By Sebastian Stehn, Norigin Media
Streaming TV has existed for around 15 years on mobile, and now nearly a decade on Smart TV. There has been a huge amount of digital growth in this time, from the sheer amount of streaming services on offer, to a major shift in consumer behaviour. Such digital services offer an abundance of data and information to track, analyse and make use of.
Aspects such as broadband connection, video quality, content search or navigation, ratings and usage patterns make the big data of TV or business intelligence an intricate art form. This becomes more complex when you consider that TV is streamed on several devices, from browser, mobile and tablet to Connected TVs and Chromecast.
According to NPAW, a prominent provider of quality of experience tracking on video services; “4.9% of VoD views experience some sort of error, and for live views the number is 7.6%. If we compare this across different devices, on browsers, for example, start-up errors are 22% more likely, and in-stream errors are 40% less likely for VoD content. On Smart TVs, startup errors are 45% less likely and in-stream errors are 69% less likely for VoD content. Thus, you can see that you cannot take the same approach when you have the same service on multiple devices.”
ANALYTICS ON CONNECTED TV
Different devices support different technologies. Just looking at Smart TVs for example, the matrix of support for video streaming formats, and the DRMs you can combine it with, is very complex (see this guide). What’s more, each of the formats supported natively by each smart TV device are tracked with different analytics SDKs and providers. These have an impact not only on how data is aggregated, but also complicate aspects of monetisation with adverts and tracking of user experience due to the OS capabilities.
Although Connected TV has seen major growth globally, broadcasters are delaying its presence due to the lack of good analytics and its integrations, which provide the foundation for advertising reporting and revenue.
TOO MANY TOOLS TO TRACK TV?
Big data mining and analytics have no agreed standards within the TV streaming industry. The input for analytics are often overlapping, but with very little technical similarities from source to consumption.
There is a huge variety of choice regarding the technology that collects data for video quality, internet connectivity, user behaviour and consumption patterns. Collecting this data is one thing, but collating the information and making sense of it is a whole other ballgame.
To broadly categorise the analytics sources, the most important items to track are the video, the experience and the ratings.
Video quality majorly impacts user experience. This category of analytics includes video tracking, Quality of Experience (QoE) and Quality of Service (QoS), which focus on tracking problems in the delivery chain of the video stream, and measurement of its eventual quality.
The value chain for video delivery extends from content original formats, transcoding choices, streaming protocols, origin servers, content delivery networks (CDN), internet service providers (ISP), network operators, home routers, devices, players and encryption, to mention the basics.
This is a ‘very’ simplified flow of where a video stream originates and ends. Each of the sections require its own expertise and technical adaptations where one can foresee many different risks for failures or service impacts.
Analytics providers for QoE (which measures the delight or annoyment of a service) and QoS (which measures the overall performance of the service), have in recent years succeeded in supporting the collection of data around this value chain. However, they still have a long way to go, where such data is exploited to its fullest potential.
Companies which provide QoE and QoS monitoring, including Agama and Bridgetech, give OTT TV service providers the ability to monitor and collect data in real time.
Collecting user data from video streaming services is complex as there are many non-video aspects to it. This includes how users find, navigate and consume content, and how they experience the flow of accessibility. Search and discovery, hence become key business intelligence that results in effective monetisation or service improvements.
Understanding user behaviour in terms of churn, satisfaction, subscriber numbers, consumption patterns and advertisement tracking, are all examples of analysing the experience.
Examples of such analytics providers including Nice People at Work (NPAW), ad-servers, Comscore, Google Analytics, Tableau, and more. Such providers are normally experts at aggregating and visualising data to help make sense of it all.
TV ratings are key for audience measurement across different services, and have been (and still are) the most important data source for advertisers on where to invest their advertisement budgets. In Norway, Kantar Media is responsible for this while Nielsen is the main actor in the US. For OTT, these companies will require you to integrate streaming statistics and sometimes ads impressions to bespoke APIs on top of your internal trackings.
Below is a simplified snapshot of different analytics tools and providers for the evolving world of TV. This matrix is just an example of some key providers which come to mind, but is hardly comprehensive – it’s just a starting point for tracking video, experience and ratings.
The multitude of sources of where and how data can be collected are questions of technology. The use of such analytics spreads across product, business or technology, impacting strategic decisions. The plethora of data and its potential is only limited by privacy and investment issues.
Data drives everything. Ultimately, the most important question is not how to aggregate this data, but how to improve the service quality and value proposition for consumers, advertising and the video industry as a whole. Big Data tools are being developed and improved every day, as streaming services evolve fast and technology is only a means to an end.