Get a demo Do I subscribe? News sign-up
Print

Brands risk 'creepy' personalisation

News, 13 June 2016

LONDON: Amazon recommendations, Spotify Discover Weekly playlists and Google's Content Delivery Network are doing something profoundly different to previous marketing, according to the winner of the 2016 Admap Prize.

The topic for this year's Prize was: 'How Should Marketing Adapt to the Era of Personalisation?'. Admap also will be hosting a discussion on this subject at Cannes with Marc Mathieu (Samsung Electronics America), Guy Murphy (JWT) and Steve Hatch (Facebook UK & Ireland) among the panel. Readers can register to attend here.


Unfortunately, few brands are approaching personalisation with the level of expertise shown by the likes of Amazon, says Oliver Feldwick, Head of Digital Strategy at CHI & Partners.

"The promise is still a long way away from the reality," he notes in his Gold-awarded essay, The Uncanny Valley of Personalisation.

Nor are consumers as excited by this world as marketers are, not least as most will have had experience of poor personalisation, from being followed around by ads for products they've already bought to downright intrusive and weird targeting.

"In the rampant march towards marketing automation, Big Data and personalisation, we have lost some intuitive understanding of what consumers actually want out of all this," says Feldwick.

He argues that marketers are stuck in an "uncanny valley" – a reference to the term used in Robotics when attempts at getting something to look and act like a human end up having the opposite effect.

"As robots get closer to being human, they become creepy, profoundly unhuman and unlikeable … That's what's happening in marketing personalisation."

One way to avoid this, Feldwick suggests, is not to try and hide the fact: "Wear your tech with pride and show your workings."

By putting some distance between a brand and a personalisation engine – 'we think you may like' at least admits some room for error – "it shields the brand from the negative impact of getting it wrong".

Feldwick further notes that current efforts at personalisation omit opportunities for serendipitous discovery.

"The best examples of personalisation marry human insight and oversight, with the power of data and algorithms," he says.

The Spotify Discover Weekly engine, for example, uses real people's tastes and insights. Each playlist contains songs discovered by a human, but with an algorithm joining the dots weekly for 75m people.

Data sourced from Admap