Chris Seline's Searchles: Social Search Beyond the Limits of Collective Wisdom
"Collective wisdom has its limits."

This is my favorite quote from my interview with Searchles CEO Chris Seline.
Because it's true. And the Searchles solution (a three month old solution at this point) indicates a compelling direction for social search.

First, Searchles built a social network for its users in addition to enabling them to add comments, content, votes and tags.

Second, Searchles enables users to pick and choose who from their network - or outside their network - affects the relevance of their individual searches.

And by the way they're hiring.

Without further ado I give you Chris Seline. And, Mr. Seline, I have some follow up questions for you at the end of the interview that you're welcome to email me answers to or just respond in comments :)

1) Please describe your aspirations for Searchles.
We want users to discover Searchles and spend time getting to know us. We want them to tell us what they like, what we can improve and what they can’t do but would like to be able to do. We want them to tell their friends about it. And we want to do a great job of meeting and exceeding their expectations so they’ll continue to come back and tell their friends about it. Along the way, we want to make it profitable.

If we can be the impetus for world peace, invent the magic bullet that prevents every computer in the world from ever crashing again, and ensure that everyone in the world has an unlimited supply of Vitamin Water – well that would be good too – but we try not to wear rose-colored glasses or drink the Kool-Aid of irrational ambition.

2) Do you think a "Google Killer" is possible at this point?
Anyone who says yes is going to sound like a moron in some circles, albeit one day, they might just be a prophetic one. That said, MySpace and Youtube have shown that viral sites can become insanely popular within a year or two and so the idea is not far fetched.

3) Do you think it will come from the social search ranks?
That’s a big piece of the puzzle but there are others. I think you’re more likely to see a new hybrid that combines several disruptive technologies and ideas for implementing them where social search is the centerpiece.

4) What differentiates Searchles from other social search sites besides its social network?
Our search is better – we built it in house - and it is the fundamental component that drives the growth of our platform. Bookmarking, tagging, sharing and rating sites are a dime a dozen. But I’ve seen few answer the question “what for?” – what’s in it for the user? That’s where we’re different and where search comes into play.

Through search, we make it very easy for our community to connect with their existing friends and find new friends who share common interests when and how they want. We also give our community more control over who influences their search for people or content, when and how. i.e. focus only on your own content, friends, friends of friends, groups, etc. Taking this approach is a much more intelligent application of features and functionality that are otherwise ubiquitous and adds more value to the user experience.

Collective wisdom has its limits. Tapping into the masses has value - and we allow that - but we also enable users to filter out the static and focus on searching and interacting using content and/or people they trust. New features we’re working on will further advance this concept.

5) What do you see as the key challenges of search and how does Searchles address these?
Search is always going to be about relevancy. Google has done an amazing job making search relevant for a good percentage of queries, but they really only offer one angle on search. On Searchles, by leveraging the knowledge of the people you trust, you can perform searches that are personally relevant. But we don't stop there, we also let you borrow the persona of someone else and let you perform searches that are relevant to them. What if I'm not a Linux expert, and none of my friends are, but I want to do some searches from the point of view of someone who is? I can locate a Linux expert and conduct a search that leverages their knowledge and connections. This is the future of search.

6) What is the hardest part about social search?
Achieving a critical mass of users so that there is enough data in the system to dig as deep as you want to into topics of interest and tap into the best expertise. Right now we offer tremendous value in the ability to submit, tag, and share favorite sites and quickly search to identify trends and experts that might otherwise be unapparent. Beyond any one individual, this has huge applications for research, business intelligence and knowledge synthesis even with a small community if that community consists of others you trust. But social "search" will not be a substitute for traditional search until there are enough users in the system.

7) How do you address that?
Partnerships. Searchles can add value to any site with original content that has an established user base by providing a more engaging and interactive experience for their users.

8) I spoke with Eurekster's CEO about his attempted work with social network Friendster. He indicated that social network data was too horizontal to be made useful to relevance. What are your thoughts on the value of a social network to relevance?
Huh? Tell Friendster to call us. We can help. Eurekster is taking a different approach to social search and has a different customer. For their purposes, perhaps that might be true, but when you have the ability to leverage the actual connections between people, that changes the ballgame completely. Not only can you create more personally relevant results, but you can figure out who the "experts" are by analyzing the network using algorithms akin to Google's PageRank.

9) How many users do you have?
When you register, we’ll have one more! It’s not as large as others in the space, but respectable for where we are. We only came out of the box three months ago and have done practically no marketing although we’re ramping up for that.

10) What is your user growth rate?
Hard to put a number on it that will have any meaning a month from now. We’re only 3 months old. Registered users currently tend to surge when we announce something new. We also are fortunate that many of them have really taken a liking to the site and are driving their friends to it. Check back in three months and I think we can give you a better picture of our current and projected growth rate.

11) What are you doing to increase this growth rate?
Staying focused on delivering great features and functions that add value to what users and potential partners want to do. We’re less concerned about the quantity right now and more concerned about the quality – i.e. actively engaged users. Having a million registered users is worthless if the majority of them never come back.

12) How many users do you project you will need to have in order to be profitable?
We can make Searchles profitable with as few as 50,000 actively engaged users. That of course involves introducing advertising and other revenue generating mechanisms. We’re focusing on doing that in a way that adds value to rather than annoys our community and by doing so, provides greater value to advertisers as well.

13) Who are your role models or mentors? What have you learned from them?
I started my first company right out of college so I’ve never really had a mentor within the workplace. I admire anyone that isn't afraid to try something new, regardless of whether they fail or succeed. The tech community in the Washington, D.C. region is also incredibly supportive. When I need help, an introduction, advice or just about anything, people bend over backwards, open their rolodexes and make it happen. I don’t think that’s true everywhere and I think it’s one of the real advantages here.

As for role models, I would say my Dad. He has done well by creating many opportunities for himself over the years rather than waiting for something to fall in his lap. He’s probably the reason I caught the entrepreneurial bug.

14) How many employees?
Eight and we’re currently looking for more: http://www.searchles.com/misc/careers

15) How does your leadership style direct organizational change?
Well, there’s a reason I call myself dumbfounder! I know enough from my past lives to know that the key to building great products and companies is to have a solid team of people you trust that work well together, bring unique skills to the table and have a track record of getting things done. You can’t grow a company if only one person is competent to make decisions. I’m actively involved in all the major decisions but I also trust our team to help lead the company forward.

16) What do you look for in selecting people who report to you?
They need to be self-motivated and work hard when crunch time comes. They need to be honest and know when to ask for help. They should be confident in their ability without taking themselves too seriously. And they have to be brilliant of course!

17) What do you see as your responsibilities in reporting to someone else and/or a board?
I get along extremely well with all parties involved with the company. All of my major decisions have been supported 100% and they bring a lot of expertise in areas I don’t have. I’m very honest with my board and timely when it comes to reporting to them or seeking advice on opportunities, challenges and progress. They in return are very engaged and have been very helpful. I suppose we’re lucky.

18) How do you decide to say "yes" or "no" to requests for your time?
I consult with my Homer Simpson chia pet and let him make the decision. ;) Actually, I’m very accessible and I don’t see that changing anytime soon. But like any startup, we’re juggling multiple priorities to drive innovation. So it’s very simple. If someone needs me to do something and it makes sense, I do it. And if it doesn’t make sense, we talk it through so they understand the reasons why or if needed, identify another member of the team that can help. We are a team and any team member who is asked to put more time into one activity may have to take time away from another. We do our best to make informed decisions to get the most out of everyone.

19)Follow up questions:
What was the inspiration for Searchles?
What is the Searchles relationship to DumbFound?
How do you build your index?
How would you describe the nature or quality of your algorithm?

Garrett, here are the answers to your follow-ups, let me know if you have anymore!

What was the inspiration for Searchles?
There were several and then they just sort of clicked and came together. I had been working on our DumbFind search engine (www.dumbfind.com) – one of our other brands - and wanted to find an application for it that would demonstrate its power without having to invest prematurely in the big infrastructure necessary to make it a viable alternative to major search engines like Google.

And then MySpace came on my radar about the time they were acquired by NewsCorp. At first, the acquisition was very puzzling to me, so I started using MySpace to figure out what they wanted it for. I saw the power of the social network, the way it spread virally, and the fact that the network had calculable momentum. This was VERY fascinating to me on its own, but then I thought about how you could harness these people to enhance the relevancy in search and create a greater sense of community at the same time. And that’s what really inspired me.

Taking advantage of just the sheer number of people is compelling, but we can also analyze the networks so that they can create what is basically a vertical search engine based on their peers, interests, or anyone's peers and interests for that matter. When you give users this kind of power, you’ve given them the incentive to spend more time on your site because you’ve increased the quality of their experience. Rupert, did you hear that?

What is the Searchles relationship to DumbFind?
Much of the backend technology from Searchles is derived from DumbFind technology. Eventually, they plan to get married.

How do you build your index?
Right now Searchles' index is entirely populated by people and DumbFind’s index is fed exclusively through our crawler. The real magic will happen when they stop dating and consummate the relationship.

How would you describe the nature or quality of your algorithm?
Wow, um... do you have a few months, a projector, 10 years of computer science background, and a degree in cosmology? :) This is the type of question where the English language fails me.

There are a lot of different algorithms that piece together to form the relevancy score for each result in each search. We have proprietary algorithms that determine the relationships between various entities (people, links, tags, etc) in ways that are unique compared to anything else I’ve seen. Then there are some fairly standard social network analysis algorithms similar to Google's PageRank that analyze the link structure between people. But your basic relevancy balancing act comes in to play when you want to set who determines what, how much, and when, when someone does a particular search. That is something that will be continually refined.

Clarification on Voting Features
The voting/ranking/rating feature will be rolled out with our next release . Stay tuned. We’re taking a different path with this than other sites. Knowing the “collective popularity” of a particular person, content or webpage is useful, but when you know nothing about all of the people who voted for it or viewed it or why, it’s not as useful a measurement. Context is key.

A good simple-stupid example is pizza. A lot of people think Dominos is great because it’s quick and a good value for the money. But it’s not authentic Italian and if that’s what you’re looking for, you’re going to be disappointed. Searchles gives our community better, faster tools for making informed judgments based on criteria that matter to them and people they trust.
Garrett French - MarketSmart Interactive | permalink | +marktd |