I won’t be wrong if I say that Google is totally irrelevant when it comes to discovering great quality content. Thanks to SEO, there is very little juice left for those who are particular about reading/discovering quality content.
Content discovery still remains a big hurdle for most companies and Pugmarks, Bangalore based contextual discovery engine has launched a new avatar combining all of its parent company, i.e. Insieve’s capabilities – right from Dhiti’s recommendation engine to contextual engine (based on twitter shares) they launched earlier.
How Does Pugmarks work?
Pugmarks.me is contextual discovery inside the browser. The Bangalore based startup has launched smart browser (for Android and Chome) that understands the kind of pages you are reading, and brings high quality recommendations, without you going to Google.
Through events and callbacks on the browser, Pugmarks finds out the address of the page you are visiting. On a page that’s deemed worthy of recommendations, Pugmarks will send the URL to its servers – through https (read the privacy note). These recommendations are served via Reading Cards (and also via ribbons on the page).
Ojas Ventures funded Pugmarks delivers great quality content and if you are somebody who is researching a specific topic, this comes like a god-send utility.
What Goes Under the Pugmarks Hood?
“We are a search engine for the real time web, under the hood.” [Bharath, Pugmarks Cofounder]
On an average, Pugmarks gets about 10-30% CTR (click through rate) on its recommendations – which is extremely high (though the current traction is limited to early adopter audience). The startup is crawling about 350,000 sources on Twitter and LinkedIn all the time, and have indexed 20 million articles now – growing at 200,000 every day.
What about the Android play? The Google Now Game?
Cofounder Bharath is quite bullish that Pugmarks will give a tough competition to Google Now in the next few months.
“Google Now mostly covers structured context. Eg: Tickets in your email. Weather in your location. Places near by. etc. They are still behind on taking signals from the browser.
At Pugmarks, we’ve done a lot of work mining unstructured context – converting URLs you visit on your browser into topics worthy of reading cards.
i) Observing the difference between your navigational searches, and information searches. Eg: If you search for netflix on google, mainly to click on the first result and go to netflix.com, your intent is to watch a movie there – and not to get news about netflix. However, if you search on “gravity” and go to the IMDB link, your intent is to know about gravity the movie.
ii) Mining the most important aspects of all kinds of pages – be it eCommerce pages, articles, etc.
iii) We also recognize your long term interests, versus immediate spikes in interests – and respond appropriately. Eg: If you have a long term interest, you dont want reference material, but care about news. If you have a short term spike on a topic that’s unfamiliar, you are open to reference material. “
Short Qna/NextBigWhat: What’s your vision behind the product?
Bharath: The future of search, and especially mobile search is not about “searching”. Typing is hard, and prone with typos. Iteration is even harder. At the same time, there’s more information available about what users have just done and what they care about. Our vision is to make search – not feel like search, but help people find good content, all the time. We call this “anticipatory content discovery”.
People often do their work across devices. You may read something on your laptop browser, and then run to a meeting. Our recommendations carry context across the devices you use.
In the future, we want to cover as many actions and events you go through – browsing, searching, location, calendar, even TV, and recommend content to you. All of this is safe and opt-in.
Feedback/Suggestions to the team:
I have been playing with the product (chrome extension specifically) for over 2 weeks and must say, am blown away with the quality of recommendations. However, the challenge I face is when I really start respecting the smart browser tab as my source of truth (and content) – the card display takes away the experience (from a UI perspective) – the vertical placed cards expects me to scroll through and find useful content.
And that brigns me to the next suggestion, i.e. no way to remove specific content.
The team needs to understand that not every piece of content that one is browsing through is intent based reading – sometimes, I just checkout random content on soccer by reading about it on Facebook – but that doesn’t mean that I am really interested in the game. What am I trying to say? There has to be an easy way to mute the extension (temporarily) and in the process, train the algorithm.
Remember that the product’s biggest challenge is to ensure that the user keeps it running on the browser and if there are more occurrences of irrelevant suggestions, it stands a chance to be simply removed from the browser (with no looking back).
Unclear Call to action: I am a firm believer in curation and one of the features I’d really like the team to work is to build ‘call to action’.
So I found a great content, what do I do with it? Add to Evernote/Read it Later? This I believe, will open up the revenue stream for Pugmarks (think Curata).
TL;DR: Startups like Pugmarks are the ones who are pushing the limits of realtime web and I just hope/believe that we have had more early adopters in the country who can really help such companies with actionable and fast feedback.