Startup Village incubatee Riafy's app seeks to predict box office performance before a movie is released

Riafy Technologies Pvt Ltd, a student startup providing relational intelligence, predictive analysis and Big Data solutions, has rolled out a mobile app called Movie Tarot  for predicting box office sales of movies before their official release.

An incubatee of Kochi-based Startup Village, Riafy was set up in 2012 by a group of engineering students, mostly from Sree Narayana Gurukulam College of Engineering (Kochi). In a bid to predict a movie's box office performance, Riafy's movie app analyses the online presence of flicks using its core relationship system. Relevant data collected from the internet (news, blogs, forums, videos, Facebook pages, Twitter feeds, etc.,) are assimilated and then crunched to arrive at crisp insights.

"Our app Movie Tarot predicted that Bollywood flick Kai Po Che would rake in almost 18 crore in the opening weekend. And the collection was almost Rs 18.1 crore, much to our expectation. There was a fluctuation of just 0.55 per cent," said John Mathew, co-founder and CEO of Riafy.

Movie Tarot is a free app and currently available only on Android OS, but it will be on other platforms soon. The app was earlier picked by BlackBerry for its newly launched BlackBerry10 handset. The startup claims that the app has seen close to 2,000 downloads since its launch five days ago. The company is also planning to introduce a freemium model once the app gets enough scale where subscribers will find additional information (trailers, reviews, etc.) via the app.

Riafy claims its technology will not only help producers and distributors, but movie-makers as well. "The data that we process will be useful for film-makers too. We make predictions based on a lot of factors and we will share it at different stages of movie promotion. The data can help them chalk out most effective promotion strategies," explained Mathew.

Is that the business model the co-founders have in mind? The traditional prediction market is already seeking more and more sophisticated algorithms for better market insight and unless the new app is positioned well, the B2B leverage could be a far cry. Also, in practice, only hugely buzzed-about flicks are open to this sort of analysis and the less-hyped regional movies may not opt for such forecast results.

In November 2012, a group of scientists at Budapest University of Technology and Economics came out with a mathematical model through which they can predict whether a blockbuster film is going to be a hit or a miss by looking at activity patterns on Wikipedia, the online encyclopaedia. But Riafy's differentiator is that it can predict how much a movie can collect at the box office, as well as the percentage of people who are going to like it.

Unlike other prediction tools, the app can determine the number of satisfied viewers, the startup claims. There is an index that tells you the percentage of people who are likely to be happy after watching a particular movie. "We believe that critic ratings can't determine 'true happiness'. When a person rates a film 2 or 3 on a scale of 5, it can be biased opinion. For example, an action movie fan will give 4 ratings to an action flick while the same person may rate a romantic movie at just 2. So it can't be taken as a measure for happiness," said Mathew.

According to him, movies high on 'true happiness' are the ones that will sustain the collections after the initial hype. "Consider Aamir Khan's Talaash. There was a huge expectation about the movie and it did well in the opening weekend. But our engine picked a relatively low 'true happiness' index for the flick and predicted it wouldn't hit the Rs 100 crore mark in domestic collections. Eventually, the film stopped at Rs 93 crore," he added.

What's in it for movie fans?

Riafy says it can provide a more tangible gauge of engagement than a traditional review or trailer, and it's obviously much more in real time. For this, the app analyses the online buzz around a movie and then quantifies it in terms of positive and negative buzz. The predictions are then posted into the app prior to the movie release and it will tell you if the movie you would like to watch is going to be a blockbuster, super hit, hit, average, flop or a sleeper, explained Mathew. "Now movie-goers need not spend time reading reviews or watching trailers. Our app can tell him in a minute or so whether it's going to be a hit or a flop," said Mathew.

Still, one cannot ignore the flip side of the logic. Given the algorithm scans existing 'positive' or 'negative' buzz, it may be as good or bad as other clairvoyants in predicting a surprise hit. Moreover, the app doesn't take into account the personal preference of the would-be viewer seeking help. Hence, a movie fan may get a positive ranking for a flick (generated by popular filter) and still be disappointed as it does not match his/her personal choice. Such loopholes must be plugged if Movie Tarot wants to make its mark in both B2B and B2C markets.

(Edited by Sanghamitra Mandal)