BK: What is Notiora, and what is your text analytics technology about?
JM: Notiora solutions transform unstructured text into actionable information, economically answering business-critical challenges in Supply Chain Management, Web and Email Filtering, and Literature Mining. Basically, if a company has a lot of unstructured text... We help them make sense of it and then make business decisions on it. The heart of the technology is "Non-metric optimization", which is a broadly applicable capability to take a traditional algorithm, such as naive bayes or mixture of experts for classification, and transform it in to an optimal classifier.
BK: How is your technology different from those available from Verity, Recommind, and others?
JM: This requires a technical answer. Notiora's Non-Metric Optimization is unique in that it is not a single classification model. Instead, it is a technology that is applied to other classification models to achieve optimized classification performance. It can be applied to Logistic Regression classifiers (the LRC model from Verity). It can be applied to matrix approaches like Latent Semantic Indexing (related to PLSI from Recommind). It can also be applied to Bayesian and mixture of experts classifiers (common in email filtering), as well as most any other classifier method used for text classification today.
BK: I see you recently signed a deal with Singlefin to use your technology in Singlefin's managed services. How did that deal come about, and how is Singlefin using your technology?
JM: The Singlefin / Notiora announcement is a large deal in the sense of combining what we feel to be the most advanced Machine Learning technologies with a services-based web filtering application. The deal came about because we're a "core technology OEM" company looking for applications to plug into and Singlefin is a "managed protection services" company looking for competitive advantage and distinction from the field.
BK: How else is your technology used?
JM: Notiora is used in one of the largest business-to-business catalog companies. We haven't publicly announced the end customer's name but it was sold by the company Saqqara, whose a leader in services-oriented catalog management. We're also being used in the Chem/Biotech sector, again we haven't announced this partnership but we're currently going into Beta with a number of biotech firms. This application will enable scientists to use our technology to query public databases, private databases, the entire web and other sources to link together relationships between other scientific research, patient trials, chemical compounds and other pieces of a puzzle related to solving a scientific research question or advancing a medical cause.
BK: How have you come across your existing customers, and how big is the market for text classification?
JM: We've come across our existing customers by being "industry insiders", by word-of-mouth. As our mutual professional lives have been in somewhat concentrated areas we all have existing networks of professional friends and associates. In some cases our partners have come to us via a pass-along referral or via our direct knowledge of a problem within an industry where we know leaders.
BK: How long has Notiora been around, and what's the history of the company?
JM: Notiora has been around since 2001 and came together as a result of a core of us being together at Change.com, tackling the fairly difficult task of categorizing business-to-business SKUs into multi-thousand node taxonomies. Interesting problem and in the end it's just another text-classification application. Truth is there are a few companies addressing the same problem we are but we feel we take it a few steps further. Today we're a classic "inch wide, mile deep" company and to-date that's on purpose. Our current business model is to OEM into other's solution sets so to some degree it's in our best interest to offer amazing, razor-sharp tech rather than wide me-too applications.
BK: How is Notiora funded?
JM: We've never sought outside capital, intentionally. I've personally raised millions for start-ups from some of the best VC firms out there. I've been involved in public and private fund raising, mergers and acquisitions. I've been CEO of companies that went from zero to 90 employees in a few months time. That was then and this is now. When we started Notiora we first wanted to validate the technologies muscle and worth. We did that. Then we wanted to seek the right partners in vastly different industries. We're in the midst of doing that now. We may, or may not, seek funding in the future.
BK: What has been the hardest part of bootstrapping the business?
JM: Having been in positions of massive cash reserves in the past one can appreciate the advantages and challenges of bootstrapping a business. On the positive side, and this side outweighs the other, we control every aspect about our destiny. We've done deals with companies with larger customer bases, cash reserves, etc and yet we were the more cautious one in contract negotiations as to some degree when you're negotiating with a VC-funded company you're not really talking to the boss until you talk with the party that owns the majority share in their business and that's usually not the CEO. The challenge is "speed of business". For the same reason Sam Walton drove an old pickup truck to and from work as he was building Wal-Mart we're more conservative with money spent. Sometimes that's fine as you just save money but as we all know it also means that your business is built slower. This actually has been fine in the years past which is why we didn't seek funding, the market hasn't been strong for start-ups and only in the past 12 months have things started to normalize back.
BK: Finally, what's the next big step for Notiora?
JM: Land more OEM partners. The only thing as important as new sales is making those customers satisfied. We're razor-focused on those two things.