(ThinkAnalytics will showcase its solutions April 8-11 at the NAB Show, booth SU7525)
ThinkAnalytics’ solutions — including its personalized content discovery platform and its industry-first viewer and video insights platform — are all geared toward one thing: giving pay TV and OTT players a leg up, by providing a comprehensive view of their video business, and enabling them to better address key performance indicators (KPIs) that boost engagement resulting in increased viewer loyalty.
With more than 80 customers representing more than 250 million subscribers and supporting 43 languages worldwide, the El Segundo, Calif.-based company’s machine learning and AI technologies power content discovery, personalization and viewer insights, across all devices.
Gabriel Berger, CEO of ThinkAnalytics, spoke with the Media & Entertainment Services Alliance (MESA) about the challenges operators face in delivering personalized content, how AI and machine learning techniques are playing a crucial part in analytics, and how the industry is “only in the second inning of content discovery.”
MESA: How did ThinkAnalytics first come on the scene, what was missing in the industry that the company has looked to offer?
Berger: ThinkAnalytics was founded in 2001 with a mission to develop real-time data mining software, a machine learning platform to help big corporations understand their customers and their value to the business. As the market evolved, the company changed tack to target the fast-growing pay TV sector – focusing on churn risk and personalized experiences.
As far back as 2005, we began incorporating new and innovative machine learning techniques into our personalized content recommendation solution and deployed with our first customer (Sky in the UK) to over 8 million viewers, thousands of VOD assets and over 500 channels of linear TV. This was the first TV operator anywhere in the world to deploy a personalized discovery solution for live TV and VOD content, helping subscribers find content they want to watch.
From the beginning the business case was solid: helping our operator customers demonstrate value for money, boost engagement, reduce churn, and increase ARPU by promoting premium content to their customers.
The team has been evolving our AI/machine learning algorithms and best practices techniques ever since to help our 80-plus customers worldwide meet business KPIs including those related to uplift in viewer engagement, loyalty, and ARPU.
We are the global leader in personalized content discovery and real-time viewer insights solutions and services, with customers scaling to tens of millions of subscribers and serving over 3 billion multi-language recommendations every day. Our customers include Liberty Global, BBC, Proximus, Cox, Rogers, Sky, Swisscom, Astro, Singtel, TataSky, Viacom18, Vodafone, and others.
MESA: In recent years, what are some of the biggest changes in how personalized content recommendations are done?
Berger: Really, one of the biggest industry challenges today is not the volume of content a viewer has to choose from, it is the speed with which that content is coming online and the impact of changing release windows. All that influences how you merchandise the experience with your viewers and really drive viewer engagement.
Of course, there are other changes as well, like the evolution of the TV user experience. The days of clunky EPGs are long gone, and viewers now demand a seamless experience across multiple devices with content personalized to their individual tastes. Add voice navigation and discovery to the mix and the experience is totally changed — decreasing content discovery times and boosting overall viewer engagement.
Another change is the rise in use of real-time analytics to measure the effectiveness of editorial, promotions, and advertising. When operators and content creators can see the impact of campaigns on viewer behavior in real-time, they can quickly make adjustments to fine tune their marketing efforts and maximize their investment in content.
All of this is underpinned by the value of enhanced metadata – including multi-language support. Accurate, in-depth metadata is essential in providing personalized experiences that effectively enable the ability to ‘merchandise the store’ and deliver the right content and services at the right time to each viewer.
MESA: Additionally, how has AI and machine learning impacted viewer insight data offerings, and how has ThinkAnalytics incorporated these technologies?
Berger: Our solution portfolio includes viewer analytics and audience insights that use AI and machine learning techniques to give our customers a complete viewer lifecycle management system.
Armed with a holistic, real-time view of their business, our customers become more agile and are better able to meet business KPIs such as improved viewer engagement and an uplift in ARPU. This ensures that they are continually driving their business forward.
These insights solve specific problems — such as how to boost engagement, stop a subscriber cancelling their subscription, or entice them back to their service. They are also used to drive decisions on content buying, pricing, marketing campaigns and editorial promotions. And, we are starting to see this data being used as a basis for new revenue-generating opportunities such as dynamic ad insertion.
MESA: What are some of ThinkAnalytics favorite use-case examples, where clients made especially good use of the company’s offerings?
Berger: Over the years we have enabled some of the largest pay TV and OTT operators and content brands around the world to maximize their customer value and loyalty. Our customers range from large operators like Liberty Global to public broadcasters like the BBC (for BBC iPlayer and BBC Store), OTT players such as Viacom18 in India, and multi-country, content-specific OTT players like DAZN with its focus on sports. Each of these leverage our machine learning and AI technologies to power personalized content discovery and real-time viewer insights and drive increased engagement across all devices.
For many of our customers, we help upgrade content discovery experiences across the board. For example, we help deliver personalized recommendations to Liberty Global’s Horizon platform so that the most relevant content is presented based on individual viewing habits; we also provide support for multi-language voice search. For the Horizon Go TV Everywhere app, we ensure relevant content recommendations are consistent across web, mobile and tablet. And on the Horizon OTT platform we help broaden the overall repertoire of content a viewer consumes.
This level of personalized content discovery is designed to help viewers easily find what they are looking for, and also discover new content. This results in better retention and increases overall engagement.
MESA: What advances in content discovery and viewer lifecycle management are on the horizon, not just for your company, but the industry as a whole?
Berger: Capturing and managing TV/video platform data from multiple sources for analysis using advanced, predictive algorithms is a key focus for the entire industry.
By using AI and machine learning predictive algorithms to understand what is happening across the entertainment platform — for example, what the consumer is doing on what device, and what type and genre of entertainment content they are engaging with (e.g., TV, video, music, games) — you can predict what the consumer wants to do next and surface the right content at the right time. Adding a feedback loop means that you can continuously improve the relevance and success over time.
The average consumer subscribes to four video/TV services and is forced to search and navigate each one individually. This is unsustainable. The industry is moving towards universal federated search, recommendations and navigation. This is possible today from a technical perspective, but we need a level of convergence between services to deliver the reality.
Voice search is quickly becoming a vital, component of the overall content discovery experience. Today, on many services, it may only be possible to voice search for example, actors’ names or on “I want to watch Mission Impossible 5”, but the pace and development of voice search will no doubt become more intelligent and nuanced. Ultimately, consumers will want to be able to navigate around the content on all TV platforms, gaming apps and radio channels using just their voice, so companies will need to work together to offer a truly connected experience to the consumer.
MESA: What’s next for ThinkAnalytics, what can we look forward to from the company in the next year?
Berger: We are only in the second inning of content discovery. We will continue to refine our portfolio of products and services to help our growing roster of customers better achieve their business goals. The industry will see us talk more broadly about how our solutions help customers to ‘merchandise the store,’ boost premium content discovery and continue to drive the new engagement economy.