Big data analytics could transform the video experience as we know it – but only if video service players can do it securely and transparently
For ages, perhaps the holiest of Holy Grails for television broadcasters and advertisers has been audience data – knowing not only how many people watched a given television program (and saw whatever ads were shown during the breaks), but who they were (male, female, age, etc.), how long they watched, what they were doing while they watched, and whether the ad successfully convinced them to buy whatever product or service it was hawking.
Up to now, gathering that data has been limited to ratings surveys, and the results were more guesswork than exact science. That’s about to change thanks to big data analytics. As everything goes digital, consumers are generating vast amounts of digital data – and that includes video services delivered over IP networks, whether in the form of IPTV services from telcos or OTT video services like Netflix, Hooq and Viu, to name a few.
The data generated from such services can potentially deliver all the audience stats advertisers could ever dream of, but the application goes way beyond that. Big data analytics can also be a customer engagement tool to deliver a better video experience by helping content consumers discover content by pushing targeted recommendations to them. As analytics technologies become increasingly predictive as machine learning and AI enter the picture, content providers hope to eventually be able to anticipate what their subscribers will want to watch next – possibly even before the subscribers know it themselves.
Not that it’s that easy – or simple. Applying big data analytics to video services (or almost any service, actually) is a challenging task, not least because of the sheer scale of data involved. Netflix, for example, estimates that its 86+ million members globally streaming over 125 million hours of content per day has resulted in a data warehouse over 60 petabytes in size.
An even bigger challenge – and a risky one at that – is data protection and security. The last couple of years have demonstrated all too clearly that personal customer data is a target for hackers. And when it’s your customer data being compromised, the damage in reputation alone can be staggering – just ask Yahoo. Or Target. Or Sony Pictures. Et cetera.
A related risk with big data analytics is customer concerns about privacy and how their data is used. The collection, buying and selling of big data has been going on for years, but consumers are only just becoming aware of the extent to which this is going on and – more importantly – how little control they have over it. And that anxiety can result in legal action – just this week, we learned Bose is being sued over an app for its wireless headphones that allegedly tracks whatever you listen to and gathers the data, which Bose then allegedly sells to third parties without permission.
Whatever the merits of that particular case, it reflects a real risk of collecting and analyzing big data. Video service providers may be able to deliver fantastic benefits to customers via big data analytics, but their customers won’t thank them for it if they feel their privacy is being unduly violated – or if the data they trust them with gets stolen.
(And we haven’t even gone into the potential creep factor of throwing machine learning and AI into the mix – not everyone is going to be thrilled by the prospect of their video service making a recommendation based on something they said out loud that was picked up by their Amazon Echo Dot, for example,)
To be sure, big data analytics has the ability to transform the television experience as we know it, but it needs to be executed responsibly by following best practices and complying with local legal requirements for data collection and privacy (and going above and beyond those requirements in markets where regulators are technologically behind the times). That’s arguably as important – if not more so – than the technological wizardry enabling a great experience.
First Published by John Tanner disruptive.asia 21.4.2017