Roger and I fell into the NOSQL movement early on, circa 2009 — he as a former Apple/Oracle exec turned EVP at MongoDB, me as a lowly community guy at Cloudant hanging out with devs from Datastax/Couchbase/Neo4J/Basho. Seven years later, I was surprised to reconnect with him, and realized we had drastically different reads on whether the NOSQL movement had succeeded or failed in its aims. I caught up with him outside an incubator in Menlo Park…
Tim Anglade, Executive in Residence at Scale Venture Partners: So you’re in Menlo now, like back in the incubator? What’s that like, the incubator? Did you ever even go to an incubator before?
Roger Bodamer, Chief Product Officer at MariaDB: No, this is my very first time, and it’s quite a big adjustment. I must confess, I really do like my office, you know, way more than this type of stuff, but it’s a good place, it’s easy. We have a few desks and so it makes it easy to have things around. You don’t have to have your own kitchen and all these type of things, it’s all there.
Tim: Yeah, and there’s some energy, right? There’s some stuff you can overhear that’s always funny. I like incubators, But more importantly, you’re back, fully in the database game. I mean, you could say you never left, but you know, you did some time in Oracle, or you were in MongoDB and now you’re helping out MariaDB. And so, I wanted to chat with you because we both went through the NOSQL phase, and I just realized this morning, that we had very, very different opinions as to what that trend was about and whether or not it actually succeeded or failed or whatever that even means, and so I wanted to get that on camera, ‘cause I was in there at Cloudant and a few other companies. You were in there at MongoDB.
I was a firm believer of NOSQL in a technical sense and a business sense and I kind of feel like it didn’t really deliver everything that I hoped for it. But this morning, you were telling me you were like, “No, NOSQL succeeded,” so I want to hear that out and I wanted to talk about like, in general, I guess, how those tech trends can succeed and fail, and what that means in Silicon Valley.
Roger: Yeah, so I think NOSQL succeeded. I think there was a hype phase in which people thought that everything was going to be NOSQL. I never believed that. I mean, that was a little too much, but this is a general statement. I think that NOSQL actually has succeeded. I think there’s several companies that are going to be very successful and are being very successful in this space. I also believe, at the same time, that relational databases will continue to be the workforce, like off the database industry. But where NOSQL has succeeded is I think that relational databases, for a while, sort of took a step in place, you know, they were what they were and there was very little innovation. They didn’t know how to talk to developers anymore, and so I think they lost some grounds there.
Tim: Right, that’s the kind of part about it that I feel like failed, but looking at the start of it, I felt like from a technical standpoint, they were the real chance for it to kind of take over. Because we had to add the relational data to the standard for many, many years. But before we used relational databases, we used a lot of different things before, right? So it felt like, why not change that again? Why not have a new edge, and why not have NOSQL become the new standard the way like maybe self-driving cars are going to be the new standard in the years to come, and that didn’t quite happen, right? Let’s put it like this. And so from a technical standpoint, I kind of felt like there was an opportunity lost, but then also from a business standpoint, the lack of cooperation kind of really prevented it from really making a dent in the market, from really getting a lot of adoption in the enterprise, and I kind of feel a bit of a letdown there. So how do you feel about those two aspects, the technical aspect and the business aspect?
Roger: Yeah, so a couple of things. I mean, you and I were on stage in the very, very beginning of the NOSQL phase. And the thing that I found very interesting is I was on stage with six or seven other vendors, and they were all attacking each other, and I was looking around me on stage, and I was going like, amongst us, amongst the vendors on stage at that point in time, we barely had maybe five or 10 million dollars of revenue, and so for me, it was always about we’re trying to make a market, and so what you do is you collaborate. All the various different vendors on stage had different used cases that they supported better and some that they didn’t support as well, but it was more about the notion that relational didn’t have to be everything. Relational is very, very good for most things, but there are some areas where it’s not as good, and that’s where NOSQL has a very legitimate place. From a business perspective, too, I think what people sort of underestimate is that it takes a long time for people to change, to change their development practices and to adopt newer technologies. This is not something that happens in a couple of years. This takes easily a decade. Oracle, it was, if I remember correctly, it was founded in 1984, or something like that, and only in 1990 did they have diverse, viable product. This was five, six years. Now, things have changed, development goes faster, there’s more things available, people are more productive, but still, it takes a long time before things become available.
Tim: So there’s few things to unpack here, the first thing is, yeah, I do think there wasn’t a long enough runway, I guess, for a lot of these companies, for the space in general. Because it would take a lot of time to change, trending the enterprise, that always does, right? You look at contenders now, JavaScript, things that are really, really hot and are getting really good traction, it always takes a long time to get to that core revenue that’s powering a lot of those technical trends. Then the other thing is that the infighting, and you could say for example what’s also happening was continued right now. You hear about rumors of a split in Docker and all those kind of vendors back-stabbing each other, and trying to do competing things when considering the market isn’t completely cemented yet, or even as a tech trend for getting the market for a second. And I kind of feel like that’s kind of something we did wrong in NOSQL where we felt like for sure it was going to happen, so it was a question of who was going to get the bigger part of the pie. That’s how every company felt, maybe on the inside of like, “Well, NOSQL’s gonna be standard. “I just want to be the 50% vendor of NOSQL,” as opposed to “Let’s cooperate, “make NOSQL happen on a large scale, “and then after that or in parallel to that, “we can focus on building our own differences, “our own staff, our own message.”
Roger: I think what I try to bring, like at the time when I was at Mongo, what I was trying to really bring, and the Mongo team themselves were really focused on this, was not to be focused too much on the competition, but more focused on customer success, and I think that in the open source world, there is often a lot of sort of religion around how to do open source and how open source are you and there’s all these litmus tests, like “On the open source scale, how open source are you?” “Are you this open source or that?” and I think at some level it’s immaterial. I think that what’s very important is to be customer-focused and to be customer successful, because then I think if you look at the success of the companies that are going to make it in a NOSQL space, what you’ll find in common with them is that they have good technology and they solve certain use cases really well, but in addition to that, they’ve had a maniacal focus on making sure that their customers are being successful, and I think specifically, like in the NOSQL days, I think one of the things that was done really well there was the initial focus on developers, like making developers productive and then those developers, as they started deploying it, and starting to use it for significant use cases in the enterprise, at that point in time they became customers and started the revenue, and that’s what you’re seeing now. So I think that for NOSQL, I think again, there’s going to be two or three companies successful, and I think they’re going to do well.
Tim: Right, customer success is really key, and it makes sense, because even a trend around you kind of crumbles, or doesn’t get as big, if you actually deliver value and you actually get paying customers, you’ll probably make it. You may not be as big as if a wave really caught on, but you’ll still get somewhere. So that’s a good one.
Roger: It’s sort of the land-and-expand.
Tim: Right, exactly, and you can do it on your own individual level even if the rest of the field doesn’t really cooperate in that way. But let’s talk about larger success. So what do you think NOSQL did succeed at on a large scale? It definitely kind of forced people to reconsider the use cases to think about different ways of doing data storage, so on a technical level, I do think it bolds the conversation quite a bit, but what are some of the other lasting successes or lessons from NOSQL?
Roger: Yeah, it’s funny because if you think about relationals, so relational databases store maybe 10% of the world’s database data and that number hasn’t really changed a whole lot, even with Oracle’s dominance, and with all the other vendors’ dominance in that space, the amount of data, the amount of store transactional database data and those sort of databases, relational databases never was as large as you would imagine it to be. It was never like 80 percent, or 70 percent or anything like it, it was always sort of 10%. I think what NOSQL has done is they have consolidated, in a way, they have fragmented part of the RDBMS market, because people are for certain use cases, moving away from RDBMS, and it has cemented an area where there was a lot of fragmentation, a lot of files, a lot of ad hoc databases before, so I think that that’s what’s now happened is I think you’re gonna have a document store, you’re gonna have a standardized style and you’re gonna have relational and I still maintain that relational will continue to be the vast majority of databases, because relational is exceptionally good at most use cases, like at the very edges of it. There are older technologies that do better than relational, but if you have a general purpose at a really, really significant scale, relational is just awesome and I think there’s a lot of innovation that can be done there.
Tim: And I’m not saying we should have gotten rid of relational, I think as a useful engineer, it’s kind of like you want to take over and change everything, but in the inside, it isn’t necessarily worse on a technical level or business level, but I guess the largest thing on NOSQL so far, not that the move is done yet, there’s still companies doing well, is that maybe NOSQL didn’t remake the database world in its image, but it did create a new normal. The market now is different. People’s behavior towards their data is different, and it creates opportunities for companies, not just in the NOSQL category, but also in analytics and in-memory processing and all kinds of connected areas, big data, machine-learning.
Roger: Absolutely, and a lot of those will be powered by all kinds of different technologies, including relational, but also including non-relational. One of the things was that when we sort of got started with the NOSQL movement, it was very interesting, because at that point if you were to go to your boss and you say, “Hey, I’m going to store my data in Redis,” then people will go looking at you like, “Are you on crack? “Are you completely insane?” The question was, is it going to be MySQL or Oracle, or any of those guys, like Redis or like Mongo, I mean like are you insane? and now it’s a choice, it’s an option for certain types of use cases, and I think that’s the right thing to happen. So I think that that is success, actually.
Tim: Right, it is, you’re right. It’s not necessarily a personal success for the trend, like the trend didn’t take over the world, but it did create a new situation, a better situation and I think the companies to your point, they were kind of ready for that. They were ready, whether NOSQL was going to win or NOSQL was going to change things or falter. I think they’re the ones that will end up sticking around, and so that’s maybe the best way to kind of think about a trend and how to live within it. You’ve got to be ready to live on, even if the trend doesn’t succeed at the highest level of your expectation.
Roger: You know, sometimes, I think that maybe the trend was no pragmatism, because what happened was that people had use cases and they were trying shoehorn them into certain existing solutions, and at the same time, there were people going like, “You know what? “We’re not going to do this anymore. “We’re going to build alternative solutions for this,” and there’s a certain type of pragmatism that’s embedded in that that I actually really appreciate and it’s happening.
Tim: And same thing for the vendors that were able to add that pragmatism and recognize it.
Roger: And make non-traditional choices.
Tim: Well, cool, thanks so much.
Roger: Yeah, thank you.
