A common fallacy among investors and entrepreneurs is the notion that the ‘clear winner’, in any given market, will be able to sustain its leading position. History has proven that, time and time again, markets which seem crowded, unattractive, and boring, can be reinvented with a burst of technological and/or business model innovation.
The search market in the late 1990s provides a case in point. Sergey Brin and Larry Page, Google’s founders, were often greeted with skepticism when they discussed what they were working on at Stanford – “search is solved” they were told. Of course in hindsight we know that it wasn’t (and in many respects still isn’t). Further, to draw on a more recent example, many considered 2007 to be too late for the launch of a social networking business. But look at what Twitter has managed to achieve since then. Both Twitter and Google demonstrate an important aspect of entrepreneurship: the ability to smash perceptions of ‘solved problems’ by doing things differently.
Fast forward to 2009 and many would argue that online movie search and discovery is solved. Consumers have all kinds of options when it comes to making moving viewing decisions, including Rotten Tomatoes, IMDB and Netflix. In order for to upend the status quo and gain some consumer mindshare, new entrants will need to do things very differently.
Israel-based startup Jinni, a taste engine for movies and TV shows, appears to be innovating in a way that will resonate with consumers. The company, which also maintains a presence in London and the US, differentiates itself with the Movie Genome Project. Through this initiative, the company is developing an understanding of the attributes of films and television shows across an evolving set of ‘genes’. The insight obtained through the decoding of premium video content is then used to power Jinni’s recommendation service.
The concept of a ‘Movie Genome’ obviously runs parallel to Pandora’s Music Genome Project. Both projects focus on decoding the fundamental building blocks of content and employing this insight to match users with the most appropriate content given their unique preferences. But, where Pandora decodes music manually using a team of musical experts, Jinni decodes movies using an automated system.
For any given piece of premium video content, Jinni analyzes the metadata and relevant text using proprietary algorithms. Based on their analysis, Jinni is able to identify up to several thousand genes for each piece of content. The taxonomy of genes – which includes variables like mood, time and place – was developed by movie industry veterans. According to Roi Ophir, Jinni’s London-based VP Marketing and Product, this approach to profiling films offers two advantages. First, the process is consistent – limiting the amount of human subjectivity makes for a more consistent offering. Second, the process is scalable and fast – it can be done quickly and required limited human resources.
The company’s resources are divided between B2C and B2B activities. Both sets of activities focus on leveraging the same core intellectual property, for which the company is currently preparing to lodge patent applications. On the consumer side, Jinni offers a movie and TV recommendation service which is aimed at helping users answer the question: “What should I watch next based on my user profile?”
Following sign-up, users are presented with a Movie Personality Test. This test allows users to indicate some of their preferences at the outset, so that future recommendations will be more targeted. The Movie Personality Test would be more engaging if it presented users with a short video containing a series of clips from a diverse set of films, prompting users to ‘like’ or ‘dislike’ the films as part of the viewing experience (if using movie content in this way would be considered fair use, of course). This approach could be a better way of rapidly collecting data on user preferences, thereby avoiding the ‘cold start’ problem associated with recommendation engines generally. Nonetheless, as it stands, the Movie Personality Test helps the user get started and provides Jinni with a foundation for delivering personalized recommendations, and Jinni’s understanding of the similarities between different content is most impressive.
Users are able to use Jinni to conduct elaborate searches. For example, one can search for an “action-packed romantic movie” or a “sci-fi save the world with special effects movie”. Clearly, this is a different way to think about discovering content than the traditional genre-based approach. Once users identify the content that they like, they are presented with a number of options for obtaining it: rent a DVD/Blu-Ray from Netflix, Blockbuster, or LOVEFILM (in the EU); purchase physical copies of the content on Amazon or Blockbuster; or purchase downloads from numerous online services, including iTunes. The lead generation model represents the core revenue stream for Jinni’s B2C offering.
Users can also search for free online content – an area where a number of startups, including Clicker and Yidio, are now exclusively focused. Jinni’s free online content discovery component performed less well than the other parts of the service. For example, running a search for Southpark yielded no results, while on Clicker the search returned 191 episodes. Similarly, for Seinfeld zero results were returned on Jinni (though Jerry Seinfeld was suggested as a possible entry for ‘people’ search) while Clicker returned 9 episodes. Admittedly, free online video search has not been Jinni’s primary focus, yet there is still room for improvement, simply because the content index does not appear comprehensive relative to the competition.
On the B2C side, Jinni is in partnership discussions with OTT (Over-the-Top Video) providers, principally in the United States. (OTT is simply video content delivered to a television set via an internet connection, but is distinct IPTV as it is delivered via the unmanaged internet.) Jinni is offering OTT providers TV Guide-like functionality which leverages the core recommendation technology, as implemented on Jinni.com. The OTT market is evolving rapidly, with new services like the Roku Channel Store opening up the TV ecosystem, enabling content creators to bypass cable operators, satellite networks and TV affiliates. By securing partnerships with OTT providers Jinni is positioning itself to play a vital role in the growth of the market for premium content being streamed directly to consumers’ television sets.
Jinni was founded in January 2007 by Yosi Glick, COO, and Izik Ben-Zaken, CTO. The founders, and the rest of the team, have diverse backgrounds in the software/internet and media industries. Michael Pohl was appointed as CEO in April of this year, and is based at Jinni’s office in Portland, Oregon. When asked about the importance of having a US presence, Ophir suggested that it was essential, mostly due to the need to foster relationships with partners and prospective clients. But it also helps the company to stay closely in tune with the needs of the consumer market. While the internet makes it technically possible to operate from anywhere in the world, the Israel-based R&D center coupled with an external presence in the US and/or UK seems to be a winning model for companies operating in this space.
Jinni is a startup that offers an engaging user experience and it is well-positioned to take advantage of rapid growth in multiple, albeit rapidly converging, market segments. The company has raised a total of $1.2 million to date from a combination of individual investors and StartupFactory, and is currently negotiating a Series A investment.
by Geoffrey Mugliston