We are not just one person. Well, what I mean is that I am many people in one person. OK – I need to take a moment to explain.
One of the challenges of personalisation is that each of us combines many moods, modes and contexts in a single human being. My mood differs from one day to the next as I move from happy, to concerned, to stressed. My mode can change faster: on a Friday morning, I’m a businessman; that evening I’m a father; the next morning a cyclist; and that evening a husband or a friend. And my context might switch from interactively using my mobile on the train to passively watching TV content on my tablet at home.
It’s difficult enough to get personalisation right for a single person. But when that person has multiple moods, modes and contexts, the challenges of getting personalisation right at that moment multiply very quickly. And we all have many such states. We all ‘have our moments’.
This is where machine learning comes in. Looking at very large data sets and identifying appropriate outputs not just for the individual, but for the individual’s mood, mode and context, produces a huge leap forward in the value of personalisation. Presenting something that is relevant to ‘generic me’ is valuable. Presenting something that’s specific to my mood, mode and context is invaluable.
That’s why The Filter started out 14 years ago using machine learning for music and media businesses. It’s what has driven our success in working for some of the world’s largest media companies, from: NBC, Vudu and Warner Brothers in the USA; BT TV, UKTV, Orange France and Nokia in Europe, to SBS in Australia. And it’s the application of machine learning in this way that has made us successful in developing our highly effective online retail offer in the UK and Europe in the last four years.
In the last couple of years, machine learning has become a buzzword across media and online retail sectors. We’re excited by the prospect of more companies recognising the transformative effect that machine learning can have, happy that more people are going beyond the buzzword to look at the substance beneath. It’s our long track record that enables us to use machine learning to get closer to the right recommendation or action for a combination of mood, mode and context, so we move the definition of personalisation beyond the person alone.