The song sounds familiar, but the words are missing. The listener thinks for a minute, trying to fit lyrics to the tune; she decides to hazard a guess. Her voice quivers as she sings into her cellphone; then, suddenly, a smile spreads across her face. The words are right. That’s one reward point earned.
The prompt at the end of the line gives her a few seconds to recover before launching into the next song. The exercise is repeated with redoubled enthusiasm. It makes for an odd sight; it also seems to be immense fun.
In the crowded world of mobile phone apps, Mobile Antakshari stands out. The game, created by Hexolabs, a Chennai-based firm, is one of India’s most popular apps, having attracted 110,000 customers within the first three months of its launch in May 2009.
Since then the app’s user base has been growing at 10% every quarter. And the game is, clearly, addictive—approximately 40% of its players are repeat customers. Mobile Antakshari, says Vijay Shekhar Sharma, chairman of One97 Communications, one of India’s largest mobile value-added services (VAS) companies, is an “excellent incarnation of the game” and makes “optimum use of resources”.
Mass appeal: Raja Manohar, managing director of Hexolabs Interactive, which created Mobile Antakshari, at his Chennai office. M. Lakshman/Mint
In a country where the Internet user base is growing sluggishly, even as the mobile user base has exploded, mobile apps may serve as the best bridge between the two. “In rural India, VAS and apps are fast taking the space of the Internet, access to which remains limited,” Sharma explains. “They’re easier to use, and they’re often available in vernacular languages.” The success of Mobile Antakshari is only one of the cases in point.
Incredible as it may sound, Mobile Antakshari happened largely by chance. It wasn’t what Raja Manohar, Hexolabs’ managing director, had in mind when he started his company in 2006, when he was still a student at the Indian Institute of Technology, Kanpur. At the time, Manohar had been focusing on creating educational apps, to help children spell or recognize prime numbers, for example.
Those apps were successful, but only moderately; Manohar knew that he’d need to create an app with a much wider appeal. What that would be, he wasn’t sure.
Then, one day, a maid working in the Hexolabs office asked Manohar what exactly was it that his firm did. This woman, who had studied till class V, spoke only a smattering of Tamil, and she needed her daughter’s help to make calls on her cellphone.
Also Read The Future of the Internet (Complete Series)
That blunt question set off a train of thought in Manohar’s mind. “Her only access to information seemed to be voice,” he says. “That’s when I realized that to reach out to the bulk of people, I’d need to venture into voice-based services.”
The choice, he claims, was narrow. Mass appeal in apps was limited to “astrology, Bollywood and cricket”. Astrology was not quite his area of expertise, and he wasn’t clear what kind of cricket app he could create. Bollywood, though: that might work.
Manohar recalls that it took a 10-member team—graduates of IIT and the National Institute of Design—5,500 lines of code, six months of work, and “between Rs 30 lakh and Rs 40 lakh” to create the app.
The heart of the game—the very first stumbling block—was the automatic speech recognition (ASR) technology to decode the players’ voices. Licences for these advanced technologies were prohibitively expensive, so Hexolabs tied up with OnMobile to use one of its ASR technologies.
Next came the “exhausting and funny” process of creating a database of songs for four languages—Hindi, Tamil, Telugu and Malayalam. Each language would have 1,500 songs, every one of which needed to have a “grammar” written for it, which would give the system points of comparison with what players were singing.
The complexities only increased. “We’d noticed,” Manohar says, “that when players weren’t sure of the song, they’d start in a wavering, slightly husky voice.” The ASR would have to compensate for these vocal quirks. Then came the testing, for which, says Manohar, “friends and a few freelancers” were roped in.
Mobile Antakshari can be played either against the computer or against friends. In the former, a player is presented with four modes of play. All four modes expect players to recognize a song and sing it back, although the clues in each differ.
The mutliplayer mode, for “Antakshari Ustads”, allows players to key in the telephone numbers of their friends, inviting them to join in. These friends-turned-competitors now buzz in every time they recognize the song, to get first shot at singing the lyrics. Winners redeem their points for prizes such as caller tunes or more apps.
The work on Mobile Antakshari continues apace. Every quarter, around 100-150 new songs are added in each language. A massive back-end system logs every single song a particular player has heard, so that songs aren’t repeated. The system also logs language preferences: A player who plays in Hindi three times is taken directly to the Hindi version the fourth time around.
The success of the app, however, hasn’t translated into financial returns, which leaves Manohar a little bitter. “The irony in app development is that the creator gets 10% of the revenue while everyone else, including mobile service providers, get 90%,” he says. “There really isn’t very much money to be made from content.”
Part of the problem, he thinks, is the fragmented nature of the Indian market. “In countries like the US, where handset markets are a lot more homogenous, it’s possible to survive by creating apps for phones like the iPhone,” he says. “But here, we have to customize apps for a range of handsets, a time-consuming and expensive process.”
Sharma agrees that the initial marketing for any app is tricky, especially for small companies, but the initial struggle, he believes, will pay off.
The silver lining for Hexolabs is that two more domestic mobile operators will soon offer Mobile Antakshari; operators in Pakistan, Bangladesh and Singapore have also shown interest, although song copyrights make those tie-ups more complicated.
Manohar is cautious, but he agrees that Mobile Antakshari was a great learning experience. What kind of apps does Manohar see in the future? “I want to reach out to rural markets directly through healthcare and educational apps,” he says. It’s time, he thinks, to start singing a slightly different tune.