It was 17 years ago that he founded the pioneering Internet advertising company, DoubleClick - and over four years since he helped sell it.
But, original DoubleClick founder Kevin O’Connor can’t get the entrepreneurial bug out of his system and has targeted "recommendations" with a 60-person firm he co-founded in 2009 called FindTheBest.
Like the self-explanatory name, O’Connor is blunt on the problem he set out to solve with FindTheBest : “The Internet was full of garbage - a lot of distrustful information, affiliate shills, people telling you what was best even though they were getting paid to say it.”
He believed that air travel site Kayak was an early turning point against the flotsam by putting recommendation tools in the consumers’ hands. Inspired, and with the backing of venture capital firm Kleiner Perkins Caufield Byers, he set out to help with thousands of other decisions in life saying, “Marketers used to have all the power of information, but it's been shifting over the last 10 years. Today, consumers have all the information to make big decisions.”
This often used phrase "big data" -- is that something you’re thinking about with FindTheBest?
KEVIN O’CONNOR: I've been obsessed with data for 20 years. With DoubleClick, it was all about who could predict which ad the consumer was going to respond to. FindTheBest is using data to drive decisions. But, the real challenge with big data is - what useful information can you use to either predict an outcome or drive a decision?
We’re focused on taking big data and making it small data - manageable. We’re figuring out the taxonomy for every decision. How people make decisions is the same across categories, but obviously the data's completely different.
So we take big data and make it what we call “consumerable,” making it usable by the consumer or business.
How is recommendation different than something like search?
Search is great for finding any bit of information that you want. But, if you want to find what ski resort in Colorado has the most snow, search is terrible. It's terrible for driving a decision.
This is the next evolution of services - vertical services. How do you parse down web pages and organize them along a topic so you can compare things?
People have done this in travel and shopping comparisons, but they haven't been able to do it in a broad area. They try. Google tried it in Googlesphere. It didn't work very well.
Part of it’s a megabit problem, too.
How does this all relate to your previous work with DoubleClick?
One thing we saw with DoubleClick was that was one piece of overwhelmingly valuable data - purchase intent data. Interests, psychographics, or even context, it really comes down to purchase intent.
So is FindTheBest data, purchase intent data?
Yes. For example, it’s someone comparing a Tacoma [truck] to an F‑150 - it's about as strong an indicator as you can get that users are interested in buying a truck. If the user is comparing nursing homes in Santa Barbara, they're probably looking for a nursing home.
What is the most popular use case right now that you're seeing with FindTheBest?
Our most popular comparisons are things like cars, colleges, financial advisors, CRM software. We have over 300 [categories], so it's spread out.
Is there's a B2B opportunity with this?
It's probably one of our biggest opportunities. For example, on enterprise software, we do comparisons, and that's been very successful. If someone's about to make a half‑million or million‑dollar purchase decision, they want to narrow it down to two or three players, and figure out who meets their needs. That's what we do. B2B is probably one of the most difficult audiences to find.
One thing we noticed here, and at DoubleClick, too, is that when you do find somebody with purchase intent, there’s an exponential decay in their performance level but you've got about 30 days where they're going to make a decision.
For a big purchase intent decision, they're probably never going to make that decision again, so for a marketer, you've got one shot or you'll lose this contender, probably forever.
Cars, you get every seven years. Colleges, hopefully, only once. CRM software, at least once per job.
Given your knowledge of FindTheBest and DoubleClick, what’s the mistake that web publishers are making today?
The basic idea of a publisher is their content is a proxy for their audience. It tells you what that audience is about.
The publishers know that the closer you get to the purchase funnel, the more valuable the advertising, so the bigger the purchase intent advertising they can get, content, they're going to get bigger advertising dollars.
They always want to mix commerce with content. I always thought that was one of the most idiotic movements. Its head keeps emerging every couple of years.
The problem, of course, is that once you start doing that, you destroy your editorial integrity, and it just doesn't work. Someone's reading a story, that doesn't mean they exhibit any kind of purchase intent.
One of the things we've had great success with, working with over 70 publishers right now, is a program we launched seven months ago. Our content is all about commerce. That's all the content is. If someone's making a decision comparing automobiles, we're not providing them any editorial or very little editorial. It's all about driving a decision.
We're having a lot of luck with publishers bringing product guides together with their content and audience. Not mixing it, but exposing it in a reasonable way. That's probably the biggest mistake I see.
How does FindTheBest make money today? - a referral business?
So, we just started monetization. We make money anywhere from advertising all the way down to lead generation and actual sale of the product. We like to take risks because our audience is very high converting.
Who's the target market today in terms of clients?
Probably our biggest challenge is that we're extremely broad and have nine different categories that cover 90 percent of the GDP of the US. It's broad. There are more than 350 different product service comparisons.
For us, the big opportunity is going to be working directly with the clients and performance marketers. We have a lot of people yet to come in. “Crowdsource” is the wrong word, but vendor source is a part of it where they update their listings and make sure it's well presented and up-to-date. So, that works out great for us from a lead generation point-of-view - the person responsible for marketing, updates the listing. That happens a couple hundred times a day.
Any plans to go and get more funding?
We're exploring that right now. We don't need more funds necessarily and can get to profitability in the next five months. We have plenty of money, but we're a little torn because the opportunity's so huge. There's no shortage of data out there, and we haven't touched any international at all.
Given your DoubleClick days, what are some of those key learnings that you're using today in FindTheBest?
Probably three things.
First, constant innovation and forcing it on every level - figuring out 100 things we could do, and finding the three things that are most important.
Another is just test and scale.
And then, there’s brainstorm prioritization technique – BPT. It’s a technique where we identify a problem, get a bunch of people in a room, and brainstorm. Everyone gets a number of votes that they can throw at what they believe is the best answer. Usually we find that it coalesces around three or four ideas. It eliminates all the politics and builds a strong consensus, and creates interesting ideas.
A big problem that startups have is that they focus on 100 things as opposed to three things they must do. They get de-focused. I've found that most markets are determined by one or two significant, technology features. If startups need to focus on the important features and get rid of the politics.
Finally, the DoubleClick platform as it exists today - many in the industry see it as an unstoppable force. What do you think?
I'm thrilled that it's been so successful at Google. Love it.
Our approach when we built it was to be respective of the advertiser and the publisher.
Publishers want to make the most money, and advertisers want to have the most effective advertising. That was the real challenge at DoubleClick. We approached it from everyone's got to be happy.
And, it really gets back to purchase intent. If someone raises their hand and says, "Why did Google build a $40 billion business?” In three words, it’s “purchase intent rules.”
We're working on a purchase intent network. I don't think that's been done well, yet. Some think purchasers get mixed with interest and it comes down to the averaging. People want larger audiences, but it just gets diluted. Purchase intent needs to be broken out very separately.