Accurate analysis of social data is becoming increasingly crucial for companies to create significant value for consumers. Transforming the relevant data into actionable intelligence still remains a big challenge due to the high-frequency nature of social data. In this complex scenario, where extracting precise content can be a difficult task, Bangalore-based Big Data startup Frrole helps media companies and brands discover insightful information from Twitter conversations.
Frrole’s Social Intelligence platform is powered by a semantic engine with the capability to mine social data in real-time. The firm owes its origin to an earlier travel product that the founders were developing called Travelomy, which was supposed to provide real-time updates to a traveller about the happenings in various city, with data from Twitter. This made them realise that they can provide real-time information about other categories as well, including news, sports, business and technology. This is how the Frrole social local newspaper Frrole News was born in April 2012.
“We worked on this idea for one year before we realised that we can use the underlying technology of Frrole News to create products for enterprise customers, especially in the media industry. We did initial experiments with a couple of media customers and saw phenomenal response. In the last three to four months, we have grown to more than 15 customers, which include almost all major media companies in India,” said Nishith Sharma, Co-founder, Frrole.
Frrole recently raised US$2,45,000 from a mix of angels and entities, both from India and US. Funding was generated from Sharad Sharma, former Head, Yahoo! R&D Centre in India; Rajan Anandan, MD, Google India; Manav Garg, Founder, Eka Software; Atul Singh, President APAC, Coca-Cola; Sunil Kalra, independent investor; Bhupen Shah, Co-founder, SlingMedia (which was bought by EchoStar) as angels; and the entity India Internet Group (IIG).
The startup raised this capital on letsventure.com, a crowd funding platform that helps Indian startups raise funds from accredited investors across the globe, signalling the growth of crowdfunding platforms in India.
Twitter the gold mine
Frrole is trying to solve the problem of signal to noise ratio, which is the single largest problem with social data. Typically, data analysis can be broken down across four levels namely, statistical, linguistic, semantic and contextual. Majority of data analytics products have an ability to answer questions of statistical or linguistic nature, but not the latter two. Frrole claims to be able to leverage semantic and contextual analysis by breaking down each tweet into nouns, adjectives and verbs, and understanding the context around it. It also understands whether the noun is a person, place, sports team,etc.
“Our major offering is very precise, filtered, customised data streams suited to our clients’ very specific needs. Our customers can use these streams to extract curated, precise content related to topics, locations, products, people or brands, and integrate it within their applications. An example is a live Twitter debate on Indian TV news channel, Headlines Today, where winner of the debate show is decided based on analysis of people’s reaction on Twitter. Our data streams can also be plugged into web applications easily to improve user engagement and interaction by brands, TV shows and campaigns,” explained Sharma.
Customers can also derive insights from social conversations, for example to determine who is the most talked about celebrity in India, which are the most buzzing products from each city and who are the thought leaders across product categories.
Elections and social prominence
Sharma is thankful to the current General Elections in India, because the importance of social data came out in the open very blatantly, and Frrole’s offering got its initial set of customers and appreciation without waiting for too long. Along with Headlines Today, it also powers social data for the show Newshour on the English news channel, Times Now and has built election hubs for media companies such as Times of India, Indian Express and MSN to track and display social buzz around elections. It is in process of developing elections hub for two other major media companies. The news app Newshunt also uses Frrole’s data streams to power its elections section.
Access to elevated Twitter data
Twitter is growing at a very fast pace and Frrole is currently processing just two to three per cent of the microblogging service data. Still Frrole’s Social Intelligence platform analyses almost half a billion data units every month. “From having access of only the freely available data six months ago, we are now the track partners of Twitter and have access to elevated data. We plan to work extensively with Twitter data, and are working on many different kinds of use-cases. Scaling up infrastructure to build robust systems, which are capable of handling larger data access, is a challenge and we’re hiring best industry talent to keep pace with our growth,” Sharma said.
“Technology is our topmost priority, and our major focus has always been to ensure that we are always ready to scale to support bigger, better, more extensive use-cases,” he added.
The pricing of Frrole is based on the number of data streams a customer is using for a specific application/programme/campaign. A data stream is a well-defined and specific collection of tweets/insights pertaining to any hashtag/location/category/keyword, etc. Currently, Frrole charges US$100 per data stream per month, and a typical brand generally uses 8-10 data streams, making the service cost about US$800 to US$1000 per month.
It is important for Frrole to go beyond Twitter-based analytics to compete globally. It is planning to build algorithms targeted beyond media and brands as well. It has started working on a product called ‘Behavior-as-a-Service’. This will be useful especially for retail and e-commerce companies, who want to target very specific interests of their customers. The product will help these clients understand what their customers like, dislike, do, plan to do, etc. almost in real-time, helping them recommend right products to their customers exactly when they need them.
“At the moment, we’ve decided to work and be really strong on Twitter data first. We chose Twitter over other networks, before because it is public data set, unlike other networks which have most of the data as private. It is almost a universally preferred choice for discussions about TV shows, sports tournaments, general events and news consumption, and this suits our use-cases perfectly. Having said that, we do plan to integrate other networks in a few months from now,” Sharma added.
Expanding to Southeast Asia
According to the Co-founder, Southeast Asia is one of the most active social media regions in the world, and there is huge potential to integrate social into the products/brands. Though all its clients are Indian so far, the firm has started exploring Southeast Asia. “We’ve talked to some major media companies in Singapore and Indonesia, and are going to meet some of them next month. We are not hiring any team there yet, but we plan to make regular client visits to SEA as we get more active and grow our operations,” concluded Sharma.