Josh Zelman is a Video Producer for AOL Tech. Prior to joining AOL he was a producer for CNN/HLN where he produced thousands of live studio technology segments for primetime programming. He was sent on assignment to cover such high profile events as Fortune’s Global Economic Forum, the Consumer Electronics Show and the Electronic Entertainment Expo.
Josh graduated from... → Learn More
Chris Dixon begins this episode of Founder Stories with DoubleClick and FindTheBest Co-founder Kevin O’Conner by telling O’Connor that “DoubleClick is probably the closest thing New York has to a PayPal.” Meaning the two companies share an aptitude for hiring employees that go on to start innovative businesses. Just as Paypal spawned Yelp, YouTube, and LinkedIn, DoubleClick spawned dozens of startups in New York City like Right Media.
With this in mind, Dixon asks O’Connor if he intentionally created an environment that encouraged innovation while at DoubleClick, before going on to ask O’Connor what he considers to be the defining characteristics of successful entrepreneurs.
O’Connor describes his systemized way of spurring innovation. He lists a few signs of a successful entrepreneur. A desire for control and the ability to challenge authority are near top of the list. A way to detect the second trait? “Have you told your boss to shove it?”
In the video below, O’Connor discusses his genesis of his latest startup, FindTheBest, a service that offers side-by-side comparisons of everything from cigars to business schools. At the intersection of trying to solve a “big obvious” idea and not finding a site that served his needs, O’Connor began to build what would become the company. He describes it as a “decision engine” to help you make complicated decisions where search falls flat. What dog should I buy? Where should I send my kid to summer camp? What investment adviser should I hire?
Make sure to watch both videos for additional insights along with the first two installments of this interview (Part I and Part II), as well as past episodes of Founder Stories here.
you're focused full time on Find the Best? Is that right?
Yes. So Find the Best is, yeah, I wasn't sure if I could actually go through another startup again.
In the meantime by the way we've invested in a bunch of companies and you've done a lot of other things. I mean a contact with that stuff.
Yes, I figured I'd you know take the totally ignore my family for 20 years. Thought it'd be a good time to get to know my kids and what easier way to do that than to do angel investing.
Yeah.
But angel investing is tough, right, it's tough to get, you know, you're not, just not the same as you know cause you've done it. You don't get obsessed with what you're doing. You're like a ham and egg breakfast. You're like the chicken. You're just participating.
Yeah.
Sometimes it is a lot better to be the pig.
Yeah.
Fully committed. So I was just really looking. Where there any big obvious ideas out there that no one did.
So that was Find the Best.
Yeah, so Find the Best. Yeah, so what is Find the Best, People might know, but just so people in case they don't.
We launched about a year ago. But Find the Best is a decision engine. We help people and businesses make really complicated decisions easily. So it could be anything from "where do I send my kid to college", "which dog should I buy for the family" or "which document management system should I get for the law firm" ?
About three or four years ago I kept running into these same problems but I didn't realize they were connected. I was "Which school should I send my kid to? Which ski resort is best?" Deer Valley always shows up as like number 1, if you're a skier.
Yeah, it's just, it's not number one. I was looking for web hosting company for my venture thing I wad doing, I except running into one I was compiling these spreadsheets to be able to compare stuff or for like web-hosting, you find these affiliate shills, right? Guys telling you what's the ten best and it was just bull.
I just felt like the internet was going down this path of... it's kind of a cesspool. All the searches showed up with affiliate.
Everything has been so heavily gamed now in the SEO side and everything else it's like you know.
I think the Panda, the Demand Medias of the world etcetera who just create these like, content farms. And Panda's helped. Would you say that?
It's helped out a lot, yeah, I think.
Panda, is the, just so people know, is the latest Google core algorithm update that was meant to eliminate a lot of the content farms and it seems to have helped.
So first is, how do you compile this information structure data so people can make a side by side comparison of something, right? So Kayak, around the same time too Kayak had come out. I thought Kayak was brilliant for travel.
Yeah.
Really taking a complicated system giving you all the filters to control it. Why can't that be done for the other thousand big decisions in your life? It's all the same, the same problem. Mm-hm.
Giving people choices, narrowing it down.
And so how do you do that? How do you build the data-? How do you figure out what the best documenting system is, for example? You mentioned that as an example.
So, this problem people have been trying to address for a long time. Because Google and Bing, they've all realized that search is great but sometimes you get structured information that you want to be able to compare. And people have done a whole bunch of different approaches to it. Google tried Google Squared where they crawl and scrape and... I always thought that was an interesting - that was a very little noticed project that was pretty - I thought it was from a technical point of view very interesting.
Yes.
Google Squared.
But it could never...
Especially where they- I just want people to know... it was like it would take - you'd say "Show me the U.S. Presidents" and it would try to actually provide a structured spread sheet of U.S. Presidents and their birth dates and other things and I thought it was just, as a techie I was like, "That's cool". It was interesting.
It's a really non trivial problem.
And we were very aware when they came out because when that came out we're like, "Oh my God the last thing for them they didn't seem to push it. I don't know.
It can't be done. I mean our view is semantic stuff.
It works, like, 80% of the time.
Yeah.
There's a joke and in machine learning circles that you can or natural language circles that it's very good in solving every problem 80 percent of the way or something like this not a funny joke. Yeah, but it's, like...
Yeah.
...so it sort of gets you very quickly part of the way there but then you end up with these ridiculous errors that you need humans or so you do have a human, you guys do have humans involved in it them.
Absolutely, so and then the other sort of approach people have used is either purely semantic way or the other way is toward these open data based projects where you create this platform.
Crowd sourced It just doesn't work, our feeling is that a human needs to lay the foundation. How does someone decide which dog to get for the family? What's the important decision criteria? How does someone decide which summer camp to send their, send their kid?. to, which investment adviser I should hire for my charity.
You know, what residential property management system should I use for managing my small, small apartment complex.
So you actually go through for each category and sort of have some smart person figure out what are the key criteria that you can technically use in a good decision.
Then you got to figure out how do I get that data and for us the data comes with three sources. We either have databases and we scoured every sort of government database. Governments put out some pretty interesting data it's just not in consumable format.
Yeah. It's not normalized. It's like a total mess.
Yeah.
It's in flat files or whatever, some random format.
Yes. Yeah, and so we really specialize in bring that stuff in. The other one is we go out and we just gather it, you know, we go out and we outsource a lot of data collection. So for something like ski resorts, you know, we'll go 1200 different ski resorts and get their data and just normalize it.
Yeah.
Maybe get on a normalizaton part. I mean, that's the problem.
Um-hmm.
How people represent information that's why scrapers don't really work.
Yeah.
They're just not represented. Or their in multiple sources.
Um-hmm. Then the third one is crowd sourcing. We want people, businesses to come in and modify, add their listings.
I see, so you provide sort of the core, maybe the seeding and sort of the core structure and then you allow some kind of crowd sourcing layer on top of that?
Kevin O’Connor is the co-founder and CEO of FindTheBest.com. Before founding FindTheBest in 2009, O’Connor started O’Connor Ventures in 2001, a company specializing in tech startups. In 1995, he co-founded DoubleClick, an Internet advertisement-technology company which was acquired by Google in April 2007. Prior to founding DoubleClick, O’Connor co-founded the Intercomputer Communications Corporation in 1983, a microcomputer to mainframe inter-connectivity company. When the ICC was acquired by DCA in 1992, O’Connor eventually became its CTO...
FindTheBest, based in Santa Barbara, CA, and New York City, is an unbiased, data-driven comparison engine. We organize and present data in a consumer-friendly format so that you can make quick and informed decisions based on what’s important to you. What Kayak does for travel, FindTheBest does for the thousands of other decisions in your life.
FindTheBest was founded by DoubleClick founder and CEO Kevin O’Connor and CeBotics CEO Scott Leonard and is backed by Kleiner Perkins.
