FEED Issue 12



ADVERTISING As the privacy model changes, the

advertising model will change with it. At its best, advertising gives the viewer new options they hadn’t thought of before. They become aware of a new product, a new brand – or an idea – that wasn’t on their horizon initially. However, we all know advertising often falls short of that. As content owners using the ad model, we have to play the agent between our advertisers and our viewers. Media businesses using the advertising model are all too often stuck in a situation where the tail wags the dog, with the advertiser’s needs taking priority over the viewer’s. At first glance, targeted advertising promises a solution, giving customers more relevant ads that are, if not perfectly aligned with their needs, more palatable. But this type of targeted advertising relies on deeper customer data, which, as we’ve said, viewers are becoming increasingly suspicious of. The Cambridge Analytica scandal, in which Facebook customer data was used to micro-target political ads in a number of elections globally, raised awareness that targeted advertising is not all about selling cars and cleaning products. It’s possible that the new suspicion of targeted advertising may lead to a greater emphasis on SVOD content, with people choosing to pay a little more to avoid being hit with messaging. But it should at least encourage us to double our efforts in making targeted advertising tech something that benefits customers first and advertisers second. GAINING CONTROL OF THE ROBOTS Artificial intelligence and machine learning are on the verge of revolutionising just about everything we do. Simple machines – the wheel, the pulley, the inclined plane – have been the basis for all of past human engineering, but the future is going to be built by digital machines, configurations of artificial intelligence that will, like those classical simple machines of your secondary school physics class, amplify human will. You couldn’t design a more perfect testing ground for AI than the digital media

sphere. Machine-learning applications need data – situational data – to hone and perfect the skills they’re being assigned to perform. Our connected media world is a gold mine of data about human behaviour, language, thoughts, desires – even secrets. Machine learning is going to enormously augment already existing technologies and is going to create a host of new technologies we haven’t even thought of yet. From concept development to production to distribution to rights management and accounting, what once took a team of humans could be reduced to a single individual ticking off a completed process on a dashboard. It’s possible that there will be a time when you just ‘spin up’ an entire programme, or series, all instantly scripted, designed, rendered and

distributed on command – and probably customised for each viewer. On the other hand, AI is going to make it much easier for people with bad ideas – content pirates and cybercriminals – to accomplish their goals more easily and to, in a short time, iterate better and better means of doing so. We might enter an AI arms race, with AI employed to counter threats to business, in turn countered by even more powerful AIs. A company’s security team will just be a coach for AI to silently battle with an ever-present enemy you can’t see. Algorithms and machine learning already have a powerful effect on the customer experience. It’s true customers are inclined to notice the negative experiences with AI. We’ve all gotten those way off base content recommendations (‘If you liked The Shining , you’ll like The Grand Budapest Hotel ’), but when good technology is working well, it’s not noticed at all, and that should be the AI experience going forward.


feedzine feed.zine feedmagazine.tv

Powered by