I love how in the past year or two, more and more book related startups have popped up. Last week I had the pleasure of speaking with the co-founder of Parakweet, Ramesh Haridas. Parakweet is a company that extracts valuable information from different social media streams and makes that data accessible and understandable. One way to do this is through BookVibe, a site that offers accurate book recommendations to users by extracting data from their Twitter handle.
BookVibe uses a few different tools for making book recommendations. The first way is to show a user a recommendations feed, once they sign up for the free service with their Twitter handle. After that, the site sends a weekly email with the top six book recommendations tailored to the user’s Twitter account.
The recommendations feed is very clean and easy to understand. Each book recommended has an image of the cover, and if you click on the cover, the site takes you to a page with a link to buy the book on Amazon, a description of the book, a graph to show how many mentions it has on Twitter, and a list of Tweets where the book was discussed. There is also a five-star rating system that shows social sentiment, and a button that tells you whether social buzz around the book is low, medium, or high.
Recommendations are based on a variety of signals, including how closely a user is affiliated with another person and how closely a user is affiliated with a particular book genre. BookVibe factors in the types of people users follow, what’s in their self-reported bio, and the profiles of the people they follow.
“So if you say you’re interested in young adult fiction, we would weigh books in that category higher up,” Ramesh said. “And if we thought there was a strong affinity with a person tweeting, we would weigh up that.”
This means that people with strong connections, though they may have different tastes in books, will still be able to provide interesting book recommendations. Todd Barish, who works with Parakweet and BookVibe, explained it further:
“BookVibe has tackled an extremely complex problem and uses natural language processing to cull through your tweets and make smart, targeted suggestions based on the data. For instance, if I tweeted that a particular book is very boring and not worth the read, it would not make sense for a book recommendation engine to suggest that book. BookVibe technology is set up to recognize the words in a tweet and know if it’s worth recommending or not. Another example would be if I said, ‘I am inside reading The 13th Cycle because the weather outside is lousy,’ BookVibe will know that lousy refers to weather and not the book so it may show up in someone’s recommendations.”
BookVibe also has a section on the site reserved for industry leaders. For example, Ramesh said that Bill Gates is influential in the tech sector, so any tech book he recommends would be ranked high.
Currently, Ramesh said they are manually choosing industry leaders and they factor in influence on Twitter, Google, and other criteria.
The BookVibe team is fairly small, with only six people working on it full time and three people part-time. But they only launched less than two months ago, and they are already getting great feedback. Many people use their “thank you” feature, which allows them to send a tweet to thank someone for their recommendation.
“People think these are great recommendations,” Ramesh said. “We have personally bought a lot of books that have been recommended to us. The conversions on the emails we’ve been sending have been very, very high […] So overall people seem to be very happy with the quality of recommendations.”
In the future, Ramesh said BookVibe will add recommendations from Facebook, Tumblr, and Instagram. There may also be support for Goodreads and LibraryThing.
“Facebook is next on our roadmap,” Ramesh said. Public likes are easy to extract, but pulling updates from a user’s Facebook stream can be tricky and inconsistent, he said.
BookVibe is built to be easier to use than a Facebook app like Bookscout, which recommends books to read based on a user’s likes. Some apps require a lot of time spent rating and adding books to shelves, Ramesh said, which can be a high barrier. But many people naturally post updates on Twitter and Facebook, he said.
“So we thought it made a lot more sense to tap into that and not try and create another social network for people to remember to go to.”
He said BookVibe has about 100,000 updates posted per day about books on Twitter, and he expects that number to increase to 300,000 in the next few weeks.
There is another similar site, called BookRX. It was developed by Northwestern and also analyzes a user’s Twitter feed, but personally I think BookVibe’s recommendations are more sophisticated.
BookVibe will always be free to use, Ramesh said. The main company, Parakweet, is about providing analytics to publishers and to authors. Currently, he said they are running pilots with a large retailer and a large publisher. But the company is also looking to work with self-published authors, at a price affordable for individuals.
“Authors are very interested in this too, to help them promote their books better, figure out who the influencers are on Twitter that speak about their books, and to better engage with these influencers,” he said.
Eventually, authors will be able to log in to a dashboard and see the books they want to track, Ramesh said. They can also compare books written by competitors or books in certain genres, as well as get analytics on the discussion around those books on Twitter. This will include the sentiment on the books and how quickly that sentiment is changing.
Ramesh said he plans to help authors and publishers with several different areas. The first is engagement optimization.
“Here we help authors identify the most high value users on Twitter to target,” he said. “So this would include ranking of an author’s followers by influence.”
Influence would be measured by engagement, audience size, and number of retweets, he said. “And also we could enable authors to similarly rank other authors and blogs, publications, etc. […] So we can figure out the influence of the followers of other blogs, other authors, and we can expand your own reach.
The second area will help authors better understand the profiles of their followers. Metrics would include the number of books in a follower’s stream, the number of books a follower purchase, their e-reader or tablet ownership, and the top genres and titles of books discussed, Ramesh said.
The third area will help authors promote their work, by enabling them to identify specific users or groups of users to target for promotion, Ramesh said
“So, for example, the ability to identify a group of users who follow author A and author B, and who are interested in young adult fiction,” he said
Publishers can also use these tools to better understand their audiences and promote more effectively, he said
“Today most of these publishers get very little data,” he said. “They don’t know who their customers are. Amazon gets all the data. Publishers get very little data.
Parakweet will roll out a beta in the next few weeks for authors to try out, Ramesh said. In the meantime, check out my affinity page on BookVibe. I don’t interact as much on Twitter as I should, but it shows my most influential friends (to me) and my book category interests (though to be honest, some don’t quite fit but that’s probably due to my eclectic book reviews and inconsistent use of Twitter). Soon everyone on BookVibe will be easily able to see their own affinity pages, and it will allow them to downgrade categories or influencers if BookVibe’s recommendations are off
Parakweet also offers another service for recommending movies, called Trendfinder.
“I prefer books,” Ramesh said. “I find that a lot more relaxing, to read a book […] I really like reading fiction.”
Unfortunately, like many entrepreneurs, Ramesh said with the launch of Parakweet he doesn’t have as much time as he’d like to read books. “When I walk in a café and I see people reading books, I envy them,” he said. “But I’m helping other people discover what books to read now.”
[…] to help promote relevant content to the sentiment. An example of a site that uses sentiment is BookVibe, a book recommendation system that uses […]