, “We’re starting a company.” I didn’t say, “Here’s exactly the product we’re going to do.” I just started exploring with them, “What do you do today when you take photos and you take movies and you’ve got stuff on your computer, what do you do?”

The peer to peer side of things, I just started asking people, “Do you do anything peer to peer today? Would you consider a peer to peer backup?” The looks I was getting was kind of like that deer in a headlights like, “Uh… I think I know someone who does. I don’t know.” So, it was just having those conversations over and over again, you’re like, “This is never going to fly.”

Andrew: All right. That makes sense. Let me do a quick sponsorship message and then some back and talk about the set of photos I saw on Brian, your cofounder’s website. You probably know what I’m talking about or you have a sense of it.

The sponsorship message is for a company called Toptal. These guys are just really changing the whole recruiting process. The old recruiting process is you go and hire a recruiter and the recruiter gets a big cut of what you’re paying–not a cut, but they get a big payment based on what you pay the person and they’re going to spend some time looking for a new hire for you and they do all the recruiting for you and it’s going to take a little bit of time, but eventually you might find someone and if you don’t, well, too bad.

Or you go online and you start buying a bunch of ads and someone internally–actually, what’s recruiting like for you, Gleb? You guys hire really solid people. You’ve heard me talk about Toptal in the past. You said, “This might be something for us. What’s your recruiting process like?

Gleb: So, we have [inaudible 00:19:13] very, very slowly. So, a lot of our recruiting, frankly, has been through each other and through our networks. One of the things that’s actually helped us a lot which we didn’t really expect is our blog. We are very open on our blog. We publish a lot on it. So, we get a lot of people that come in and say, “I read your blog. I’m interested in a position.” We actually hired one person because we did an AMA on Reddit. We were just in the comments area and this guy was chatting with us and said, “I want to work at this company.”

So, for the most part it’s been that kind of thing. As we’ve been scaling up and needing to hire more people a little bit quicker, we’ve worked with a couple of recruiters here and there, but it’s not any singular process. On the Toptal thing, since I’ve listened to your interviews, I was listening to one of them, I don’t know which one, where you were talking about Toptal. It’s not necessarily the perfect thing for Backblaze because at the end of the day we want people who are employees of the company and it sounded like Toptal was more an agency where they worked for them but maybe you can clarify that.

But I have a friend that I was literally having coffee with last week. He’s not a tech guy. He’s not a developer. He’s a real estate guy. But he’s got this app he wants to build. He’s like, “How do I find somebody to do this?” I’m like, “I don’t know even know where do I get you started on this path?” When I was listening to your interview, I was like, “These Toptal might be a path I would send him to.”

Andrew: Yeah. They definitely are for someone like that. They just have to go to Toptal.com/Mixergy and sign up. To your point, I think anyone who has the kind of network that you do and the blog and everything else is going to have a much easier time hiring. If the person listening to me does not have that, they’re going to really stuck finding the right person who’s a good cultural fit, who’s a good fit, who knows the technology they’re working on. I think for them it’s even better to go to Toptal than it would be for you.

For you the advantage is that it speeds things up. You get to talk to Toptal. You get to tell them what you’re looking for. They put someone in front of you. Chances are really good that they’re going to be the right person first or second shot because they’ve got them in their network and if it’s not, you don’t have to work with them.

The other thing you brought up is can you hire them? Well, for speed, Toptal will hire them themselves. Actually, they’re working for Toptal. But if you want to hire them yourself and just bring them on the team and say, “They’re long-term. We don’t ever want to talk to Toptal again.” You can work that out with Toptal also. Many people do.

The speed is why people go to Toptal, that your friend or you can say, “We have this project. We need to get started right now,” or, “We have this need for an extra developer. We need them to start on Monday.” You go to Toptal. They will get that developer for you on Monday. They have this incredible personal process of making sure that they find the absolute best person for you.

In fact, they said, “Andrew, we want to make sure that this is so personal that if someone wants an introduction, make an introduction on their behalf.” So, I’m going to do that right now. Anyone who’s listening to me who wants an introduction to Toptal should just email me, Andrew@Mixergy.com and I’ll introduce you to my guy over at Toptal. You can have a conversation with them, see if they’re the right fit for you, think through your process, think through what you’re working on.

If it’s a good fit, you can get started with that person quickly. If it’s not, you’ll have talked to someone really smart about your hiring process and about your project–Toptal.com/Mixergy. If you want to see the incredible deal they have for Mixergy listeners including 80 free developer hours and a guarantee that you like them or you won’t pay or frankly just email me, Andrew@Mixergy.com. And I’m glad that they’re back as sponsors.

Coming back to your story, you’ve got the idea now. You’re starting to talk to people. The last thing you want to know before you get started–actually two things–is price and how you create this for them. Let’s talk a little bit about price. How’d you come up with the price you had?

Gleb: So, it was more of just talking to lots of people. What we found was–again, I go back to this hike that I was one because I remember it really struck me on that hike. I talked to a bunch of people over the course of a few months. I asked them–again, didn’t tell them we were starting a company and didn’t tell them anything–just described, “Here’s what we’re thinking. Here’s what would be out there.”

And I was like, “What would you pay for something like that?” We got some people that said, “Well, it’s the internet. It should be free. Everything on the internet is free.” I’m like, “Great,” and internally I’m thinking, “Oh shit, this is going to be terrible. They’re going to want this for free and we can’t do it. We’re going to go bankrupt.”

On the other side, there would be people like, “Oh my god, that would be amazing. I’ve been living in fear of all my most valuable photos and everything just vaporizing. If there were a service out there that would do that, it’s worth $50-$100 a month for me easy.” So, there would be these two extremes. But then there would be this huge bell curve around $5 a month, where people were just like, “I don’t know. If it were like $5 a month, it would be a no-brainer.” It was huge. It made it easy from that perspective.

The funny thing is that most of these people, they had no idea other backup services existed. It wasn’t the kind of thing where they were doing some sort of competitive matrix in their head. We would have people literally that would say things like, “That is a brilliant idea, an online backup service? Why has no one ever thought of this? Oh my god. That’s awesome.” We’re like, “We just came up with that.”

Andrew: I kind of feel like there’s someone listening to this interview still saying the same thing. “Oh yeah, I don’t have to use Dropbox. I don’t have to use Google in order to back my stuff up.” Then the second part is the part that I brought up at the top of the interview, which is you guys decided to not use Amazon. Why not just go with Amazon? Netflix is using Amazon.

Gleb: Yeah. It seemed–it probably still seems insane. So, it was not our intention. The team that founded was a software team. So, our plan was use Amazon. That was absolutely what we were going to do. We were going to build this beautiful, great, easy to use experience. But the backend storage was all going to be on Amazon. There was no discussion about doing it any other way.

Then what happened was when we came up with the product concept, it was going to be unlimited backup. It was going to be $5 a month. We were going to sort of all of your data plus your external drives. We started thinking, “Where we would put that?” When we were looking at Amazon, Amazon was charging $0.15 per gig per month.

So, $5, we said, “Well, that’s 30 gigabytes.” We’re like, “I think on average people are going to store more data than that. So, we’re going to lose money on every single solitary customer even before accounting for development or marketing or sales or support or anything.” It was just a non-starter. We could not get started with it.

We talked about that if we had raised venture funding, what we probably would have done is said, “Screw it. We’re going to just start with Amazon and hopefully figure out how to make it cheaper later. But for now we’re just going to go.” Since we bootstrapped, that wasn’t an option. We didn’t have these deep pockets that we could just go, “We’ll figure it out.” We had to make the unit economics work on a per customer basis. So, we had to throw Amazon out. It just wasn’t an option.

Andrew: And then you decided, “We’re going to build our own thing.” And the first one was like the first Apple computer, which was in a box of wood.

Gleb: So, there was one other step before that. We actually didn’t decide to build our own at that point. We’re still on this path like, “No. We’re not going to build our own storage. That’s crazy.” So, once we decided no, then we decided, “We’ll buy equipment from EMC, NetApp, Dell, somebody.”

There are storage companies out there. We’ll buy equipment from them. When we looked at all those systems, they were all $1,000 a terabyte and drives were $100 a terabyte. So, at that point, we were like, “Okay, none of these are options.” So, what we want at the end of the day is we want to take a hard drive and plug it into the internet and somehow plug a hard drive into the internet and do it as cheaply as possible. That got us down this path. We tried a bunch of different things. We tried attaching a bunch of USB hard drives. It’s supposed to work. The spec says 256 devices.

Andrew: To one computer?

Gleb: To one computer.

Andrew: I see.

Gleb: It does not work. Nobody does that. So, none of the manufacturers that build all these USB hubs support that, so it doesn’t actually work. So, we tried a variety of things. We ended up evolving this concept of these NAS drives that seemed to work. We plugged nine NAS boxes into one computer. We’re like, “This seems to work.”

And we’d put five drives in each NAS box and that seemed to work. And then we found the manufacturer that makes the card inside the of the NAS drive, put all those into this wooden box you’re describing and built this wooden storage server which we actually put into the data center to see if it worked.

Andrew: And it worked?

Gleb: And it worked.

Andrew: Eventually. What about backups? The idea with Amazon, if stuff goes bad, they backed it up. And if they didn’t, it’s on them.

Gleb: So, they don’t do backups per se. All of it is too much data to do “back up.” So, what you do is you do some sort of repetitive sweep using either RAID or some other type of what’s called a parody system. So, that’s what Amazon does. That’s what we did. We did RAID inside of a box. So, even though it was one box initially, we could lose several hard drives and there was redundancy for the drives.

Andrew: How does that work? To be honest, I don’t know. I’ve heard the word RAID forever with regard to hard drive and I don’t know how that works.

Gleb: So, RAID used to be–what it stood for, redundant array of inexpensive disks. So, that was the whole idea. What you wanted to do was take a bunch of disks and use math to break a file up into pieces in a way that you could assemble them from some set of those pieces. So, the most expensive way to do that is to just double it.

So, you have one copy and you put all the same data on a second copy. But instead what you can do is you can do some math where you break it up into three drives where you don’t have all the data on all three drives. It’s more than one third of the data on each of the drives. So, it’s a little bit more expensive than having it on one drive, but you can reassemble it from any two of the three.

Andrew: I see.

Gleb: Depending on the algorithm you use, you can change the number of parody. So, what we had for example was that out of nine drives, you could lose three at the exact same time and reassemble from the other six, things of that sort.

Andrew: I see. Then it’s time to launch. We’ll talk about getting customers in a moment. But first, I’m looking at this photo on your cofounder’s website. At first, I thought, “Aha, a Palo Alto launch party from 2007. I can’t wait to see how crazy things got because this was like the height of Web 2.0. Michael Arrington reigned supreme. What’s going on?” Do you remember what was going on that day?

Gleb: We were hanging out in the living room and–

Andrew: I see Casey making a sign for Backblaze on what looks like a tiny kitchen table, with, for some reason, two rolls of tin foil there. I scroll down and I see the party table is like a paper plate with pretzels on it and a couple of other things.

Gleb: We were huge!

Andrew: Yeah. There’s a roll of paper towels in case you need to wipe your mouth after one of the pretzels. It’s just like people hanging out. A lot of them are just drinking water. This is not a wild and crazy party. If I look carefully, I can see Mail Frontier’s sign somewhere on this place.

Gleb: Yeah.

Andrew: Why is Mail Frontier in your office?

Gleb: So, first of all, this “office,” this was a one-bedroom apartment that our CTO lived in. When we decided to start Backblaze, two of us lived in San Francisco, two of lived in San Jose and he lived in this place in Palo Alto and was willing to let us started the company there. The original plan was there were going to be five of us that started the company. We were only going to be there for six months and then we get out. It turned out six months became a year and five people became seven and seven became nine.

Andrew: You just stayed at this guy’s apartment?

Gleb: We stayed at his apartment until at some point he looked up and he’s like, “You guys are obviously not leaving. So, I am.” He moved out. We took over his bedroom, which was great, because we needed more space.

Andrew: I kept seeing more photos. The 2009 Backblaze Macintosh launch party, also over there, the Backblaze from 2008 was there. Backblaze VIP party from 2009 was there. I think the VIPs got pizza. That’s what it is. This time there’s a tray. So, it seems like this was still lean times for you guys, right?

Gleb: It was very intense. Just to answer your question on Mail Frontier, Mail Frontier was the last company that we did together. So, the group that started Backblaze, we all worked together during Mail Frontier, which was an anti-spam company. So, we did that one from 2002 to 2006 before it got acquired.

Andrew: And you sold it?

Gleb: We sold it to SonicWALL. So, SonicWALL is where we were worked when we quit to start Backblaze.

Andrew: Is it inappropriate for me to ask you if you became a millionaire from that sell?

Gleb: I did not. The Mail Frontier exit was mediocre. It was a $31 million sale. So, the company was sold for multiple millions, but there was quite a bit of venture funding that went into it. In part, it’s why we decided to bootstrap Backblaze to get started. We kind of saw firsthand what happens through the whole venture process, that even though the company sold for $30 million and even though we had thousands of enterprise customers, very few people made any real amount of money off it.

Andrew: I see. You said, “Let’s do this brand new way.”

Gleb: We said, “There are positives and negatives to venture and there are positives and negatives to bootstrap. Let’s positive the bootstrap side of it.” When we started, it was very lean. We quit our jobs and committed to a year without salary between each other. The idea was that if all of us quit our jobs and ten a week in, two of us go, “You know what? This whole bootstrapping thing sucks. I’m going to get a job,” then the other three are left holding the bag.

So, by committing to each other for a year without salary, we figured we would have a chance of getting it off the ground and it would effectively be the angel round simply by us starting. We put a little bit of money in. It wasn’t a lot to get the company started. But the main thing was the deferred salaries.

So, the thing was it ended up being a year and a half before we took a single penny in salary and that was minimum wage, which was like, “Woohoo, we’re making minimum wage.” It was a very exciting time. It was lean times. We raised no funding. We were paying for mortgages and rent. My wife was in law school. I had a mortgage to pay. It was lean times. So, pizza was a splurge.

Andrew: So you really needed customers. You needed them to pay. You finally launched. It looks like the launch happened in 2008. You said that getting into TechCrunch was important to you. I see the TechCrunch article here, June 2nd, 2008, Jason Kincaid, who I really like.

Partially I really like him because I was at South by Southwest and he walked over to me and said, “Andrew, I’m a huge fan of your work. Let me tell you what I like about it.” Then I started asking if I should be more like Michael Arrington. He said, “Andrew, just be more like yourself. You’re good.” But you guys get so much attention. He said, “You do good work.”

Here’s the weird thing. You needed people to buy and you still offered 500 free accounts to TechCrunch. Why did you decide to offer so much for free?

Gleb: So, the accounts weren’t free. We offered them 500 invites into the beta. So, what we did was we had a private beta. We charged during beta, which was some amount of internal gnashing of teeth about whether we should charge during beta. But the idea was that on June 2nd, if you showed up to the website, you could not sign up for the service. You could learn about it, but you couldn’t actually download the application to start backing up.

So, what we told was TechCrunch was that we didn’t have room for anybody and everybody to show up and just start signing up. So, we said, “We will you have 500 invites. The first 500 people that come through your site, we’ll actually let into the beta to start.”

Andrew: I see, and allow them to buy. So, it wasn’t free.

Gleb: We gave them a two-week free trial. But after that, they’d have to pay. It was still–we had to shut it off at some point.

Andrew: So, that’s the other issue that I wonder how you dealt with. You basically have inventory management here. If you don’t have enough customers using your hard drives, you’ve paid for hard drives without making the money back on it. If you have too many customers, then you can’t back their stuff up. How do you account for that math?

Gleb: So, it was one of the hardest problems early on. For the first six months or so, we were in partial private beta period, which was from June through September, we were in a private beta where we could manage how many people were allowed in. Then we opened up the Windows product but the Mac was still in private beta. It was still in private beta for a number of months. So, we were able to manage a certain amount of demand by who we let into the private beta and we didn’t for a little while. But after that, once the service is open, it was charging.

Andrew: And even then, you’re basically giving up on customers as a way of making sure that you have enough hard drives.

Gleb: Yeah. And it was deeply painful, believe you me.

Andrew: Yeah.

Gleb: So, there are certain things we did to smooth demand. So, some of those things, for example, so, Amit Gupta from Photojojo signed up for the service early on. He went an email to customer support. I’d never heard of Photojojo at the time, but I was doing support. There were only five of us. So, we were doing round robin tech support. I happen to be doing support that day.

Every single person who emailed into support, if they had a signature, I would look at who were they. So, I checked that out and I’m like, “That’s interesting.” I chatted with him a little bit and he’s like “I really like what you guys are doing and it would be great for photographers. I’m thinking about sending it out in a newsletter to my audience about this.”

So, one of the things we did was we worked a little bit in terms of, “Hey, can you wait a week, let us put in some more drives.” So, for those events that we could plan around, we would try to adjust when they happened and when we had drives and we actually did capacity planning on those kinds of things.

There were certainly things where we had no idea. There was a while where we were getting a bunch of customers and we couldn’t figure out where they were coming from because we didn’t see anything in Google Analytics that would say where they were coming from. So, I just started asking people, “How did you hear about us?”

Andrew: Wow.

Gleb: A number of people kept saying, “TWIP.” I’m like, “TWIP?” So, I started Googling TWIP and I’m like, “This Week in Photography.” We’re like, “Oh, it’s a podcast.” That’s why we don’t see any analytics traffic. They’re hearing it on a podcast and then coming to the site. So, there were things like that that we didn’t have any control over.

Those that we did have control over we would try to plan and adjust. It was still literally this thing where I think we came pretty close to causing a nervous breakdown for our VP of engineering because it was just this constant for a long time. It was this constant chasing the customer demand, which is not to say that we had an easy time getting customers. We didn’t. But it was also this thing of like you never want it to have too much storage space. So, it was constantly chasing the right amount.

Andrew: That’s where if you were on Amazon, that problem would be solved because you could just take up as much space as you want and only pay for what you use.

Gleb: It would have been amazing. We would have kept our sanity.

Andrew: But still the math made sense and you guys decided, “We’re going to do it this way. It’s going to be profitable in every customer or else we’re not going to build this business up.” So, you got TechCrunch. You had other people write about you. Life was good for how long before the crickets came in because TechCrunch doesn’t last forever?

Gleb: About a week.

Andrew: A week? That was it. Okay.

Gleb: We had been working on it for nine months or so at that point before TechCrunch. We had friends and family beta testing but nobody outside using it. Then the day of launch, both TechCrunch and Ars Technica wrote about it. It was huge for us. Literally, we were like, “Oh my god, the servers are burning down. We’ve never seen so much traffic at Backblaze at that time.” We’re like, “Woohoo, party.”

So, the first day, a lot of people showed up. The second day, it was about half as many people. The third day, about a quarter as many. And then by the end of the week, it was really quiet. What kept happening was we had this background rate of customers that came. But it was like single digits a day. It was like three customers would show up every day. We didn’t know why they showed up. We were like, “That’s great they showed up.” But at $5 a month, that wasn’t going to pay salaries any time soon.

It was lean for a long time. That was in June/July. Pretty much the next time we had a lot of customers show up was, I think, in December. So, it was months.

Andrew: What did you try during that period that didn’t work? Sorry, the audio went a little funky there. What did you try in that period that didn’t work for you?

Gleb: You know, I can’t say that we had a great process at the time. So, we tried to come up with reasons for the press to write about us. But we didn’t have much. So, we came out of beta, which was a reason for the press to write about us. So, we had an article in CNET about that, but that was it.

Andrew: You just went back to PR. You said, “That’s what got us our first customers. Let’s see if we can get those guys to write more about us.”

Gleb: That was kind of the main thing. We started blogging. But at the time, it was like we didn’t really know what we were doing. We just writing about things that seemed interesting and semi-related to storage. Maybe my mom would read the blog post each day. So, it was very lean times. We tried going to some events. So, we went to SF New Tech and presented there. I think we might have gone to SF Beta around the same time and demoed there.

We actually, I remember with SF New Tech, we were really excited. We got in. I’d been to SF New Tech a number of times before and I liked that it was a good event and a couple hundred people would show up. We were going to be on stage. So, we spent all this time crafting a demo and a presentation. We even came up with a custom code so that we could track people that came through it and everything.

After the event, I came back to the office and was staring at the logs for this URL that we came out. It was like 0000. A week later, 1. It was painful. So, we tried some events. We tried going back to PR. We tried blogging. But frankly, for like four months, it was somewhere between zero and 15 people a day would show up.

Andrew: I’m smiling because I’m looking at some of your earlier blog posts. Here’s what you got. “Here’s the Gift of Backblaze.” I get it. You’re blogging and you say, “Let’s make this blog pay off for us.” There’s “Digital 007: Outwitting the Thief.” It’s about how someone noticed that their computer was stolen and what you guys could do to help out. This one made me laugh. “Weird Doggies Need Online Backup Too.” That’s just a YouTube link.

I could see that you hadn’t nailed it. You were still trying to figure out what to do. You talked about when you guys spoke at events like FailCon. You talked about new features. You talked about new launches. You talked about what you liked. None of it was working and then–you tell me, is that when you discovered this new way of blogging which is to show what’s going on behind the scenes and to appeal to a tech audience that’s especially curious about it?

Gleb: So, some of that was thought through and some of that was luck. So, in 2009, we had been using–the wooden box server evolved. So, by 2009, it was a steel server and it had kind of momentum behind it internally. We had a lot of the severs deployed. We felt pretty comfortable. They seemed to work well. We kind of got into a rhythm with it. We started talking about, “Should we tell people about this?” Our original thought was to tell people, “Hey, we’ve done this cool thing,” kind of a press release almost about it.

The thing was that part of the motivation for sharing a little bit more about it was that people were, at the time, they were saying things like, “There’s no way these guys could do $5 unlimited. Clearly there’s something fishy going on here.” The example they gave most of the time was they would point to Amazon S3 and say, “I’ve done the math on Amazon S3 and these guys are clearly losing money, which means it’s not possible.

So, the three common reasons they said why were one, what they must be doing, the must have a ton of venture capital and they’re just burning through cash and once the cash runs out, they’re going to raise prices or they’re going to go out of business.

Andrew: It’s happened.

Gleb: It’s happened. The other reason they said was, “Clearly there’s some sort of other mischievous plan, like they’re going to take all our data and sell it.” Or my favorite, “What must be happening is they’re not actually storing any data and they’re just hoping no one does a restore.” I’m like, “That’s brilliant. That would be awesome.”

So, what we wanted to do was show that no, we actually were really making money. We were profitable on a customer basis. The reason for that was we built this very inexpensive storage system. So, we thought, “Okay, we’ll share this storage design.” We said, “How much of it do we want to share?” There was angst internally about how much to share because it was part of the secret sauce.

Andrew: Google didn’t talk about for years about how they created their own servers. I remember when I interviewed Paul Graham years ago. I said, “What kind of questions would you ask entrepreneurs?” He said, “Don’t ask about the things that are in the news now. Think longer term about those pivotal decisions like what made Google decide to build their own hardware. That’s huge and no one cares about that to the degree they care about whatever happens to be in the news that day.”

Gleb: That was the thing. We had built it in secret, I think, at that point for a year and a half, two years maybe. We said, “Okay, well, we’re share some of it. We’ll share a little more. We want to share this piece because it’s interesting.” So, then we just after a lot of internal angst, my wife was a lawyer going, “Are you insane?” We just said, “We’re going to share the whole thing. We’re going to open source it. We’re going to tell everything and we’re going to show everything.”

We wrote this long blog post which literally detailed all of the specs of the storage pod that you saw. At the time, like I said, it was partially strategic and partially luck. We understood that people would think that this was fairly interesting. We understood that we were giving a fair amount away. We had shared this one chart which compared the cost of our storage pods versus EMC and NetApp and S3 and how those things compare and how dramatically less expensive the Backblaze storage pod was. We thought that was compelling.

Andrew: Let me actually read it. I’m looking at post that you wrote in 2009. Just to give people a sense of the difference, one petabyte of storage on your pods–by then they weren’t in wood anymore. They looked nice. They were in metal, beautiful red boxes–but one petabyte, $117,000. One petabyte on Amazon was $2.8 million is what your sheet shows.

Gleb: Yeah. It was an astounding difference in price. It was just the reality. Periodically we did the math for ourselves because we’re like, “Are we doing this right?” And we’d say, “Okay, well, based on the amount of data we’re storing today, if we had all this on Amazon, we’d be broke tomorrow. We’d be out of business in a week.”

So, we wrote that blog post and then we pitched the press on it, specifically the storage and techie press. We figured like, “This is interesting. They might like this.” But at the end of the day, we had no idea how many people would care. It’s a large storage server. We’re thought hundreds of people might care about a larger storage service. Maybe thousands of people would care about a large storage server. A million people read that blog post. We were just blown away. We’re like, “It’s a storage server, people.” People loved it. It went completely viral.

What was interesting was we had done a number of press announcements at that point and all of them, the trajectory was the same, which was day one you’d have some number of people, day two you’d have half, day three you’d have a quarter and by the fifth day, you’d pretty much be steady state.

This was the first time where that actually changed. On the first day, I think we had like a quarter-million people read the blog post. A quarter-million people reading the blog post for us, if you combined all the people on all the blog posts over the previous two years, but then the next day it was twice as many people that read it because what happened was the first day the press wrote about it, the second day it got picked up on Digg and Reddit and Slashdot. I can’t remember if Hacker News at that point.

Andrew: I see it on Hacker News. It got 401 points that day.

Gleb: That day was just astounding for us. We had fully expected it to die down and instead it picked up steam. So, that really changed our mindset about what could be possible with blogging. But frankly, it still took a number of years before we really felt like we were in a good rhythm. If you look at blog posts that we’ve written as recently probably as two years ago, a lot of blog posts we wrote about random stuff that we’re like, “Why did we write this? I don’t know.” It seemed interesting for us at the time.

Certainly, about a year and a half ago maybe, I sat down and said, “You know, what people really come to us for is information about storage. We have a lot of data about storage. If we’re willing to share it and we’re willing to be open with it, it will be something that people will read.” So, very aggressively tried to slow down and stop the blog posts about what the cat ate at the office and put more focus on what could we contribute back to the community about storage.

Andrew: So, what do I take from this? Is the big lesson that if you’re looking to figure out your marketing, just try a bunch of everything until you find the one thing that works and then keep focusing on that. Is that what you did?

Gleb: So, I would take that with a slight twist on it which is I don’t know that we were trying with the blog post to quite think of it as our marketing path in that particular blog post. I think that when we saw success with it, we didn’t jump on it as much as we should have. So, I almost feel like–there’s the old adage, “Do things that don’t scale.”

I think we looked at it at the time as saying, “This does not scale. Yes. We wrote this one blog post that was very successful, but what are we going to do tomorrow? We’re not going to write another blog post about our storage pod. It took us two years to build this thing and we wrote one blog post.”

So, we kind of had this perception that this is not the marketing path for us. This is just something that we’ll do once in a while. What took us a while to realize is that this could be a real channel if we focused on it and did it well. So, I think a lesson I would take away from it is if you see success on something, try it. Keep seeing if there’s some way to put fuel in the fire on that one thing. Finding something else that we think should work might be really hard. But that thing, whatever it was that succeeded, succeeded for some reason.

Andrew: Just a few days before we’re recording this interview, you guys have a post up about Storage Pod 5.0. You also write about–you do have blog posts about who’s working with you guys and about “Back to the Future” on “Back to the Future” day. But it seems like you also kept doing more techie posts. Hacker News is still a major source of traffic for you it looks like because of this posts.

Gleb: Definitely.

Andrew: What else did you find–actually, before we look into more places that you found new customers, your process is, “Come see our site. Either download the software and install it or subscribe to our mailing list. We promise we won’t spam you and then download our software to try it.” Eventually some number of people who try it are going to buy it. You told our producer that percentage is 70 percent of people who install the software and try it for free end up paying. What did you do to tweak that number so that you get it as high as 70 percent?

Gleb: So, part of that is a benefit of the overall model. So, things that are a benefit to the overall model is once people decide they want to back up their computer, the fact that you are trying something to back it up, it’s not clipping. It’s something that you actually want to do. So, that helps. But there are definitely some things we did to increase that conversion rate. One thing we did was early on, we had a set of emails you would get during the trial that would educate you on the product.

At the end of it, if you didn’t subscribe, we had an email that just said, “Hey, it looks like this is not for you. Please let us know why and if there’s anything we can do.” That got us two things. One, it got us feedback on how we could improve the product. And two, we found out some of the people were just having confusion or something and it actually allowed us to convert them into customers.

So, one of the things that we did was we moved that email up three days. Before we had it on the very last day of the trial, at which point the trial was over. We were going to delete your data from one system. Then we’d have to convince you, “Oh, hey, reinstall the trial.” So, we moved it up three days and we changed the wording just slightly to say, “It looks like you’re thinking of not subscribing. Let us know why.” So, that helped us be able to capture people before the trial fully expired.

Andrew: Okay.

Gleb: So, that was one thing that helped. Another thing that helped was really explaining to people what they should expect during the trial and what they should not. So, one of the things that we got as feedback all the time was, “Well, I couldn’t get fully backed up during my trial. So, I couldn’t test it and therefore it’s not for me.” We never expected–the whole point was it’s a trial. It may take you a month to get fully backed depending on how much data you have and how fast your connection was.

So, we actually put into one of the first mailings, we said, “Hey, there’s no expectation that you’re going to get backed up here. You can try every single thing in process. So, go ahead, right now, try to restore. Download a file. See how it goes for you.” So, people were able to do the full loop of install, back something up, go to the website, try it, download it and see if they could restore a file in the first couple days and that gave them much more of a sense of, “Oh, this will work for me.” And they didn’t feel like they had to get approved.

Andrew: I see. Misunderstanding the problems that they’re bound to have and just telling them about it has helped.

Gleb: Just setting those expectations. And then the last one that we did early on was we had a lot of people that said, “I have no idea how long this is going to take. How long is it going to take me to actually back up all my data?” So, we kept getting that question over and over and over again.

So, in the product, on the main screen of the product, on the very front of it, we literally wrote the question as it was phrased to most of the time, “How long will my first back up take.” When you click on that, you would do a live calculation based on the amount you stored and your bandwidth and give you the answer for your particular case. So, those kinds of things really helped to increase the conversion rate.

Andrew: You know, I wonder about a product that’s a backup, that people don’t appreciate it until they need it, right? It’s only when you lose it that you finally realize, “I should get my computer backed up.”

Gleb: Yeah.

Andrew: So, how do you sell a product like that where people don’t feel the pain until it’s too late?

Gleb: It’s hard. I think about it as they’re insurance products. So, your mortgage insurance, the bank requires it. Your auto insurance, government requires it. Health insurance now, the government requires it. So, I’m thinking the government should pass a law that mandates Backblaze installed on every computer. I’m just waiting for them to do that.

Andrew: Give Obama another year and he’ll give that to the American people.

Gleb: We don’t ask much, just Backblaze on every computer.

Andrew: The other thing that happens with insurance, I was just talking about this in a past interview, you have a major life event and then you buy it. You have a baby and you realize, “I should get life insurance.” You start a company or get a job and you realize, “I should get some kind of health insurance.” So, what do you do in this situation?

Gleb: So, what we find is a lot of people do sign up at that moment. So, about 40 percent of our customers sign up and they tell us the reason they signed up is because they just lost data or they know someone who just lost data. That’s where a lot of people come from. But the other things that we do, we try to make sure that our existing customers are good referrals and they don’t forget about us.

So, this is one of the things we found as a challenge. Something happened that got you to install and sign up. You’re really happy because you just signed up and you feel so much better. Then two years goes by before you need a restore, during which time you’ve completely forgotten about us and you haven’t told anyone about us whatsoever at all for two years. That’s a terrible viral loop.

So, we did a couple of things. One was we realized that some chunk of people felt perfectly comfortable beforehand, then one day realized the data would get lost so they installed. That day, they became really fearful until they were fully backed up.

Andrew: Okay.

Gleb: So, they were safer now than they were before. They felt more worried about they could see they weren’t fully backed up. So, we added an email at the moment they were fully backed up, which would happen three weeks in, a month, two months in. It would say, “Congratulations. You’re fully backed up. You’ll sleep peacefully. Everything is great.”

So, it gave us a touch point to remind people that everything is fine. Everything is good. People are like, “Oh my gosh, I’m so happy.” It was a moment where they would then go and tell people, “I’m so glad. I’m fully backed up.” They would even share that on Facebook and Twitter and stuff.

The other thing is we found another touch point. People would ask us, “How’s my backup doing?” We’re like, “What do you mean how’s it doing? It’s back up.” People needed a steady check in to tell them things were fine or not. So, we added an email that would send once a month and people could say, “I’ve never wanted,” or, “I want it weekly.” It comes once a month. It just gives you a summary. “Here’s how much data they have. Here’s how much their data is backed up. Here’s the status of things.”

Once a month they get this email. So, again, it reminds them of our existence as well as giving them the comfort of how their backup is. So, a lot of our process is about how do we make our current customers happy with the service and therefore and remembering to tell others about it.

Andrew: One other way I saw–I don’t know if this is your marketing or someone else’s–but Jon Gruber, when he did sponsorship spots for you guys, there was a period there where he didn’t promote Backblaze as a backup solution, but as a way of getting access to your files anywhere. Was the an internal decision? It seems like he changed back to calling it backup.

Gleb: So, we have multiple use cases for the product. Backup is the main default one. But we have an iOS app. We have an Android app. We have a website. So, people do absolutely use the service to get files just from wherever they are. One of the nicest things is because of the Backblaze where we backup everything, all of your data is in the cloud.

So, a lot of times people have told us that, “I have Dropbox but I’m on their 2 gig plan. I was out somewhere and I wanted to grab something from my phone that was on my computer but I hadn’t out it into the Dropbox folder because I didn’t want to put it in there because it was only 2 gigs. With Backblaze, all the data is in the cloud, so you don’t have to think about it beforehand.

Someone emailed us saying that we saved Christmas because there was this one particular weird something or something that their kid fell asleep to and they would only all asleep if they were listening to this song. It was on the computer at home and they flew to their in laws. They were like, “This is going to be a disaster if we can’t play this song.”

Andrew: That seems to me like more of a typical need. I may not lose my data more than once a year, but I’ll need a file when I’m away from home multiple times a month or away from the office multiple times a month. So, it seems like that’s a bigger use. But when I’m looking at your website, you’re still promoting Backblaze as personal backup.

Gleb: Yeah. So, on the Gruber side of things, it’s not that he changed it and then changed it back. We sponsored Gruber very regularly. So, a lot of the people here about Backblaze every single week.

Andrew: He’s looking for ways to change it up.

Gleb: So we gave him different parts of the product that he could talk about. From an explanation perspective, one of the things that we want to be careful about is we’re not a Dropbox replacement. We’re something that you would use in addition to it. So, access is a great feature of what we do, but it starts with backing up your computer, especially with services like Icon and GDrive and Dropbox and everything else, we’re trying to say, “Yes, you can have all those, but Backblaze will have all your data so you can go sleep peacefully.” Then as a thing you can do, you can access it.

One of the other features we have is locate my computer, which we very rarely talk about, but it is one of the things that when your computer is stolen or lost, not only do you not want your data back. You might want your ex-computer back.

Andrew: The other thing you guys recently did was you said, “We figured out how to get prices for ourselves that are lower than Amazon. What if we give that to the rest of the world?” So, now you’re comping with Amazon on their simple storage solution product by undercutting them. What’s the price difference between you and them?

Gleb: We are about one-fourth what they charge.

Andrew: So, how do you compete with Amazon. Even if you’re one-fourth, they’re giant. They can crush you on prices, can’t they? If not today, they’re getting there. If all of this is pricing, isn’t that a dangerous spot to be with Amazon?’

Gleb: It might be. I think here are several things that make it interesting for me. One is that Amazon has absolutely dropped prices year over year over year for the last eight years. But our pricing internally, our cost internally have also gone down and down year over year. We started lower cost than Amazon eight years ago when we compared our costs and their prices. We’re still, eight years later, lower cost than their prices.

So, looking at history, we can continue to lower the cost. Thy will consider a lower cost, but we’re continuing to be lower than that. One of the things people ask is, “Wont’ they just drop all their process and crush you.”

I think the interesting thing is, “I doubt it, at least not for a while.” They make plus or minus $1 billion or $2 billion from just S3 alone. If they cut their prices to a fourth, they’re going to lose $1 billion. At some point, if Backblaze is selling $1 billion worth of storage and eating all of Amazon’s lunch, they might look at that and say, “It’s worth it.” In the meantime, it’s not worth them losing $10 million from us or $15 million or $20 million and cutting their cost by $1 billion.

One of the things we also look at is they have several different tiers of service. They have Amazon S3 and then lower cost than that is reduced redundancy and lower cost than that is infrequent access. And the lowest, lowest, lowest cost service they have is called Glacier.

Glacier is this services where it requires four hours before you could even start getting new data out of it and it’s got these other various random costs around like if you delete too soon or you want your data back it’s expensive or whatever else but it’s like if you just want rock bottom cheap storage, they will offer you Glacier. Three years ago, they came out with Glacier at a penny.

A month ago, the first time in three years they’ve lowered the cost on Glacier. It was a whopping, dramatic reduction down to seven-tenths of a penny. Backblaze is already at half a penny. So, we’re 40 percent lower cost than Glacier, which is their lowest cost of to survive, for a service you can’t get to the end and back out of. That’s where our real time use lives. So, they would have to lower their normal service price below their Glacier price. I think we have a little bit of ways before they start trying to do those kinds of things.

Andrew: And for this you raised money?

Gleb: We didn’t, actually.

Andrew: I thought you raised money already, no?

Gleb: Yeah. We raised money three years ago. We were bootstrapped for five years. Profitable, making millions in revenue. We decided to raise a round of funding at that point. It was around the time of the Thailand drive crisis.

Andrew: That was another set of blog posts that was interesting.

Gleb: So, we could chat about that. It was around that time we decided we wanted to raise some funding and put some money on the books. We have not raised any money since them. We’ve announced and launched this new [inaudible 01:12:42], Backblaze v. 2.0, all just on the revenues and profits on the company.

Andrew: So, why did you raise? According to CrunchBase, you guys raised $5.3 million in two rounds. It was $5 million in 2012.

Gleb: Exactly. So, when we started the company, no way we have any reasons for funding–heads down and work on the product. There were various reasons form that and having mixed experiences in the past. One of the things we wanted was we wanted to be able to just focus on building the company, not on pitch decks and pitching and board management and all the stuff that goes along with it when have VC funding.

The other thing that was really key to us was that we built the culture to believe that money comes from customers and not that it rains from the sky.

Andrew: Did you raise money because you guys wanted to cash out a little bit?

Gleb: So, we had a couple things. The biggest thing we wanted to do was we wanted to put a little bit of cash on the balance sheet. The timeline drive crisis made us nervous. It was this moment where we looked at it and said all of a sudden if we had a cash flow issue, then the whole company here just [inaudible 01:13:58] that was fine in year one, but now in year five.

Andrew: The Taiwan drive issue is it was flooding in Taiwan. You didn’t realize hard drives are made in Taiwan. If there’s an issue in Taiwan, it’s hard to get hard drives over here. You guys couldn’t get the drives using your usual way. You guys went–I forget what it was called–drive dumping?

Gleb: Drive farming.

Andrew: You would go out, buy hard drives at local stores and I’m guessing also on Amazon, start ripping apart the cases and putting them into your devices.

Gleb: Yea. It was crazy, crazy time. So, it was in Thailand. It flooded overnight. Prices just doubled. So, what we normally buy are these internal hard drives. They’re these raw hard drives that you put into circles or inside of a desktop computer. That’s what we normally buy. Those became impossible to buy at any reasonable price. So, we looked at that and said we were going to go out of business within two or three months because we had to pay twice as much or three times as much for hard drives.

So, we started what we call drive farming. We noticed that you’re regular external hard drive that you would get as a consumer from Best Buy was not much more expensive and was available. So, we just setup a map on a whiteboard in the conference room and said, “Here is where every one of the stores around this is. Here’s where everyone that works at the company is. Each person is responsible for going to one of these stories in the morning before showing up to the office, picking up some drives and bringing them to the office.

We would then do what we call shucking, shucking like an oyster where we would take the drive out of the enclosure wand then put them into the server. We did that for a while and then the stores had limit two per person. So, we got friends and family in on it. Then it was still not enough. So, we actually put a bounty. We emailed our customers and said, “If you can buy a hard drive at your store locally, we’ll pay you the cost of the drive for every drive you can buy and ship it to us.”

Andrew: And people did that?

Gleb: People did that.

Andrew: For $5? That’s got to be because they’re fans.

Gleb: Fans. Right. I think the $5 was just a nice thank you.

Andrew: Yeah.

Gleb: But ultimately they wanted to help out. One of my trips was one Christmas day. I went to my friend’s house who happened to manage to buy just a boatload of drives from Newegg or somewhere or TigerDirect. He lived in a one-bedroom apartment in San Francisco and his apartment was shoved full of boxes of drives. He’s like, “You have to come to my house today because I can’t walk through my hallway.” So, we packed up an SUV, got all these drives, drove them down to the office. The other folks here unboxed them, shucked them. We work to live another day.

Andrew: One of the things you said in that blog post was the price of a hard drive at the local Costco or Best Buy was $169, which is $100 less than what you guys were usually paying.

Gleb: It was $100 less than what was available if we had just bought–

Andrew: I see. So, your prices still had gone up, but not as much because you went direct to the stores.

Gleb: So, they went up maybe 20 percent.

Andrew: So, that’s why you decided to raise money, part of it is to put cash on the balance sheet, but it seems like part of then is to cash out a little bit, take some money off the table.

Gleb: Yeah. There were basically three reasons. One was put some money on the balance sheet because we didn’t want to go out of business because we were running that close to the edge. One was we wanted to run some paid marketing experiments. We’d never done any paid marketing. So, we wanted to try that and see if that worked. One was because we wanted to take some cash off the table.

So, we worked for a year and a half without salary. We worked for another year basically at minimum wage and maybe another year at double minimum wage or so. We definitely had gone through quite a lot of our savings. So, we wanted to take some of that and refill the coffers.

We did something that is a little bit unusual, which was as opposed to just the founders taking money off the table, we did what’s called a tender offer, where said, “Anyone who owns stock, whether they are employees, contractors or early angels investors. Anyone who owns stock can sell a certain percentage of their shares.

It was the same amount whether you’re a founder or a contractor or anybody. You could sell the same percentage of your shares in this VC round if you want. It was completely open and you could choose to do it or not do it. About half the people said yes and half the people said no, which meant we felt like it was priced perfectly. So, it was a good way of doing that too.

Andrew: I see that you guys are doing some paid ads. In fact, you guys have an AdWord for “Walla,” which was one of your competitors that went out of business. You guys jump on top of stuff pretty fast.

Gleb: We try. We’re still a pretty small company and we try to be fairly lean on stuff. Some of these things work, some don’t.

Andrew: What are your revenues now?

Gleb: So, we don’t publish exactly, but I’ll tell you one of the things we did publish is that in 2014, we were over $10 million in revenue.

Andrew: 2014, last year, over how much?

Gleb: Over $10 million.

Andrew: $10 million. Is it fair to say you guys have doubled that this year?

Gleb: We haven’t doubled, but we are growing pretty quickly. On average, we are growing somewhere between 25 and 50 percent every year. We just actually were on the Fast 500 list. So, 225 percent growth, I think over the last four years.

Andrew: I’m looking to see what your competitors have and I couldn’t do it fast enough. As we talk, I keep Googling and Googling. One of your competitors is worth about a quarter-billion dollars. I want to see what’ their revenue to give me a sense of where you guys would end up. I can’t find it fast enough right now. You know what I’ll do instead is ask you to close out with this one great story about how when you were in high school, you decided to buy a car–oh, two stories, actually. There’s one other one that we’ve got to talk about and it’s much shorter. But in high school, you decided to buy a car and you noticed something. What did you notice?

Gleb: So, in high school, I wanted to buy a car. I didn’t have a lot of money. So, I was looking at cars that had like a $500 range. So, I was looking at the San Jose paper and the San Francisco paper, just trying to find a decent cheap car to buy. I noticed the San Francisco paper, you could by a car for like $50.

Then the San Jose paper, the minimum you could by a car for was $500. I was like, “Oh my gosh, there’s this great arbitrage opportunity. I’m going to buy all these cars for $50 in San Francisco. And I’m going to shuttle them down to San Jose and put them in the San Jose paper for $500 and I’m going to be rich. What I realized was I was just misreading the paper.

Andrew: Really?

Gleb: Yeah. So, the San Jose paper had a separate section called something like junkyard or something. It was not the car section. Flea market–they called it flea market. You could sell anything you want for under $500. So, if you had a car that was for sale, you could not advertise it for under $500 in the auto section. You could only advertise it in flea market. So, you could actually buy cars with $50 in the San Jose paper. You just had to look in the right place. So, I gave up on that.

But in the process, I started noticing that cars were being sold for different prices, even if they were same make, same model, same year, same everything. So, I called up a friend of mind because at that point, I didn’t even have a driver’s license. I was 15. So, I called up a friend of mine who was 17 and I was like, “Let’s go and buy a car together and then resell it and try to find a good deal.”

So, we bought a Camaro Firebird or something and cleaned it up a little bit, touched it up, washed it. It was painful. It took a month because there were rains and everything. But at the end of the month, we sold it for like $2,500 and bought it for like $1,500. We’re like, “Oh my god, this is amazing. This is great.” We just started buying and selling used cars for a while until we realized it was illegal.

Andrew: Why is it illegal to buy and sell used cars? Oh, you need a license don’t you?

Gleb: It turns out you need a license. But here’s the funny thing. You are absolutely allowed to buy cars. You’re absolutely to sell cars. You’re absolutely allowed to make money on the fact hat you bought and sold a car. You’re just not allowed to intentionally do that. If you happen to buy a car and you sold it and it sold for more, that’s fine. But if you’re doing this with the intention of making money, you’re running a business.

Andrew: So, what did you end up doing because of that?

Gleb: So, we said, “Okay, let’s get a dealership license.” So, we actually got a license, but in order to get a license, you had to have a lot. You had to have at least two cars on that lot. It had to have an office. It had to have a sign. The sign had to have a phone number on it. These were the requirements legally to get a license. So, we did that. We got a lot at the very end of El Camino on a dirt lot with a trailer that was yellow and we put up a sign that said “Cheap Cars” and tried to make a business out of it.

Andrew: Were you really selling it off the lot or were you selling it basically from your house but you needed the lot to justify that you were doing it?

Gleb: We actually started selling off the lot as well. We started putting ads in the newspaper because–it didn’t matter whether it was a lot or anyone else. It was a disaster, but we have a good college try.

Andrew: Why did it fail?

Gleb: What’s that?

Andrew: Why did it fail?

Gleb: Well, we didn’t really know anything about how to run a used car sales business. We knew how to do our hustle. We knew how to buy a car inexpensively from the newspaper and sell the car for a little bit more. What happened was with the whole used car sales thing, all this stuff came with it, like fixed expenses, which was a novel concept that we never had before and employees–I actually went to college at that point and could come down for the weekends.

So, we hired people. We had no money to hire people. So, either we hired them at 100 percent commission or we hired them at minimum wage plus some commission. What we realized was we probably weren’t getting the top salespeople of the land at that point. Then we had this whole thing where people would show up and they would say, “I can’t afford to buy the car, but finance it for me.”

So, we tried to work with financing companies. Nobody really wanted to do it. They would put us through all this paperwork. We thought about and we’re like, “We’re buying a car for $500, selling it for $2,000. If they give us $250, on the very first payment, we’ll break even. On the second, we’ll make money. What we found was that the people who were financing cars from us at that point, the only money that we ever saw was the money we got for the down payment. So, we learned a lot of good lessons.

Andrew: That one didn’t work out, but Backblaze did. Why don’t we close things out by–I’m curious about your environment right now. You’re doing this interview with good lighting, with a good internet connection because you followed the infographic video we sent you. I’m curious about where you are. You said earlier you couldn’t really move your camera because you might disconnect something. We’re done with the interview. Let’s take a risk. What happens if you turn your camera around? What will we see?

Gleb: What you’ll see right now is closet, kind of.

Andrew: Are you literally in a closet? Let’s see.

Gleb: You can see there are some around servers right there.

Andrew: This is a closet.

Gleb: That stuff up there is blocking what is a bunch of rack-mounted switchers that are attached to the ceiling up there for networking for the office. There’s a window here.

Andrew: Okay.

Gleb: And I can’t turn it all the way around because we’ll get disconnected. But there’s a door directly in front of me. So, what we’ve done basically is we had an open floor layout. That was great. We liked the open floor layout. But for phone calls and things like this, it’s very, very noisy.

So, originally, we just built a phone closet, not the one that I’m sitting in, but the one next to us. We just up some sheets of plywood and some foam on the inside of them. You would slide this sheet of plywood that we mountain handles on to close it up. That’s one of our phone closets when you need a little bit of privacy, but not a lot of privacy.

What I’m sitting in used to be the office data center. So, when we moved in here, we needed a place to just store all the office data center equipment, not any of the stuff that any of the data for the business goes on, but just the stuff we’re working on at the office. So, we built this completely enclosed mini-data center where it’s really sound proof. That door, actually, when you close it, it has a special sound barrier that goes on the bottom and everything.

Andrew: Everything is hacked together. I asked you in the beginning of the interview were you a millionaire from the first sale and you said no. You clearly are at this point and the company is a multimillion dollar business. You’re still like living out of a closet surrounded in your neighborhood by people who have beautiful offices, who have incredible furniture.

I talked to a guy recently and on my next call with him, he couldn’t talk long with me. I said, “Why? What’s going on?” He bought a jet and they had to take the jet off to go cross country with this freaking thing with his friends. He just wanted to chat with me real quick. Do you ever feel like, “At what point do I get to live it up after working this hard?”

Gleb: So, there are certain things that I really value and certain things I value less. One of the things I really value is I show up to the office. I like that I do and I like the people that I work with. So, at least as far at the company, that is one of the most important things. I realized it took me way, way too long to realize this. For years my thought about work was, “I will show up, work hard and as long as the work is interesting and as long as I’m just kicking ass and as long as I’m progressing in my career and making money, that’s important.”

Andrew: I get that.

Gleb: I realized at some point where I was in a job where I had a lot of those things but I was miserable every single day–that sucks. If you’re not enjoying what you’re doing day to day, all the rest of it isn’t worthwhile. So, one of the things I’m most proud of at Backblaze is we are eight years into this and we have only had two people leave the company–two in eight years. One of them just came back a week. He restarted as our VP of sales.

Andrew: Nilay Patel was one of your cofounders. Is he one of the two people who left?

Gleb: One of the two people. He’s the that just came back. So, I attribute that mostly–yes, the company is growing and yes there are perks and this and that, but I attribute that mostly to the fact that you show up and you like who you work with and you like who you work with and you like what you’re working on and it’s interesting. It’s cool. I’ve never aspired to own a jet. I do aspire to certain things. Like I love to hang glide. Luckily, hang gliding doesn’t require a ton of funds.

I love to travel. So, my budge has gone up in my travels. Before Mail Frontier, I was living on roughly $5 a day.

Andrew: In an expensive part of the country.

Gleb: No, not here. I went traveling for four months. During that four months, I was living on about $4 a day.

Andrew: Okay.

Gleb: Certainly now between me and my wife and now our daughter, it’s like–

Andrew: What’s one luxury that you now have that would have been mind-blowing to you when you were in the $5 a day days?

Gleb: Honestly, the number one luxury is peace of mind.

Andrew: I mean like when you travel. Do you have a butler follow you when you travel yet?

Gleb: No butlers.

Andrew: I’m working on that.

Gleb: I have to say, it stresses me out. When I go to nice hotels, they want to carry your luggage and they want to really help you, it stresses me out–too much. I could carry my own bags. It’s all fine. So, the luxury is more of like we feel like going and going to a show. Tickets tonight are $300 to see that show, but we want to see the show. So, we’ll buy the tickets and see the show.

It’s the peace of mind of being able to say, “I feel like doing this,” or, “I feel like going on this trip. I wasn’t able to pre-plan tickets six months ago when they were cheap. Now they’re expensive. But we’re available to go. We have free time tomorrow so we’re going to go.” So, it’s a very like–that just relaxed…

Andrew: I know what you mean. Yeah.

Gleb: It’s literally the best luxury of all the stuff.

Andrew: I remember a comedian saying when you’re broke, you do a lot of math. You sit down to eat and you’re constantly adding everything up. It’s nice not to have to do the math. It’s nice to just say, “I feel like ordering this food. I feel like going on this vacation this time. I feel like going to see a show.”

Dude, you built something fantastic. It’s so cool to have you on here. I have to admit. There was a part of me that said, “We can’t get him on.” I don’t know why. Sometimes there are like blocks in my mind that I don’t know I’m aware of. That’s why it’s good to have a team. Somebody on the team–let me see who found you, who said, “Andrew, we should be interesting someone from Backblaze?” Who was this?

Gleb: You know what Backblaze used to be.

Andrew: Yes. That’s a story I almost told. Go look up Backblaze. Google that. Toren [inaudible 01:33:37] emailed us. Oh yeah. He was doing some work with you guys and he said, “Listen you guys have got to check these people out.” Then he started sending us links. So, Toren, thanks for opening my eyes to this.

Gleb: I probably had the same blog. Toren was like, “Hey, you should be on Mixergy.” I was like, “I’ve never even though about it. I’ve listened to Mixergy a bunch of times.” It never occurred to me.

Andrew: It’s so cool that you’ve been listening. Thanks for listening. Thanks for doing the interview. That’s always been my dream. You might have heard me say it–have people who are listening to these interviews, build up their companies and then hopefully come back here and say, “Here’s how I did it, learn from what I did and then the next person will be listening and come back and do the same thing.”

Gleb, thank you so much. Thank you all for being a part of Mixergy. The sponsor, Toptal–if you need a developer, go check out Toptal.com/Mixergy. If nothing else, they have a hot freaking model on that website, Toptal.com/Mixergy is where you see the hot model.

Thanks, Gleb.

Gleb: Thanks so much.

Andrew: You bet. Bye, everyone.

Who should we feature on Mixergy? Let us know who you think would make a great interviewee.