You are here:  Home  >  Data Education  >  Current Article

SemanticWeb.com “Innovation Spotlight” Interview with Kevin O’Connor, CEO of FindTheBest and Founder of DoubleClick

By   /  October 23, 2012  /  No Comments

The founder of DoubleClick.com, purchased by Google in 2007 for around $3.7 billion,  Kevin O’Connor, spoke with me about his newest venture FindTheBest.com. Founded in 2009, FindTheBest.com makes recommendations and comparisons for just about anything of interest on the web.  Kevin tells us all about FindTheBest.com, recommendation engines, and future plans for his company.

If you would like your company to be considered for an interview please email editor[ at ]semanticweb[ dot ]com.

Hi Kevin, after leaving DoubleClick, which was sold to Google, what made you decide to start another company?

Kevin: I resigned as the CEO of DoubleClick in 2000, although I did remain the Chairman until the company was sold in 2005. I had spent 17 years working 80-hour weeks—and loved it—but ultimately decided I wanted more balance in my life; that meant spending more time with my family.

My passion for tech, however, never faded. I wanted to find a way get back into the tech world, but still have time for all the other important things in life. So I decided to start my own venture capital firm—O’Connor Ventures—and began investing in promising startups like Surfline, Meet-Up, Procore and Travidia.

I honestly didn’t think I would get back into the tech world as a founder, but I was becoming more and more frustrated by the chaos of the Web and I wanted to find a way to organize it.

 Sean:  What inspired you to start FindTheBest.com? What problem where you trying to solve?

Kevin: FindTheBest was founded out of three fundamental problems I saw with the Web:

  1. A lack of organization—There was no one place to go to find all the information you need to make quick, informed decisions across a range of topics.
  2. Users didn’t have the proper tools to narrow their choices, compare options and make a quick, unbiased decision.
  3. Finally, I kept coming across recommendation sites that were affiliate shills masquerading as unbiased review and rating sites.

I became obsessed with finding a solution to better organizing the Web—and that’s why I founded FindTheBest

Sean: The problem you are solving is very complex. Considering the endless reasons someone might choose one thing over another; how did you begin to map out an algorithm for making comparisons?

Kevin: There are three main parts to FindTheBest: technology, human curation and algorithms.

We spent two years and $2 million perfecting our data-driven comparison platform, which powers hundreds of comparisons across nine verticals. If you think about it, when you’re trying to make a big decision—no matter what the decision—there are three main steps you go through: narrowing choices based on personal needs and preferences, reviewing what experts and users think about your possible choices and comparing the results. FindTheBest’s data-driven platform does just that—helping users make quick, informed decisions.

But for something like this to work, there has to be an element of human curation. Companies have tried to do what we’re doing by scraping data or doing it completely algorithmically, but it doesn’t work. We started with the realization that everyone has a different idea of what “the best” is. For me, the best ski resort, for example, might be one with the most number of runs, or the greatest amount of snowfall. For a family with young kids, their priority might be a resort with a kid’s school. For each of our comparisons, we have vertically focused product managers who thoroughly research what people take into consideration when trying to decide on anything from college to a car. We then include those considerations, or data points, as filters for users to narrow their options. We also put a lot of effort into optimizing the placement and order of filters based on how frequently they’re being use. That is the human curation element of FindTheBest.

We use algorithms in our ratings, automated narratives and AssistMe feature—which overcomes the problem of sometimes limiting-filters that could rule out a potential best fit simply because the option was priced a dollar above the set price filter, for example.

Sean: Making comparisons seems difficult enough. You also make recommendations by asking sets of questions to users.  Do you see these as two separate problems?

Kevin: Our core focus is to help consumers make quick, informed decisions—everything we do revolves around this goal. So we want to include as much information and as many tools as possible to help consumers get to their best choice.

Having said that, helping consumers compare options and making recommendations can be seen as two separate elements depending how the recommendations are generated. In FindTheBest’s case, our AssistMe tool—which provides users with personalized recommendations—is based off of the personal preferences of our users. Instead of coming to a decision based on more objective filters—which works great when users know exactly what they want—people who might still be trying to figure out exactly what it is that they want can use AssistMe to answer subjective questions and have personalized recommendations returned to them. For example, the AssistMe feature on our colleges comparison asks questions like: “How important is the national rank of the school?” Users can choose from five options ranging from unimportant to very important.

So while providing users with the tools they need to compare options and offering them personalized recommendations based on their input can be seen as two different things, we see it more as two complimentary approaches to help users get to their best choice.

Sean:  How does the site utilize semantic technology? It seems like a great fit because you are discovering deeper connections between things that people are interested in.

Kevin: We have incorporated a lot of semantic elements into our platform, including infographics, contextual charts, related topics and comparisons and we also just rolled out automated narratives. This is an area in which traditional search still has a long way to go, though Google is trying to tackle the problem with its Knowledge Graph.

Through colors and graphics, users can easily see how a product or service stacks up to competing listings. In terms of ratings, if the color is green it’s better than average, if it’s yellow it’s average and if it’s red its below average. Other infographics show averages, ranges and where the particular product or service lies within that range. The battery life graphic on our tablets comparison is a great example.

Charts that contextualize data:
Everyone talks about the importance of data, but data needs to be put into a context for it to be useful. For example, without having access to contextual information, how do you know if the price, screen size or battery life of a smartphone is actually good? Or if the gas mileage or horsepower of a particular car is within the average range of similar cars? Through our charts, we show the average and range of related products.

For example, if you’re looking at the University of Michigan—my alma mater—our charts will show things like the average and range of similar schools in terms of things like tuition price or school size. This way, important data that’s considered in the decision-making process will be laid out within a context.

Related topics and comparisons:
For each listing, we include related side-by-side comparisons that our users have previously created and also link to related topics that might be of interest.  For example, for our car comparison, we relate specific models to other data sets like model recalls and pollution emissions.

Automated narratives:
We also just rolled out narratives that present relevant data in a text format that incorporates our raw data. Our Investment Advisors comparison, for example, includes these automated narratives in the overview section; we create a basic text and then our algorithm fills in all the data points like advisor name, the number of accounts under management, the dollar amount of assets managed, account balance and also generates a number that shows how much more or less the investment advisor’s account balance is relative to other investment advisors.

Sean: This type of problem is different than search. In some ways it appears more complex. Do you see more similarities or differences?

Kevin: Search is great for general research and information. But when you’re looking to make a big decision and want to be able to view all the relevant information, compare all options and have results tailored to your preferences—all on one page—search doesn’t always cut it. There’s a level of human curation that has to be involved for this to be successful, and traditional search just isn’t structured to be able to tackle this problem.

Search is great for finding a needle in a haystack, but sometimes you want to compare needles and search just doesn’t work so well.

Sean: What types of comparisons are getting used the most?

Kevin: Currently our top performing comparisons are smartphones, colleges, cars, dog breeds and financial advisors—we have a pretty good traffic spread across all of our categories.

Sean:  Where does FindTheBest.com get its data from?

Kevin: We get our data from three sources:

  1. Public and 3rd party databases
  2. Primary sources (manufacturer websites)
  3. Expert sources (vendors themselves and people have an expertise in a certain area)

Sean: The site also has the ability to login using Facebook. How much has using Facebook’s API improved the experience for users?

Kevin: We used to offer all the standard ways to login but found 80% of users preferred Facebook so we made it our default registration method.  We have just begun making basic recommendations to users based on their interests, however much more will be coming in this area.  FindTheBest is a utility where we want people to be able to make quick decisions, but we have to be careful integrating social since it can confuse decision-making. 

Sean: How does the company make money and how can companies ensure their products are found on your site?

Kevin: We recently began monetizing FindTheBest and are generating revenue largely through different forms of advertising including: display, affiliate and lead generation. We’re also getting ready to officially launch our new Sponsorship Program—which is a performance-based advertising solution. Vendors who are part of the program can bid to be placed at the top of relevant comparisons as sponsored listings—similar to how Google does it—and the highest bidder also gets automatic insertion into user side-by-side comparisons so that consumers take the particular product or service into consideration when making a purchase decision. Through the Sponsorship Program, we also offer other perks like having competing ads removed from a vendor’s page. Of course, the sponsored listings will be clearly labeled as such so there’s no hidden advertising.

Vendors who want to include their listings or update their information can easily do so for free once they’ve registered. To add a listing, vendors can simply scroll to the bottom of the comparison they want to be included in and click “Add Listing.” Fill out the data and then click “Add Listing” at the bottom. Our research team reviews each submission before publishing. Vendors who want to edit or add to their listing can click into their listing, click “Edit” (at the top right), update the data and then click “Update Listing” at the bottom of the page. For full instructions, visit our Adding and Editing Listings page.

Sean:  What are your current plans for FindTheBest? Are there any new features you are working on?

Kevin: We’re always working on additional features and new ways to use our data-driven platform to help people compare options and make more informed decisions. Our data-driven content platform has proven to be applicable to a large range of problems.  We’re not ready to officially announce anything, but check back in with us in a couple months.  

Thanks so much for your time Kevin!

You can follow Kevin O’Connor (@kjpoconnor), FindTheBest (@FindTheBest)


About the Author: 

 Sean Golliher (@seangolliher)  is an adjunct professor of search engines and social networks at MSU and is a member of their computer science advisory board. He is also the founder and publisher of SEMJ.org. Sean holds four engineering patents, has a B.S. in physics from the University of Washington in Seattle, and a master’s in electrical engineering from Washington State University. He is also president and director of search marketing at Future Farm, Inc., Bozeman MT, where he focuses on search marketing, internet research, and consults for large companies. He has appeared and been interviewed on well-known blogs and radio stations such as Clickz.com, Webmasterradio.com, and SEM Synergy.  To maintain a competitive edge he reads search patents, papers, and attends search marketing conferences on a regular basis.


You might also like...

Is Data Governance Solely About Controls on Data?

Read More →