Great ideas are born when there's a pressing demand. The idea to create a more personalised and effective Web search occurred to Jonathan Siddharth and Vijay Krishnan in 2007, during their academic rigours of their Masters at the Stanford University in the United States. “We found it tough to keep going back and forth to the Web, trying to remember the one great article or web page we saw,” recalls Siddharth. Their solution was Infoaxe ( >www.infoaxe.com ), an add-on to web browsers such as the Internet Explorer and the Mozilla Firefox (an extension of Google Chrome is in the works). In simple terms, it plays the role of a ‘memory' by keeping a record of a user's complete browsing history. So no need to bookmark, tag or try to think hard about that one great website you saw a week ago. Users can download the Infoaxe extension for their browsers and forget about having to remember urls and websites.
Re-finding queries
A 2007 study by a group of students from MIT (Massachusetts Institute of Technology) states that nearly 40 per cent of all search queries are re-finding queries. By banking on the entire web history to help one's search, the relevance of the results improves dramatically, the Silicon Valley entrepreneurs reason.
There is some credibility to their claim: Infoaxe already has more than seven million unique users. It is currently growing at the rate of a million users every two months.
User recommendations and glowing reviews by reputed technology blogs are helping. Added to that, the company has some of the biggest names in Silicon Valley on board as advisors. On completion of their recruitment drive in India, the duo took time off to talk to MetroPlus, Chennai. It was also a homecoming of sorts for the Chennai boys. Jonathan Siddharth is a graduate of Sri Venkateswara College of Engineering and belongs to Chennai. Vijay Krishnan, though an alumnus of IIT Bombay, has also stayed some years in the city and has roots here.
As luck would have it, the duo ended up as room-mates during their Stanford University Masters programme. Siddharth graduated with distinction in Masters in Research, bagging the Christopher Stephenson Memorial Award for graduate research for best master thesis at Stanford. His research areas included information retrieval, web search and search personalisation. Vijay received his Masters in Computer Science and his research focussed on data/text/web mining.
Though they passed out in 2007 and Infoaxe was incorporated as a company, it was officially launched in late 2008 at a time when they raised $9,00,000 as a seed fund. The environment at Stanford and later Silicon Valley was ideal for setting up a 15-people office at Sunnyvale, California. In August, last year, they raised an additional $3 million by way of venture capital and angel funding.
Goodbye, bookmarks
At the time of its launch, though drastically different in its approach, the closest similar service was offered by online bookmark management services such as Delicious. “The method had become tedious for most users,” says Siddharth. “It is taxing to create tags and label everything one wants. Besides, search options in bookmarking services were minimal and ineffective.”
Around the same period as their launch, social networks such as Facebook and Twitter were gaining popularity with users as tools to curate content. Recommendations through links on the newsfeed and re-tweets were becoming ways to discover new content. But in that too there was a hitch. “Only explicit web behaviour gets reflected in social networks. There might still be some interesting content people browse but won't want to show,” points out Siddharth.
Twitter and Facebook are estimated to receive much less than 5 million distinct webpage hyperlinks in a day through tweets and the facebook sharing functionality (since a lot of the content sharing involves retweeting and multiple users re-sharing content posted by their friends, causing a relative lack of diversity in the nature of the content) and Infoaxe bookmarks and indexes about 20 million distinct webpage hyperlinks everyday.
But more is not necessarily good. The entrepreneurs say this is where their algorithms come in. Using more than 50 signals, Infoaxe curates content based on the data memory of the user. “It will know what sort of content the user will like,” says Vijay.
Siddharth and Vijay are taking the service to the next level — a Web discovery engine called Flipora that will, in addition to tracking and storing one's web history, make recommendations based on the interests not just of the user but also a community of similar-minded users.
“All the search that happens through the search engine is based on the pull paradigm,” explains Siddarth. “You know what you want and you keep pulling information from the Web by typing in the search words. But with Flipora, we apply the push paradigm. The algorithms know the user's browsing history and will keep pushing content taking into account his likes and dislikes.”
A crude analogy for Flipora is the way television functions. The user will be able to flip through content without bothering much about typing into the address bar of the browser. Based on the Web history and preferences, the discovery engine will keep pushing out content tailored to the user. This way there is a considerable likelihood that one runs into some fantastic lesser-known content from areas of one's interest.
Infoaxe stores personal Web history on its servers. In the wake of the ongoing debate on privacy, it is important to be aware of this. Siddharth, while assuring that the private data of the users remains safe, also highlights the obvious gains. “Every service has a cost. It is up to the users to decide whether the gains exceed the costs.”
