Finding Data: A Pixar Story
Gemma Milneon 22 June, 2016 at 02:06
It’s no secret that data is a hot topic in the world of advertising. Agencies now have staple data analytics departments, are marketing themselves as powered by numbers, and are increasingly turning to the swathes of information which the world of tech has recently afforded us – to gain insight, inspiration and answers.
But speaking at Cannes Lions this morning, Matthew Luhn, Story Supervisor at Pixar Animation Studios, reminded us all that we really haven’t quite got the gist yet.
Luhn spoke about how Pixar are ‘obsessed with creating the illusion of life’. They don’t want to just create lifelike hair and fabric, they want the audience to be able to feel what the characters are feeling, and recognise their thoughts in themselves. To do this, Pixar work with psychologists to ensure they can create universally understood gestures and facial expressions.
One example is pupil movement: they researched which way a person’s pupils move when they’re thinking certain things. Top tip – if someone is creating from scratch a visual thought, the look to their top right; if they are recalling a visual memory, they look to their top left. Great for identifying the truthful from the less so.
This means that when Toy Story’s Woody is recalling times gone by with his owner Andy, we can empathise more easily to the pain of those memories as he seems more like us. Woody is real because his behaviour reflects what makes us intrinsically human.
But Pixar don’t stop at gestures and emotions for their characters. For Inside Out, Pixar ensured their teams were schooled on mood-based modes of cognition so they could create a true-to-life representation of how the brain works. For Toy Story 3, the team were taken on field trips to rubbish dumps so they could understand the process of getting the trash from the truck to the incinerator to the fire pit.
When Pixar have a problem, they identify the ‘data’ they require to answer their questions, and stop at nothing to get it.
It couldn’t be more different in the world of marketing. We don’t have as big a journey to get to the data, as we already have access to so much of it in the form of social media sentiment, purchase behaviour, online consumer journeys, customer location, audience profiles, connected objects… The list goes on, and it is ever growing.
There are so many issues that arise from our new beloved big data analytics – for starters, there is built in selection bias in social media (there are so many people who do not use or have access to social media, so there’s no way of getting a true representation of the population), and we are increasingly relying on Facebook and Twitter for our measurements and insights. We haven’t even quite decided what a ‘like’ means for the bottom line. And it’s not just social media data which is problematic; each of the data types are being mined for patterns and insights which can inform our thinking, but we’re rarely utilising the deep analytical methods required to get trusted extractions and so are acting on questionable results.
Criticisms of marketing data aside, the key issue is not so much with the way we are using the data but rather the order in which it is employed into the creative process.
Instead of us identifying a problem then finding a method of extracting an answer, using whichever data is appropriate, it seems we are starting with the data we already have and working backwards. We are being told that we must be using the data to inform what we do, but aren’t really thinking about what it is we want to do in the first place.
What Pixar really highlight is the idea that the data that informs your work isn’t always collected using APIs, geo-locators and invasions of consumer privacy. Data is simply collected information about a particular topic. And that can be photos of rubbish dumps or scientific research papers or universal truths which humans relate to.
We need to stop jumping on the big data bandwagon, assuming we’ll get to the core of our problems with the manipulation of easy to extract numbers. We must start with the real challenge, identify what it is we really need to know, and think more laterally about how we can get to the answers – whether that’s big digital data or not.