Don Dodge has had ridden the startup rollercoaster many times, perhaps most infamously at Napster. As a Developer Advocate at Google, a special advisor to Google Ventures, and prolific angel investor, he keeps very close to the startup world and can be found judging pitches and meeting with founders at conferences and developer meetups around the world.
Elliott Adams: I haven’t met many developer advocates that work alongside the full spectrum of building high-growth companies, it seems like you have a pretty broad interest in startups.
Don Dodge: Yes. Well, I am also an advisor to Google Capital and an advisor to Google Ventures and an angel investor myself in forty-six companies. So, I spend a lot of time at startup conferences and seeing new companies and finding companies that can either partner with Google, or they could be an investment for Google or in rare cases, an acquisition. So, it’s sort of my job to go out there and be an advocate for these startups back to Google, to bring the message of what Google needs to do in the product, in terms of features, to be successful with emerging companies.
EA: I was reading up on your background, and I didn’t realize this before, but your were part of Napster?
DD: Yes, the original Napster.
EA: Do you mind talking about how that came about and what that experience was like?
DD: Sure, but before I do that, I was Director of Engineering at AltaVista, one of the first search engines, and my group invented what we called “multi-media search” at the time. It’s hard to believe, but seventeen or eighteen years ago, in 1998, you could not search for an image or a photo or music or video because the technology to do that did not exist. So, my group invented the image search. We started with image search, then video search, and then music search. As part of doing that work, I was always looking around to see what the competition was doing, and there was this new thing called Napster that no one had really heard of, but they were doing music search. And I thought, oh, this is interesting.
EA: Was AltaVista metadata-based, or it was actually looking at file content?
DD: To reveal the secrets, the way that we did it was we were already indexing the page with everything on the page and doing nothing with some of this metadata. If there was an image, it might have a file name or it might have some attributes around it or words that describe it. We just started using all of that data that was on the page to try to identify what the image was because back then, you couldn't. There was no other way to do it.
EA: So, it might be “EiffelTower.JPG”, and you made an educated guess on that.
EA: So, then you saw Napster, and they were doing search and—
DD: Right. So, as we were finishing up the music search part of the equation, I saw Napster and Scour and CuteMX, and others like them. I was always looking around to see what other companies were doing, and it turned out that Napster was two blocks down the street from AltaVista, and the recruiter in HR that I worked with at AltaVista was the girlfriend of one of the founders of Napster. So she said, “Hey, you should go to Napster and help them out.” Shawn Fanning and Sean Parker were eighteen years old at the time—eighteen—and just trying to get started. Long story short, I went over to Napster when it was just eight people, and then it took off like a rocket, and the rest is history.
EA: I think there was something in your LinkedIn profile about that period of time, how you held on for dear life, or something along those lines.
DD: Yeah, it was a rocket ride, and there were no ways to plan for it. I remember Eddie Kessler would order servers for the backend server part, and he thought he was ordering enough servers to last us for a year, and they would last us three weeks or a month, and he'd have to buy more and have to plan the capacity on something that you really could not plan. It was a rocket ride.
EA: You know, that’s a great anecdote that speaks to you totally, completely computing with cloud.
DD: Well, it didn’t exist then, so we had to build our own infrastructure. There was no Google Cloud or Amazon Web Services that would scale for you. You had to do it yourself.
EA: Sure, since you guys were not even renting, you were buying servers.
EA: Were you co-locating them in your facility?
DD: No, we had a hosting service that would give you rack space, and you'd put your servers in there and manage them and that kind of thing.
EA: I’m not sure I can think of, off the top of my head, another young company in 2000 like Napster that had that kind of consumer-driven growth.
DD: Yeah, AOL took a long, long, long time. It was very slow growth. Steve Case is a personal friend of mine, and he’s told me about the first ten years of AOL. They grew very, very, very slowly, and it was only in the eleventh, twelfth, thirteenth year that they took off. But at the time, and I think for a long time after, Napster was the fastest-growing web business ever.
EA: That would seem to be the case.
DD: There were just no corollaries. There were no other examples to say, “Oh, okay, we can plan based on this or that.” It just grew so fast that we had to just hang on for dear life trying to make it work. And to some extent, Twitter was the same way. I don’t know if you recall when Twitter launched the Fail Whale?
EA: Oh, I do.
DD: I mean, Twitter would crash several times a day. So, I would guess that the early days of Twitter were like the early days at Napster, where it just grew so fast, there was no keeping up with it.
EA: So, if at the time there was no precedent for that kind of growth in the consumer web—
EA: —and utility computing from Google—
DD: Didn’t exist.
EA: Yeah, it wasn’t there at the time. When did Google get into that business?
DD: Google started building its own servers and its own data centers almost immediately, but nobody really knew about it because it was under the covers and it was only for their own internal use. It was not until much later when Amazon did the same thing, and Amazon started providing Amazon Web Services that Google decided, “Hey, we could do that too because we’re the best in the world at building data centers, and why don’t we—we have to build that infrastructure anyway for our business, why not just build it even bigger and provide Google Cloud Services?” So, it was somewhat later, but at the time of Napster, there was nothing you could do. Nothing that was scalable.
EA: What's exciting to you with all of the young startups you see pitch these days?
DD: Well, I think young companies can start cheaper than ever before because we have these cloud services and all kinds of free services to get started. Back in the day, to start a company, you had to raise $5 million and be in development for a year or two before you ever got to market. That was the traditional approach.
So, it was a long, hard road, and you had to raise significant money to do it. Now, going to consumer—and I think largely because of smartphones, which really came about in 2007, 2008, and really hit their stride in 2010—you are able to start a company with almost nothing. You could bootstrap it with almost nothing and get something working and get it distributed on a mobile platform for nothing, basically, and get enough traction to then go to investors to say, “Hey, I need some money to scale.” So, that’s become the new normal, where investors expect that these founders will start the company, build the product, get the initial customers, prove out the revenue model, and then ask for money.
In one sense, those are very high hurdles to meet. In another sense, it is very inexpensive and cheap because you don’t have to pay anything up front for these services to get your company going. So, a couple of tech founders that are great coders can do it all themselves without any money.
EA: And so, that is table stakes today?
DD: Well, right. I think it’s never been easier to start a company, but it’s never been harder to build a business.
So, it’s very easy to start a company and there are thousands of them started every year, but because the barrier of entry is so low, the hurdle to get to a point where you actually have a business, where you have customers that pay, is harder. And I think that is where the problem is. There are lots of startups that have interesting ideas and they can build a product or a working prototype and get some initial customers, but they don’t know how to scale it. They don’t know how to identify their target market, how to advertise, or how to do customer acquisitions in a cost-effective way to get to the scale that they need. And that’s a really difficult problem that a lot of startups overlook.
EA: So, as an angel, again, table stakes are much higher. If someone says to you, “I built an app,” you're probably not impressed by just that statement.
DD: No, not at all.
EA: If someone comes up to you and says, “I’m doing X.” Is there an “X” that impresses you today?
DD: No, I am never impressed with an “X” or what the product is. I’m more impressed with the team. Where do they come from, what is their background, why are they the ones to do this, what special experience or issues did they have earlier so that they completely understand this problem?
EA: The jockey, not the horse.
DD: Right, it’s the jockey, or the team of jockeys, that’s number one. Number two, what’s your go-to-market strategy? One question I always ask is, “How do you get to $10 million in sales?” It’s a simple question, but most startups have spent so much time thinking about the product and the features and the competition and the blah blah blah blah, they really haven’t really given much thought to do the math—"How do I get to $10 million in revenue? How many customers do I need to do that? What price do I need to sell it at? How am I going to target that market? What is the customer acquisition cost going to be?"—just going through those simple calculations to get a rough idea of how much it takes to get there.
EA: I think you might put a lot of startups on their heels, right on stage, don't you think?
DD: Here’s one for you. Back in the day when this started to emerge, all startups—well, not all, but many startups were advertising revenue-based. They were going to build an audience, monetize the audience with advertising. And I say, “Great. Great idea. Let’s go through the math. How many pageviews do you need to make $1 million?” And they had no idea what the CPM rates were or what the ad rates were, and how many pageviews or how many customers you needed to make just $1 million. And when I took them through the math, they were stunned. They could not believe it. They just assumed that you build the product, you create an audience, and you monetize it with advertising. Done. No, not done.
EA: Startups who say, “We’re not going to charge for this. We’re going to sell the data.” Do you hear that a lot?
DD: Sure. That’s the next version of, “I’m just going to monetize with advertising.” “Oh, we are going to monetize the data.” “Sure, yes, great idea. Let’s go through the steps how you do that.” And they just don’t know. They assume that it just sort of automatically happens, like advertising automatically pays for this audience that you created. So yes, it’s a very workable idea, data. I’m invested in several companies that that’s their business, selling data. And it’s a model that works, but it’s much more difficult than people realize, and you need much more scale than most people realize.
EA: Scale for one and I assume domain for the other.
EA: In-app data from a medium-sized game maybe won’t move the needle, right?
DD: Right. I mean these founders are jumping for joy if they have a 100,000 downloads. They think, “Oh my God! This is awesome!” Yeah, it is kind of awesome, but it’s not enough to make money. If you want to sell data or if you want to do advertising, you need hundreds of millions, not hundreds of thousands. So, I think it’s the age-old problem in a different context today than it was fifteen years ago. It was all about creating an audience and monetizing it with advertising. Now it’s creating an audience and trying to upsell or sell the data to someone, but the scale that you need to do that and the quality of the data that you need is much harder than it looks.