Dwight Merriman

**Jessica Harris** (0:05)
I'm Jessica Harris, this is From Scratch. My guest is Dwight Merriman, co-founder of MongoDB, an open-source database company that focuses on helping companies manage their data in a more flexible, scalable way. Customers in MongoDB include the insurance company MetLife, Goldman Sachs, Facebook and eBay, among several others. Dwight is also the co-founder of Doubleclick, the ad-serving platform that was acquired by Google for $3.1 billion in 2005 He's also co-founded the fashion e-commerce company Gilt Groupe and the news web platform Business Insider. Dwight is originally from Ohio and now lives in New York City, where MongoDB is based. Welcome.

**Dwight Merriman** (0:51)
Thank you.

**Jessica Harris** (0:52)
So how does MongoDB work?

**Dwight Merriman** (0:55)
So MongoDB is a database. It's an open-source database. This is a tool that engineers, software engineers use to build systems, right? To store data in. The MetLife example is, you know, MetLife is a large company, right? So you have all these divisions, all these products, all these different types of insurance. And in that bigger company, how do you get a single view of all that information? So they're pulling in data from over 60 data sources internally into a single data store that's in MongoDB, so that when you call in, the person you're talking to trying to help you can see everything.
The thing about those 60 databases they're pulling things in from is those databases are constantly changing, right? Those teams are making upgrades to those apps, and when they do that, the data in the database, the shape of that data changes, and now how do I react to that? I've got 60 inbound data streams, but they're all fluid, if you will. And it turns out with MongoDB, that's one thing it's very good at.

**Jessica Harris** (1:57)
Now, your experience at Doubleclick, which you co-founded in 1995, established the germ for MongoDB. Can you talk about some of the frustrations that you were experiencing at Doubleclick?

**Dwight Merriman** (2:11)
Yeah, so I was the tech guy, I was CTO there. So, you know, we're building this system, we're serving 20, 30 billion ads a day, half a million ads a second, you know, so we have big scalability challenges to deal with. You know, one of the challenges was data. It's so much data that the traditional way to scale is what we call it's called vertically, which means just buy a bigger server, right? But there's a limit on how big you can go vertically is a problem. And the other problem is those big servers are very expensive. So what we really want to do is we want to do what people do today with cloud computing, which is to scale horizontally, which says, okay, I got a really big database, it won't fit on one server, tie together 100 servers into one database system. So that was something we really wished we had back then. So there was all these sort of like homegrown solutions, but there wasn't really anything out there just for everyone to use. There's generally available products. At some point, myself and my co-founders at Mongo, Elliot and Kevin, there's got to be a better way to do this. There needs to be a general solution to this. So we just started building something from scratch which turned into MongoDB.

**Jessica Harris** (3:24)
Where does the name MongoDB come from?

**Dwight Merriman** (3:27)
The idea was that the word Mongo is in the middle of the word humongous. And one of the key benefits of the product is this ability to have big scale and then do big data, right? It also helps you do rapid application development very elegantly. But that it comes from that humongous was sort of the pun there.

**Jessica Harris** (3:50)
MongoDB is open source. So, developers are welcome and invited to help shape the infrastructure of MongoDB. What is the catalyst for the open source movement more generally? I mean, I think of Wikipedia, for example, as being my first encounter with what open source means.

**Dwight Merriman** (4:10)
It's actually, I think, a few things. One is I think the Free Software Foundation was a big proponent of this concept of free software, which, at first, when you never heard of it, it sounds a little silly. It's kind of like free cars or something, right?
But it actually turns out it does work. Then the internet, I think, was a big catalyst because to do this sort of community collaboration, going back to the Wikipedia analogy, you need some way for all these people to talk to each other and they're not all in the same place. So the internet facilitated open source, I think. And then another thing that helped a lot is actually IBM. In the early days of Linux, IBM came out and said we're going to invest a lot in Linux, and that was a big strategic decision by them. And that was a real milestone in open source, I think.