At the recent Future of Video: OTT, Pay TV, and Digital Media Conference hosted by Parks Associates, industry leaders came together to discuss industry trends and insights with a focus on consumer behavior and other research. The week began with pre-show workshops and a panel discussion on artificial intelligence and machine learning Innovations featuring FPT Software’s Ira Dworkin, Managing Director of Communications, Media, and Entertainment along with fellow industry experts. We’ve compiled highlighted topics from the panel discussion with insight from our representative on the panel.
What is the state of the video market, including OTT service churn? Seems like the new normal – will the industry now be in a forever state of churn?
There is more choice among streaming services than there has ever been. Over the last five years, the choices have grown exponentially from five major services to over three hundred available in the US alone. According to a report conducted with Parks Associates, 83% of US internet households subscribe to at least one streaming service but only half subscribe to traditional pay tv services. Additionally, only 27% of users exclusively subscribe to the major services which include Netflix, Hulu, Amazon Prime, Apple TV+, HBO Max, Paramount+, and Peacock. This reflects a drastic change from ten years ago, when 87% of US households subscribed to traditional pay tv services. The wide variety of choices – and the ability for consumers to stop subscriptions easily – has contributed to churn. Another impact on churn is the availability of more free options for consumers. FAST (Free Ad Supported TV) has increased in popularity, with a model similar to more traditional linear TV programming; ad-supported subscription tiers for the SVOD providers provide yet another avenue competing for consumer’s entertainment consumption. Despite the greater options to choose from, viewership time has not dramatically increased, excluding a short spike during the pandemic. Overall churn has risen from 28% to 50% over the last 4 years, and is expected to continue, given the many different services and a greater emphasis by consumers on cash-friendly entertainment options – but there are steps that service providers can take to reduce churn and retain customers.
Can OTT services build loyalty in an environment where it is so easy to cancel subscriptions and find similar content elsewhere? If so, how?
Yes – with some caveats. In our sponsored report with Parks Associates, it was noted OTT-only households are subscribing to an average of six services per household. This is double the number seen just a few years ago. Those services that have been around for a while – such as Netflix or Hulu – and, hence, have had customers that have been with them longer, tend to see lower churn than the newer services. Additionally, the services with a “built-in” fanbase, such as Disney+ or Crunchyroll, also tend to see longer subscription durations. Beyond those elements, the key factor comes down to providing ongoing value. People tend to move on from a particular streaming service for a few different reasons:
- They’ve finished a particular show or series
- Their promotional period has ended
- The service wasn’t worth what they were spending
- Budgetary restraints requiring a cut in spending
These points narrow down to how value is perceived by the consumer.
Entertainment services have enormous amounts of data that can be leveraged to improve perceived value and reduce churn. For example, data can be fed into models that help with:
- Ease of Finding Content: Beyond the most popular titles, consumers need to find relevant content quickly. By leveraging data around user preferences, how they navigate, what they watch, and more, services can surface “hidden gems” to the user – keeping them engaged.
- Better Recommendations: Services can improve upon the recommendations made to users by digging deeper into user actions. This includes going past the typical genre-based recommendations and perhaps using AI to more deeply analyze and classify what each user is watching.
- Different ways to consume content: Not all users like to discover content in the same way; some may semi-randomly browse, others may be very directed, while others may skew based upon the day of the week (for example, action movies on Friday nights but romantic comedies on Sundays). Models can be built to identify and segment these user behaviors. Different interfaces can then be built that cater better to individual needs.
- Churn Prediction and Customer Value: AI models can be built to recognize specific patterns that tend to lead to churn. This data can then be combined with customer value measurements to identify those customers a provider wants to save –then fed into automated retention campaigns.
- Pricing: Optimal pricing can be found which best matches the perceived value and rate at which people subscribe to (and leave) the service. Combinations of different models can be offered (e.g., ad-supported, subscription, etc.) with different price points to reach the optimum mix.
- Service Reliability: A repeatedly poor experience when hitting “play” will rapidly drive users from a service. Tracking end-user metrics such as buffering, and bit rate can help optimize delivery and address issues. Abandon rate can also be addressed through understanding content mix and what content users tend to begin but then leave the service.
What decisions do video services face that could be addressed with better user data?
Better data can drive improved decisions in a wide range of areas, including the following:
- Overall Content mix: Many services are already using data to some degree to determine the types of content they offer; deeper analysis can help further segment the content and users
- Improved levels of personalization
- Delivering the most relevant content to the user
- Identifying and preventing churn
- Discovering which marketing and outreach programs net the greatest results
- Ways to re-engage previous users who have left the platform
- Developing the right pricing strategy
- Narrowing down the correct ad mix
Most services have access to the right data (as outlined earlier) to inform these decisions, but often do not leverage this information to the extent possible.
How is Artificial Intelligence changing the search and recommendation process? What types of user data can inform these decisions? From what sources is this data available?
AI is making for richer, more accurate recommendations unlocked through better targeting for ads and entertainment. It also enables more and better ways to engage beyond the standard carousel and search bar. AI allows us to go much deeper in content than recommendations based strictly upon metadata. Vionlabs, a startup focusing on creating richer metadata via AI, for example, indexes and scores entire movies leading to more accurate and meaningful recommendations. Services which leverage models such as these can go beyond recommendations and search based just upon what a user watched – recommendations and search results might be based upon an overall profile of a show or movie. Leveraging click path data, such as the steps that a user takes to get to “play or parts of a title which were rewound or fast forwarded the most, coupled with a user’s watch history and the overall “profile” of a title, AI-based systems can lead to better engagement. This will ultimately lead to greater loyalty and reduced churn.
In a crowded market, what sets solution providers apart?
It comes down to one word: Value. Although overall value has many different aspects, the three components of the value equation, as discussed above, are price, content, and ease of use.
- Price: There are a lot of services competing for a finite share of a consumer’s wallet. Price of the service – especially in tough economic times – plays a part in service adoption. Many providers are moving to a hybrid model, offering a lower priced, ad-supported tier of service, and higher priced ad-free tiers that may add in more “premium features” such as 4K and specific content.
- Content: Once you get beyond the tentpole titles, a service must have enough other content that’s interesting and relevant to its users to retain them. A service provider’s ability to segment their content and present different options to different types of users will extend the length of time a user will stick with a service.
- Ease of use: The user interface and ability to personalize the platform for the individual can make an overwhelming difference in the perceived value a service offers. Leveraging AI, service providers can simplify the user experience – offering different navigation models based on the type of user. AI can also be leveraged to deliver better, more relevant advertising to users. This goes beyond the typical demographic and psychographic targeting, getting into more detailed analysis and placement within a piece of content (e.g., delivering an ad for a sports car after an intense car chase, or an ad for a couple’s getaway following a romantic scene) – increasing revenue for both the advertiser and service provider.
FPT Software and AI
FPT Software has extensive Media & Entertainment experience – in AI, machine learning, and beyond. We can help Media & Entertainment companies accelerate their AI journeys and meet broader development goals. To speak with someone to discuss your needs in more detail, click here. To download the Parks Associates whitepaper, click here.