Unruly's Dan Best talks ±±¾©Èü³µpk10 through Unruly's work in predicting social sharing, while Dr Hamish McPharlin, director of Decipher Media Research, showcases the equipment used.
Transcript
Hi and welcome to Adventures in Adland ±±¾©Èü³µpk10 has arrived on YouTube This channel is going to chart everything across Adland and the media industries and all the exciting new developments that are happening For the first video we're going to take a look inside Unruly's social video lab and find out how they're tracking video online and seeing if they can predict sharing of videos on YouTube * So you're in the Unruly video lab so, obviously, it's a real place and the idea for lab is to bring the landscape together for clients and their agency's It's obviously a very fast-moving market with loads of new things happening this is where we help clients and agencies understand what it is how to make better use of the sector *So, Ben, while you are watching this ad, we're watching your response Yep - your map's just lit right up So what you were doing there is the EEG test - really great fun you can see some really nice stuff from that Two other things we're really keen on... One is facial recognition - what we can start to see is, in a particular piece of content, you can start to understand what was the highlight in that piece of content and that can sometimes be surprised to creative agencies and their clients * Show your face, and you can see that the little web cam here was picking up where it thinks your face is and it's found your eyes and your face and it's making some estimates, here, so it's made an estimate of your age How close is that, Ben, to your age? *Erm, 29 at the moment not as close as I'd hope... 26 - a bit closer *There we go... And then eye-tracking is the other piece, which we find can add really interesting bits of extra colour and the reason for that is branded content isn't your normal 15 to 30 second slot so where, with YouTube, we're finding branded content length is getting shorter...we're finding the actual opposite to that. So we're finding it's increased roughly year-on-year for about 2.5 - 3 minutes long. You'll notice that you look at faces a lot and that you scan copy repeatedly. So if you see some copy, you'll go back and forth You'll orientate towards faces a lot. People pick up a lot of non-verbal cues [ Ad playing ] It's interesting to note, if you pick up the badge of the car Which would clearly be something that the company would like * Yep So you'll see that the badge will arrive from time-to-time. And it's nice to know... if, there we go, you had a fixation on the badge there * So, within that piece of three minutes content What's the brand prominence like? Because, otherwise, a marketeer might say look... I've got lots to show for my content, but no-one understands who my brand is. And wer'e trying to get to the point of saying, actually, no you can put prominent branding in and it won't affect your sharing behavior Of course, it's really easy then to start measuring the brand uplift and the effect of your campaign So, it was really interesting to go Inside the Social Video lab, there. It was really weird to be hooked up so many machines and obviously the accuracy isn't always 100% there but with the speed at which technology is developing its really interesting how brands and agencies are going to be able to use some of this technology, especially the coming years to kinda optimise their content for the web and really make some awesome viral videos