Krassimir Fotev scoffs at SEO with his brainchild, Peer Belt

If you’re like me then you are sick of SEO. These cookie-cutter predictive patterns that create a myopic filter for the content we see online make as much sense as casting a rapper in a film noir, or taking a three-year old to a fine china shop.

This must be what Krassimir Fotev thought when he looked at how we gather, process and retrieve information and thought — pssst… I can do better. So he did. (Don’t you just love people like that?). He created Peer Belt — a personalized search engine that organizes the content you encounter online.

After first getting a masters degree in Physics, then giving up a cushy job at Credit Suisse Krassimir created Peer Belt, and hasn’t looked back since.

We had tons of questions for Krassimir — like what the heck Peer Belt actually does, so we fired off a series of rapid-fire questions to him about his entrepreneurial experience and the technology behind the brand. He then banged his head against his desk until something brilliant came out. And brilliant it was…

For any of our users not familiar with Peer Belt could you give them the elevator pitch?

Peer Belt is a personal search engine that organizes the content you encounter online. Think of it as of a personal assistant that deduces the research context and takes note of the information nuggets you uncover. Browser extensions remember everything and make it easy, fast, convenient getting back to articles you found interesting and engaging.

Peer Belt does not expect nor asks for an explicit feedback, likes, bookmarks, tags. You are already doing hard work filtering in what matters. What you like is reflected in your actions on the Web. And this is the key – our software makes sense of your behavior in the browser and translates what it sees into something tangible without you having to do extra anything. It implicitly knows what you like and shows you what you want to see.

Peer Belt seems to be an autonomous way to improve how you find and relate to information on the internet, without giving up your online security or privacy–what was your inspiration behind this idea?

It all started with considerations about how Search Engine Optimization (SEO) could and should be removed from our lives. While running an online e-commerce site I found the SEO eating big chunk of my time. It was robbing my customers from the better experience they deserved. I thought there should be better and more user-centric way of organizing the information out there.

I also thought it would be beneficial in understanding the individual better. Perhaps the person’s behavior could bridge the multiple online identities, organize information consumed not just in the browser but in a pleiad of mobile applications. It may provide a unified view of what matters out there. Some of the information that matters is personal and should stay that way. It cannot be better protected but being kept on the device where the actions took place.

There is a content explosion on the Web with the cost of producing information down to zero. We are consuming information behind signup, rely on real-time streams. The old well connected, static Web is evolving into something new, more interesting which happens to be fragmented. Within the new environment there is nothing to hold the medium together but us – information producers and curators at the same time.

The changed information flow and the access to ton of information challenges many businesses that have performed well historically. Understanding Human behavior turns a key element (this is how it should be). But the behavioral information collected and processed does not necessarily work first hand for the person providing it.

Without giving up any trade secrets, could you explain the technology behind Peer Belt and how it works?

Of course. Peer Belt has browser extensions for Chrome, Safari, Internet Explorer (support for Firefox is coming soon). The extensions keep track of the research context and collect statistics about how the user interacts with the content currently on the screen. Then an indexing component, that is common across the multiple browsers installed, translates the collected behavioral data into a single number – a content relevance measure. Each and every page you see, except for page over secure channel or pages seen in incognito or in private browsing, ends up indexed.

When you happen to need a particular blog post, technical article, something important through Twitter or Facebook, Peer Belt takes your query and executes a lookup against the local index. The results are merged with the results your search engine of choice presents. It is extremely convenient and fast way of rediscovering what you found most engaging . And the content Peer Belt organizes for you is not limited to just the public information the Web search engines expose – it is any content you have access to. This includes Web 2.0 apps, personal data in Facebook, Enterprise applications you interact with. It works like a charm.

Through you actions, Peer Belt intuitively knows what you like and surfaces back what you want to see. For instance, I was looking into giving credits to the photographers for the visuals used in the Peer Belt’s Ignite Fluent presentation. It happened about 5-6 days after the ppt file was ready and before it went to public. Can you imagine the amount of time you need to put into this effort? Selecting images for the set in stone presentation script took good 6 hours. The credits required that I retrace my steps from near a week ago. The browsing history was not of any help, considering the amount of visuals reviewed. But Peer Belt was able to bring back the exact same images the presentation stepped on. I was done in matter of minutes.

We currently have platform extensions from Google, Bing, Yahoo, DuckDuckGo. We are opening the indexing engine and letting others create extensions for their Web sites. With the API we will provide sample implementation for Twitter for instance.

What is your educational and work background?

Studied Physics for many years – first at the National Math and Natural Sciences High-school in Sofia Bulgaria, then at Sofia University where I earned my Masters degree. Sofia University was big part of my life even during the High-School years. We had two great professors from the university – Karlukovski in 8-th grade and Maksim Maksimov for the next 3 years. Prof Maksimov was dedicating time to us each and every Saturday outside of the regular Physics classes at school. The net result – my classmates and I did great at the HighSchool Physics Olympiads. I won a bronze medal at the XXII iPhO (International Physics Olympiad). My Physics study was backed up by a scholarship from Eurika Foundation.

Came to US in 2000 working for Webmessenger. The company had spare resources that went into an Instant Messaging product. Taking pride in the initial protocol and the highly available server design. Webmessenger was acquired by Aptix in 2007. Few years before the event joined Credit Suisse IT. At CS worked for various projects at Investment Banking, Fixed Income divisions. The highlight of my career there would the the non-trivial work done for Structured Products Research.

You quit a cushy job at Credit Suisse to start this venture, why did you feel it was worth the risk, and has it indeed been worth it?

Of course it is worth it! Money-wise one may have doubts. For now… But if one broadens the horizon a bit it is the desire to help similar minded people and progress toward this goal that makes the effort worthwhile.

I finally have a chance to work on something that truly matters. Actually, it is the second time I have an opportunity to make a difference. The first time around it was the work on my PhD related to non-linear time series analysis and forecast. Because of the full time engagement with Credit Suisse I was moving too slow there and about two years of work I did ended up being scooped in the summer of 2002. A publication in Physics Review Letters introduced new recurrent time based metric that broadly gives an idea how long time series is needed to make accurate forecast for a specific dynamical system. Exactly what I was working on. It hurts.

With Peer Belt I knew I could not do part-time. This time around the life is somewhat more complicated with the kids and family I love so much. As with everything that is rewarding the family deserves attention. And this translates to time – pretty scarce resource these days.

The work on Peer Belt has introduced me to a totally different than the big Enterprise world I spent 8 years in. I am so happy for that. I finally met people making difference first hand. They are by definition open minded, sharp, internally driven. It is amazing experience being near them.

What are your biggest challenges as an entrepreneur?

Finding resources to do everything I know needs to be done. Time being of essence. There is three options: one either puts time on the work himself, finds similar minded people that care about the idea and the work done, or finds money to buy others people time. All this needs to be balanced.

It becomes more complicated with the time the family needs, the time one needs for himself. Typically something gotta give. For instance I love windsurfing and I used to go windsurfing on Sandy Hook’s bay each and every weekend. Visited Sandy Hook not more than 5 times in the last two years.

How have you worked out the technological kinks that you have undoubtedly encountered?

The past experience comes to the rescue in cases like these. Not just knowing how to inject a library in the Internet Explorer process if IE happens to not expose what you need.

In early 2011 I was not quite happy with the results Peer Belt was giving back. The instinct was to look for help from professional linguists. But it turned out the Physics/Math education helped understand and put to work methods described in Natural Language Processing papers. It was more resource efficient than going out and recruiting unknown people. In this example I went back to stuff learned back in 8th or 9th grade at high school.

In conclusion, the experience matters. The more diverse the better.

What are your plans, hopes and marketing ideas for the future of Peer Belt?

The biggest concern or push back on Peer Belt is the distribution. Though at present day there is examples of client side software that got good adoption, the World does not stand still and ripping apart and copying exactly what worked couple of years ago will not produce same results.

We have recently identified a guerrilla marketing strategy that will help Peer Belt leverage our partner distribution channels. There is work to be done and I do not want to disclose too much prematurely.

In terms of what the future plans were a friend after using Peer Belt said in writing the software should be integral part of each and every browser. I would like to think the technology behind Peer Belt should be part of each and every OS and device out there. I know people need a unified view of the most valued information out there. Peer Belt could technically provide this. As stated earlier it is resource or as people refer to it an execution game.

What are some of the mistakes and/or failures that you learned from while starting up?

Look where the future is. Let the past rest peacefully. Referring to the Windows implementation of Peer Belt. This is where Peer Belt ran initially (and is still running). Then there was a Mac version. From a common code base, but months later. Now we are moving to Android, iOS.

Might have started the other way around. But coming from the Enterprise I knew the Enterprise runs Window and Internet Explorer. To keep the Enterprise option opened for Peer Belt I started with the most challenging of all implementation – the one that targets the old IE, still with 60% share at the time. A waste from my current perspective.

My main focus today is minimizing waste.

If you had it to do all over again, what would you change about your entrepreneurial approach and why?

I would do canned demos looking for cash resources immediately after some core proof of concept functionality is ready. May not even bother finishing the proof of concept. I know I will make it running eventually. If unsure I would not consider going through the exercise in a first place.

People generally get excited about an idea and the faster one can provide visuals that demonstrate the idea the better. The problem with this approach is one shows off something that may not work at the very end. But the learning is so much faster. Plus it is much easier to pivot to something the works before being emotionally attached to anything.

What development, event, or new understanding has had the most impact on your original plans for Peer Belt and how have your plans changed in response?

Since I started, there is many. Reading Good to Great, How to Win Friends and Influence People, Linked, The Filter Bubble to name few. There are articles out there on the structure of the Web, Web size, crawling the Deep Web. Some of them came late, but they confirmed the approach taken by Peer Belt was well-grounded.

It would be unfair forgetting the conversations I had with company founders, each and every time they commented after a pitch. Typically, the harsher a grounded (one that follows what you intuitively know is not working well) comment, the more impact it will have. One just needs to listen and keep the comments in mind for future actions.

Comments have adjusted the plans. For instance opening the indexing engine via an API may not seem reasonable from a pure marketing perspective – you have to market the API. You have to market the end-user product too. But the API unlock partnerships, which is an important distinction.

How did you land your first customers?

Family, friends. But everyone has limited supply of them.

From there, presentations play significant role. Being on the social channels. Having strangers refer to the work is rewarding and keeps you going.

What made you first think that starting a company is right for you?

I was not thinking of building a cash printing machine. And I do not see the companies that way. I had identified a problem and had a solution. I wanted to help people be more efficient at discovering information. Facebook and Twitter are awesome channels, but there is no good search engine on top of them to date. I understood it was not a single person endeavor. Forming a self-sufficient entity capable of rewarding the project contributors (employees, investors, advisors) seems a good idea.