Flipora is the fastest-growing social surfing service you’ve never heard of (interview)

wow

Never heard of Flipora? Join the club.

But you’re likely to hear more about this new service for discovering and sharing online content, as it’s growing by 150,000 users each week — and just today added significant new features to make social surfing recommendations to help users discover and enjoy new web content and sites.

Formerly called InfoAxe, Flipora grew from a service that helped users simply find sites they’ve already visited. You’ve likely experienced this problem: You found a great site, but can’t find it later. And you can’t find the data you know is available on that page. InfoAxe helped you search your history.

The company rebranded to Flipora late last year, and today it announced a significant new direction: social discovery of new web content. VentureBeat spoke to Jonathan Siddharth, one of the cofounders.

On the surface, the new Flipora features are much like StumbleUpon. Users select interests from 57 different topic areas and then surf the web within the Flipora service on those topic areas. Content inside Flipora is based on what others users find interesting.

Instead of “stumbling” to find new content they like, users “flip” web pages:

But there’s a key difference between StumbleUpon and Flipora…and indeed Twitter, Facebook, or any other service that web surfers use to find new content. Instead of articles or videos being entered via contributor’s choices — a post, a tweet, or a stumble — Flipora lives inside users’ browsers and shares an aggregated view of all users surfing habits, which builds a vast array of website popularity data.

Flipora announced the change to their own users like this:

We’ve been making it easy for our users to go back to where they’ve been in the past…what we’re announcing today is a way for users to discover where to go next ….

The key is in the web-browsing history that originally was only for your own personal consumption, to help you find previously viewed pages and information. Now that data is collected, uploaded, and stored with others users’ data. And similar to an Amazon.com people-who-liked-X-product-also-liked-Y way, it generates new suggested sites for all users.

And that’s lot of data. Flipora currently has 8 million users. To operate the social-discovery engine, the company indexes 20 million web pages daily — an amount equivalent to six times the size of Wikipedia.

The key questions, of course, are the same as those Facebook faced when it enabled “frictionless sharing“: What happens to privacy when everything is shared? And similarly, what happens to quality when there is no discrimination between what is shared and what is unshared?

VentureBeat asked Siddharth and his co-CEO, and Vijay Krishnan these a few questions:

VentureBeat:
You seem to be almost half the size of StumbleUpon without near its attention. How have you grown so large so quietly?

Siddharth:
Correct. Flipora, previously Infoaxe, is probably the largest consumer Internet company that no one in Silicon Valley has heard of. [Smiles] We are a small team relative to the size of our user base, and we’ve just not had enough time to focus on anything besides working on the product and scaling for the growth we’re seeing.

Until we raised our $3 million round in late 2010, our full-time team in Silicon Valley was just Vijay and myself. We had some really great engineers working remotely. Once we raised the round, we hired them full time. In hindsight, we should have probably done a better job of getting the word out in Silicon Valley, since it helps greatly with hiring. Additionally, our user growth is also extremely spread out across the world, rather than being confined to the U.S. alone. We have users from over 200 countries worldwide.

Our growth has been largely organic. Flipora’s web history search engine is a unique product with very few competitors who are able to do it at scale.

We’ve been beta testing the discovery feature for a while and that has been really good at drawing new users. We offer great recommendations and give users easy tools to share them on Facebook and Twitter. People who click through to the recommended webpage are shown other related recommendations that keeps them engaged, in a manner similar to YouTube where users who come to see a video [and] end up clicking through and consuming related videos shown on the side.

VentureBeat:
Flipora has enjoyed growth in 2012. Why is it happening now?.

Siddharth:
We did a massive redesign of the core web history product in late 2011, vastly improving its ease of use and beta testing the discovery feature. Both of these are the key contributors to growth. Flipora’s web history product was already a uniquely differentiated product with very few alternatives. The redesign and product enhancements greatly helped usage and growth.

We crossed 8 million users last week and are growing by 25,000 users/day. We see users using the beta version of the discovery product flipping as many as 300 pages a day which is extremely encouraging to us. In our beta tests, our unique product innovations like the side bar that lets you preview the recommendations coming next while also allowing a user to easily skip over items that are not interesting greatly increased usage.

Our recommendation engine quality is also going to keep improving. What we have live now is just version one. We are working on big updates to our algorithms that will drastically improve recommendation quality in the coming months. The product does get better the more you use it, as it learns your interests from use.

VentureBeat:
How does Flipora address privacy concerns.

Siddharth:
The answer is transparency and offering easy ways for users to opt-out of sharing any data they are not comfortable sharing. We take privacy very seriously at Flipora and have had a very meticulous approach to this right from day 1.

1. A user’s web history is private by default and is never shared with anyone. We don’t share data with third parties. You explicitly choose what you want to share. The fact that right from the time a user signs up, we are upfront about being a web-history search engine, and the fact that the user experience with our product is most effective when they save their browsing history with us, has greatly enabled users to be comfortable with the data that we collect.

2. We offer a “record” button that can be turned “off” if you want to go off the record.

3. We offer a private “block list” of sites that you can set up. Page views on those sites are not recorded even if you accidentally had record on. More important, this “block list” resides on your browser. Even Flipora does not know what the block list is.

4. The recommendation engine only uses aggregate, anonymized browsing data.

5. In the last decade, Internet users as a whole have fundamentally got more comfortable with the idea of some of their information being accessible to consumer Internet companies, which has indeed powered the growth of several billion dollar consumer internet businesses.

Web-based e-mail providers have made Internet users much more comfortable with having their private e-mail stored on remote servers, rather than on their local machines. Social networks like Facebook and Twitter, personal finance management services like mint.com…have made users much more comfortable with the idea of sharing their personal information with consumer Internet companies, as long as the data is secure and the product offers them compelling value. These, coupled with the fact that we are up front about what data you collect and our clear privacy policy and terms of service, have gone a long way in alleviating privacy concerns regarding our product.

VentureBeat:
What about reality bubble issues? Can or do you introduce serendipity into the pages you show people?

Krishnan:
We try to balance exploration and exploitation in our recommendations. We recommend websites that are related to what the user is in the mood for at the moment while still showing a small number of recommendations from other topics. For example, if we think the user is enjoying recommendations from the “cars” topic more, the user receives more recommendations from “cars” while still seeing recommendations from other topics in case she wants to explore other interests.

Algorithmic discovery has greater potential to widen a user’s horizons than social discovery, discovery based on what websites your friends share on Facebook or Twitter. With social discovery, there is a lot less serendipity since your friends are unlikely to share certain types of content. Your friends typically would not share content on Facebook or Twitter that they think would be uninteresting to their friends, politically sensitive, etc. So you see a much narrower spread of topics.

VentureBeat:
What are your long-term goals?

Krishnan:
Web history search will be the platform that helps us build an amazing Discovery Engine that truly understands a user’s interests, both explicitly declared and implicitly learned. We believe Discovery to be the natural evolution of web search itself. Web search works only when you know what you’re looking for and you know it exists. With our Discovery Engine, we can automatically push the most relevant websites to a user at the right time when she’s in the mood for it. With web history search, Flipora makes it easy for users to keep track of the great websites they discover and the data generated by this will continue to improve the discovery engine.

We think, just like Web search, Discovery will be an integral part of the web browser, answering the question of “where should I go next?” With the data advantage that we have — attention data from web browsing history — and the technology to mine that data, we are superexcited about some unique innovations we can offer in the discovery space. The discovery space is still very nascent making it ripe for disruption.

We are at 8 million users now and expect to double that in a year.

 VentureBeat:
Thank you for your time!

Image credit: ShutterStock