|
Yactraq Online Inc. |
|
|
July 22, 2013 Issue |
||
|
The Most Powerful Name In Corporate News and Information |
||
|
Focused on Monetizing Video, Through Revolutionary Cloud Software Yactraq Online Inc. Extracts Meaning from Audiovisual Content for Ad Targeting, Content Recommendation and User Behavior Analysis |
||
|
Yactraq Online Inc. is
focused on helping publishers monetize video. They do this by extracting
meaning from the audio track of the audiovisual content to generate rich
metadata. Customers then use Yactraq metadata to target ads, recommend
content or analyze user behavior. Yactraq cloud software is based on its
patent pending Speech2Topics technology. Market traction includes millions
of minutes of audiovisual production content processed, and early revenue
from global customers that include Top20 VC backed businesses.
Founder &
CEO
Before migrating to North America in 2002, Jeh played a key role in incubating Customer Asset, an India based IT infrastructure services venture that evolved into the $100 million, Firstsource Solutions (Bombay Stock Exchange – BOM:532809). Jeh sourced the acquisition of Customer Leverage, the software solution on which Firstsource offers its customer management services, and the team that developed it, into Customer Asset. Prior to that while working on Delsys, his first startup, Jeh sold the first-ever ethernet switch in India (Kalpana-Cisco).
Jeh has a Masters in Management Studies (Finance) & a Bachelors in Engineering (Electronics) from Bombay University. He is also a lifelong manned spaceflight enthusiast and believer in technology innovation as the key driver of corporate and economic growth. |
Online Audio & Video Speech Recognition
Yactraq Online Inc.
1400- 1055 W Hastings St
604-341-6217
|
|
|
Interview
conducted by: Lynn Fosse, Senior Editor, CEOCFO Magazine, Published – July
22, 2013 CEOCFO: Mr. Daruvala, what is the concept at Yactraq Online Inc.? Mr. Daruvala: What we are all about is video monetization. We help our customers monetize their video. The way we do it, is through a lot of powerful technology related to speech recognition, natural language processing, and all kinds of semantic extraction. At a high level, we extract meaning from the audiovisual content.
Essentially it is like an interface where on the input side, audiovisual content goes in, and on the output side, we send out a file containing the metadata that describes the audiovisual content.
Then our customers are essentially using that to drive more and better revenue by more effectively targeting ads, profiling users based on their video watching behavior, or perhaps recommending content to them. We have seen many OTT (Apple TV type) products that recommend content to users. But they do that imperfectly often missing the mark in terms of what the user is really looking for. And we are correcting that. The vision of the future of search is that it is going to be essentially content discovery and recommendation, where the machine proactively suggests things to you. And that's where our technology provides the essential recipe for success. We can also use our software for video search or general analytics on the entire audiovisual content.
Additionally, we can adapt to our customer's deployment needs. One very simple way to visualize our service is that either it can be deployed in the cloud or it can be used as on-premise software.
CEOCFO: Would you explain a little bit more about the technology? Mr. Daruvala: At a high level, what we are doing is first passing the content through a process of speech recognition, and then we pass it through a second layer of what we call semantic extraction, which is the extraction of meaning in terms of topic metadata. Our output file contains information by the minute, and each minute will contain say five or twenty topics that occurred during the speech of that minute. For each topic, we will provide some numbers saying that a topic was more dominant than the other topic, because we have topic dominance metrics and similar metrics for sentiment- whether it was a positive sentiment for a topic or not. Essentially, we are developing a semantic signature per minute within that video object, and if it happened to be a five or 10 minute, or two-hour video object, there would be a semantic signature for the entire object itself. We have somewhat of a B2D2B business, and what that means is business to developer to business. We are currently selling a solution to a product developer, who can then develop a very rich and meaningful coating or in other words applications using our metadata. In that sense, you can view our product as a very horizontal product. We enable many things, but we are increasingly working on a game plan to develop some of our own vertical applications. We already have some rudiments of an ad-targeting platform and so on.
CEOCFO: What were some of the challenges in putting this technology together? Mr. Daruvala: The technical challenge in particular is that speech recognition is a very difficult subject. You have to deal with content that is not just very pristine such as the kind we have on this phone call, but it could be a YouTube video that is at a trade show floor with a lot of background noise and music. If it is a produced video clip that has a lot of background music or music interspersed, those things are generally very challenging for speech recognition systems. Those are the hardest problems to solve.
CEOCFO: Who is using your technology today? Mr. Daruvala: In terms of what we can declare publicly, I will talk about where we are and the status of the company as well as the product. We have been in the market for about a year. You could view us as a very late beta stage product by now, or you could call us a true production product. We have indexed several millions of minutes for one very interesting customer whom I can tell you all about, who happens to be Michael Robertson of DAR.FM. In the past, Michael founded MP3.com in the late 90s, so has quite a story as an entrepreneur and is quite an icon in many spheres. Today, DAR.FM is thought of as a TiVo for internet radio. We work with them in terms of helping them contextualize ads, helping users navigate through long-form content and 2-3 hour long programs. This example and others, have helped us amass a huge amount of experience with broadcast quality content. We have about four or five customers signed up, and most of them are revenue producing customers.
But we still view ourselves a pre-revenue company because our revenue is in the thousands of dollars, while our opportunity is in the hundreds of millions once we get really going. What does count at this stage, is the quality of the customers we are already serving. We have several customers that are funded by global top 10 VC’s. These companies are building applications and doing work in areas such as face recognition. They might be creating content recommendation engines or video content exchanges. Some are NASDAQ listed companies which already use social media analytics on Facebook and Twitter text, to extract brand sentiment. Now they want to do the same on YouTube video and our software provides them with the perfect answer to this exploding need. In other words there are many impressive names that we are working with, on very novel solutions, and we managed to generate revenue from it even if you could assimilate that to some kind of experimentation.
Some of our customers are testing our system and performance under heavy load conditions for robustness and accuracy. And we’ve done particularly well. Most of the time, we have been working with many of them for a period of months and they are helping us improve our product. I think we have a pretty amazing product by now.
Another important consideration is the way we have chosen to go to market. Although the end-beneficiaries of our solutions are business to consumer type video publishers, , we have seen that it is more effective to work with companies that serve them, such as video infrastructure companies that are either building video platforms or advertising technologies, and so are supplying data analytics etc.
CEOCFO: Could you tell us about how the use of video is growing, and what the opportunity is for Yactraq? Mr. Daruvala: We all know that video is exploding, and we are talking about billions of minutes and billions of hours of video very soon in the next four or five years. It is equally well known that this is an industry that has had and continues to have issues with monetization when it comes to video internet in particular. One of the major issues is that computers have been designed and continue to essentially be optimally designed for processing numbers and text. That is what computers do incredibly well, so when you throw foreign objects such as audiovisual content at machines, from their perspective it is a black box. The inside is dark content, and they cannot particularly see inside of it the way we as humans can use our eyes and ears to do that. From the computer’s perspective, all it knows is what the external label tells me. That external label might be stuff that a ser uploaded to their YouTube in terms of a subject line, or some data that Google/YouTube asked you to put in. It could also be some sort of an editorial paragraph that says it is a movie and who is starring in it. That is all that the machine ever knows about it, so that is what the basic limitation is. If we can help computers using artificial intelligence software to fully understand what the audiovisual content is all about and do this inexpensively enough that it can become a universal phenomenon that completely changes the nature of the video internet at a very fundamental level. It is our hypothesis that this would help solve many of the issues in terms of user experience, user engagement, and yes monetization.
CEOCFO: How are you reaching potential customers? Mr. Daruvala: We are a self-funded,, early stage startup, so within the financial constraints that we have, we use a combination of networking and I spend half my time in the Bay Area, which happens to be a strategic location. Our VP of Bizdev is based out of that area, and we try to get in front of as many investors and potential customers as we can to speak about our product at public forums and so on. We try to get as many interviews such as this as we can get, and that is basically what we do. It is a networking style approach combined with some corporate communications effort.
CEOCFO: You certainly have experience in the business community. What do you bring from previous ventures that is most helpful here as you are developing Yactraq? Mr. Daruvala: I will talk more about the team than merely myself, but I will start with me first. My story begins in India. I have a background in electronics and financial management, and I spent the first ten years in the 1990s building a network consulting company in India just as the global community was essentially arriving. There were many financial institutions, which were mostly American, that were landing at India’s doorstep in the early 90s, and we served those companies and did a lot of work for them. That was my first entrepreneurial venture, which ran for about ten years. The one achievement that I love to talk about there is that I had the good luck to sell the very first stand-alone Ethernet switch that was ever sold in India as part of that process. Toward the last few years of my career in India, I did some work in the late 90s tech boom, and we are talking about doing some boutique, high-tech investment banking work. It was actually a company comprised of me, another colleague, and a whiteboard with the word Newco written on it, and a direct descendant of that is now essentially a company called Firstsource. It is listed on the Bombay stock exchange, and they do business process outsourcing and are worth well over a hundred million dollars on the stock exchange. In 2002, I moved with my wife and kids to Vancouver, British Columbia. My first job was actually in Seattle, and in the last ten or so years that I have been here in North America, most of my time has been spent with Microsoft, T-Mobile, and Telus here in Vancouver, doing product management, business analysis, and functional systems analysis type of work. This was excellent preparation to work with people like my CTO, because I was grappling with some incredibly complex software systems of these large telephone companies and companies such as Microsoft. This is a good segway to talk about Lee Iverson, who is our CTO. Lee joined us some time in December of 2010, which is essentially the time that the concept had been formulated and we had just about started coding when he joined us. He is a Princeton undergrad in engineering, with a PhD in machine vision from McGill of Montreal. He spent about eight years at SRI in the Silicon Valley, essentially as a defense scientist doing work on machine vision. Finally, he moved back here to British Columbia because his family was here and he wanted to be closer to his parents. He took a gig at UBC teaching computer engineering, but after a couple of years decided that it was not for him. He just turned to entrepreneurship about a year or two before I met him. That first venture had not really gone anywhere, so he was available and that is when he joined us. That is his story, and he is a world-class computer scientist. He is an amazing guy, although the principle product vision has been all mine, all of the implementation is to Lee’s credit and it is extremely complex software development work that we have done. Our VP of Bizdev Francois Modarresse is a more recent entrant, he joined us in the last three or four months. He is based out of San Mateo, California. He has spent his earlier career as a big company guy with Phillips, and he has incredible experience of 25-30 years in video. He then morphed into an entrepreneur as he got involved with SkyStream Networks that was VC funded, and a couple of other entrepreneurial ventures, which were successful. He also ran product marketing for Dolby which provided him additional experience in mobile and online video He has a great all-around, deep insight into video infrastructure and consumer devices as an industry as well as an understanding of how startups work. That is essentially the team, but there are a couple of others who are doing some part time work. It is also worth mentioning our advisors. We have some truly outstanding advisors- recently, Eric Benhamou, who used to be Chairman and CEO at 3Com and Palm and is still chairman at Cypress Semiconductor, recently agreed and has formerly joined our board as of just yesterday. He is not even yet in all of our documentation, but we will add him very soon. Subhash Bal, who is a marketing specialist and has been CEO of 3Com and Synopsys India. Many years back he was also VP of Marketing at Excelan when Kanwal Rekhi was VP of Technology. Kanwal is something of an icon in the Indian tech community. They were then acquired by Novell and Subhash has had several other successful exits. We also have Randall Hounsell who is currently a VP at Comcast and who provides us a unique insight into the demand side of our business with both executive-level and state-of-the-art intelligence. This team is rounded up with Steven Forth who has worked many years with Michael Porters Monitor group up at Harvard, and he is a brilliant strategist. That rounds up what we have in terms of a team here.
CEOCFO: What is ahead for Yactraq?
Mr.
Daruvala:
We are at the point where we have built a really amazing
product. It has excellent characteristics in terms of accuracy, and more
recently, we have been able to scale it to be able to process hundreds and
potentially thousands of concurrent audiovisual streams. We are all set up
on Amazon Cloud infrastructure and it is all beautifully scalable, so it is
an outstanding, rugged product. I think we are at that point where in the
last few months we have begun to receive some serious recognition. In the
Fall and Winter of 2012, we were declared a CIX Top 20 Most Innovative in
Canada, and in May, we became a TiE50 winner. That has made a big
difference, and all of a sudden we find doors opening, we find ourselves
invited into a lot of Sand Hill road investor offices, and many things are
happening in that regard. We are getting into people’s doors, and that is
nice to have. Also, given that we have invested so much in developing such a
good product, it is quite intriguing that even in a pre revenue stage, some
of our potential customers see our product as so strategic that they are
suggesting investing. We could see the possibility of a very early exit for
Yactraq, which is almost a pre-revenue type exit, and that is quite
interesting when that happens, largely because of the technical value that
customers strategically see in terms of what we are doing. There are all
kinds of possibilities. We could either try to build this into a great
business over the next year, or there could be a possibility for an early
exit. |
||
|
“What we are all about is video monetization. We help our customers monetize their video. The way we do it, is through a lot of powerful technology related to speech recognition, natural language processing, and all kinds of semantic extraction. At a high level, we extract meaning from the audiovisual content.” - Jeh Daruvala |
||
|
|
||
|
Video Monetization, Yactraq Online Inc., Technology Companies, CEO Interviews 2013, Monetizing Video, Cloud Software, Extract Meaning from Audiovisual Content for Ad Targeting, Content Recommendation, User Behavior Analysis, Speech2Topics, Recent CEO Interviews, Technology Stock, interface on the input side where audiovisual content goes in, on the output side send out a file containing the metadata that describes the audiovisual content, Yactraq Online Inc. Press Releases, News, Companies looking for venture capital, Angel Investors, private companies looking for investors, technology companies seeking investors, video monetization companies needing investment capital |
||
