The Electronic Retailing Association

2018 MCMS Speaker Insight: Data, AI, TV and the future of commerce – Charles Dawes discusses 3 trends that are making a difference


In June 2018 I had the pleasure of participating in a panel discussion on 'Data, AI, TV and the future of commerce' at the annual MCMS in Seville, Spain. In this article we will examine 3 areas that TiVo believes are making a difference to the lives of consumers.

Charles Dawes, Senior Director International Marketing, TiVo

Charles Dawes leads TiVo's international marketing team across APAC, EMEA and Latin America. A regular speaker at industry events, he is responsible for evangelising how TiVo's world-leading products, services and technology licensing programmes can help media and entertainment companies bring amazing, award winning experiences to their consumers.

Charles began his career at Cable and Wireless, where he was responsible for electronic programme guide development. Charles also held the position of product management director at UPC/Chellomedia, where he spent a decade overseeing UPC’s digital television product development across their European footprint. Prior to TiVo, Charles was at Rovi where he held positions in product management across Europe and Latin America, Global Account management and International Marketing.




The move to IP

Over the past few years consumers have experienced an explosion in the way they can access content thanks to the rapid deployment of technologies that use IP (Internet Protocol) as the backbone of their communications. This has given rise to people being able to access almost any content, at any time, on any device, anywhere. Just over a decade ago, there were no smartphones and the only access people had to the online world was via their computer. Move forward to today and you have just under 710 million subscribers in Europe with a 3G or 4G service. This number will continue to grow and by 2022 there will be 25 percent more subscribers with a 3G, 4G or 5G connection – 70 percent of which will be 4G or better. This means that people’s lives are connected wherever they are and can access services and communicate with others at home, at work and on the go. This near-ubiquitous access – in addition to fixed-line access from fibre-based networks that are becoming standard in many households – means that consumers really do have access to more content and services than ever before.

However, this brings a whole new set of challenges for the consumer and service providers. In the content discovery realm, TiVo – an established world leader – and its predecessors have been instrumental in helping consumers find and discover entertainment content. From the first onscreen TV guides in the early 80s through to the mass deployment of the first commercially viable consumer DVR (Digital Video Recorder) in 1999 that kick started the ‘content on your terms’ revolution, we have remained at the leading edge of developing the technology that drives consumers to content. However, we still have consumers tell us that there isn’t anything on for them to watch or listen to. This leads us to our second area of disruption.

Charles Dawes (third from left) takes part in the MCMS 2018 Panel discussion on 'Data, AI, TV and the Future of Commerce', hosted by Ben Keen (far right).


For the last few years more and more of our lives are being affected by the introduction of Artificial Intelligence and Machine Learning. No more so than systems that help us to make a purchasing or content consumption decision. In fact, leading analyst firm, TDG, predicted in 2016 that within ten years seventy-five percent of all viewing would be initiated by a recommendation-based guide.

Outside of the content discovery space many online retailers have deployed basic recommendation engines that help consumers make purchasing decisions. However, many of the early implementations have been limited by the amount of data they use to make decisions, and this sometimes has led to frustration as consumers are recommended to purchase something they already have, or they have ultimately bought elsewhere, and the system doesn’t realise.

In the content discovery space, an example of this would be that you watch something on one platform, let’s say your mobile, but through limitations with data and connectivity when you return to your main TV you’re recommended that you watch it – again! Similarly, many systems have relied on one or two methodologies to ascertain their results. This can result in an ever-narrowing focus that limits the content choices and doesn’t fulfil the promise of helping consumers to expand their content horizons. As such, the serendipity is lost.

Step in the latest generation of recommendation engines that not only look at past content consumption to predict future likes. These new systems can draw on a plethora of information sources and use other indicators such as time, day, device and location, to help determine what the best recommendation for the customer is, in their moment of time.

Recommendations is a requirement for any system that is looking to implement our 3rd disruptor – Voice.


Voice interaction

Over the last few years we have seen the use of our own voices as the input method increase massively. It’s now a decade since the first voice services were introduced, and by 2016 Google was already reporting that there was a 35x growth in voice search queries on its platforms. Going forwards, Comscore has predicted that by 2020, half of all searches will be initiated by us speaking to a device or digital assistant.

This represents a fundamental shift in the way that we interact and search for content, and how computers will adapt to our needs and desires. As Brian Roemmele, one of the world’s leading commentators on the topic has said;

The last 60 years of computing humans were adapting to the computer. The next 60 years the computer will adapt to us. It will be our voices that will lead the way; it will be a revolution.

Brian Roemmele

When we interact with our voices, we step away from the lists of disjointed keywords that we have subconsciously taught ourselves to type and move to using natural language, expressing our desires with an immediacy, convenience and intimacy that text-only search could never provide. Subtle nuances can affect the result, and we find that a greater level of personalisation is required as users imply that they want results ‘for me’ even if they don’t say ‘for me’ every time.

In the content discovery space, we are finding that voice is opening up new opportunities for operators and content providers. The industry has spent a huge amount of money and effort to give consumers access to content where and when they desire it, but as we have already discussed, consumers can’t easily find that content because the tools they have to hand are limiting. In the TV world this conundrum is exacerbated further by having a remote control as the input device. With a keyboard an average person can typically type about 20 words in 30 seconds. With a remote control and an on-screen keyboard, it is significantly less even with predictive text, but with your voice you can speak over 50 words in the same time frame.

When you tie this to a system that understands natural language and has a word recognition rate that equals the human ear, you give the consumer access to content like never before. Ironically, in the world of operators the very device that was limiting the interaction is emerging as the one that reopens the door. The humble remote control is being equipped with a microphone. Rethink Research, a UK based consultancy, has predicted that deployments of voice enabled remotes will rise from 21.3 million in 2017 to 142.9 million globally by 2022.

Additionally, we’re seeing more integration with smart speakers and other devices that are the natural home for content discovery, enabling new use cases for the consumer who is looking to be entertained.

'Show me action movies on tonight'

So rather than limiting a customer to searching for a specific title or actor, they can use language that is similar to how they would ask their partner on the sofa next to them. This means that rather than trying to type in “Arnold Schwarzenegger” with all the frailties of trying to spell it right, you can simply say “find action films from the 80’s that star Arnie” or “What’s the film where the cyborg says, “I’ll be back”?”.

To enable this, the systems need to have access to more information than ever before. The days of a piece of content only needing a title and short description are no more. Now we equate the data for each item to an iceberg. There is a small amount of data that is seen by the consumer – the title, synopsis, list of actors and images like the box art or episodic images for TV series – but under the surface, a wealth of further information exists that allows the recommendation and voice engines to deliver the right results. Much of this information is created using AI and Machine Learning, where systems are able to make sense of huge sets of unstructured data about content, categorise it, and link it to information that has traditionally been created by human editors. This then forms the basis of deep, dynamic, graph-based information stores – such as TiVo’s Knowledge Graph, aka ‘the Entertainment Brain’ – that enable voice and recommendation systems to deliver the right result at the right moment, ensuring that when the consumer is looking for something specific they find it quickly, and when they are looking to have their content horizons broadened and discover new content they aren’t narrowed down into an ever smaller funnel, leaving a feeling of serendipity remaining.

To find out more about TiVo visit them at or follow on Twitter @TiVoForBusiness