By Ian Cameron, VP of Digital Media, and Daniel Mayer, CMO and SVP for Products, Expert System Enterprise –
In 2016, the seven biggest film studios were hit by an estimated 15 percent decline in profit, a drop of $700 million from 2015. This decline is occurring, in large part, because of changing content consumption patterns and distribution disintermediation.
As three example applications show, artificial intelligence may prove precious to the entertainment industry as it adapts to this disruption.
Compelling consumer experiences
From the consumer’s point of view, choice is no longer limited by networks’ programming decisions. Quite the contrary, thanks to OTT, consumers have a proverbial long tail of tens of thousands of options available within a few clicks on their remote. This trend has gone so far that if they’re willing to pay, consumers can watch a movie at home that is in the theaters, at least for some independent titles.
While “more to pick from” has proven attractive for consumers, it carries new implications for the industry. As a content producer, how can you ensure that your programs will be findable within a long tail content catalog? As a distributor, how can you ensure that your catalog will be as easy to navigate as it is deep? Indeed, in this disrupted landscape, compelling access to content is becoming as important as compelling content itself. Offering a better experience to consumers is the new imperative.
The good news is that the entertainment industry can look at how other industries – such as online retail or news – have successfully dealt with the same long tail issue. In short, what these industries have found is that compelling access to content – typically through smarter forms of search, navigation and recommendation – hinges on a defining principle: metadata.
Metadata is well-known in entertainment but what is less well-known is that new artificial intelligence platforms enable content providers to apply descriptive tags to their assets more deeply and in a fraction of the time required when indexers do it. Positioned as a plug-in to your chosen repository, such a platform processes your content, classifies it according to your schemas and recognizes the personalities, locations, brands, characters and topics that appear. In the consumer’s living room, deep metadata translates into captivating, personalized recommendations and fewer clicks to find what they want to watch.
AI therefore represents a strategic lever for content producers and distributors to ensure that consumers enjoy the best content and a differentiated access experience.
Connect to consumers with audience analytics
AI-driven metadata carries a second implication for entertainment. Imagine knowing in detail what your audience wants, when they want it, and why they gravitate towards you. What are their sentiments before, during, and after a new release? What motivates your audience to consume content? AI-driven metadata can help answer these questions.
Indeed, enriched asset metadata puts traditional analytics on steroids. Imagine mapping how specific parts of the audience or particular distribution channels react to any specific part of your schemas: asset types, characters, locations, or any of the other descriptive dimensions that you will apply as metadata. This is an immediate outcome of AI-driven metadata and can produce trend and “early signal” insights that will help make quicker and more informed decisions in the creative and distribution departments.
A slightly different, yet potentially game-changing application of AI is in the analysis of consumer comments, whether in the context of audience feedback to a trailer or to social media comments on the finished product. Corralling such large, unstructured datasets is a challenge with traditional methods, but can be readily automated with AI, generating answers to such questions as: Which aspects generate the strongest emotional reactions from the audience, and how positive or negative are these reactions? Which words or themes is the audience using to describe their reactions reactions (often very different from what a content creator would expect)?
The impact of such analytics again enables better creative decisions, and ultimately a better connection with the audience.
Drive efficiencies in licensing
Licensing offers a third snapshot of how content providers may exploit AI. The use of clips from movies or television shows in advertising is an attractive path to brand recognition. When a brand approaches a studio to create a 30-second ad for a show, will the studio sift through hundreds of thousands of clips to identify the relevant ones?
The answer again is that the use of AI to tag assets is now the more realistic and efficient option, as opposed to an immense manual process consisting of hours upon hours of physically screening clips to find those where the brand’s products, or a desirable context, are prominently displayed. An AI platform on the other hand processes entire archives to identify the desired characteristics that appear in its assets, letting the content provider pinpoint the relevant assets in a few clicks. This ensures content providers can meet advertisers’ evolving demands on the double.
In conclusion, while asset metadata is a known implement in the industry’s workflows, it has taken on a new importance in the increasingly disinter-mediated distribution landscape.
In this context, AI offers a proven and efficient path to unprecedented metadata granularity at a fraction of traditional methods. As the three examples above show, AI holds promises for studios and distributors alike, whether to deliver a differentiated consumer experience in accessing content, deep audience analytics supporting creative and distribution decisions, or more effective licensing capabilities.
The secret is out!