Personalisation has become one of retail’s favourite buzzwords, but how far can it go before it’s just creepy? Retail Week investigates.

Personalisation has become one of retail’s favourite buzzwords, but how far can it go before it’s just creepy?

Personalisation offers enormous potential for retailers. The sky is the limit – consumers might, for instance, walk into a store to be shown a map of the shop on their mobile, highlighting the products they’ll need to cook dinner. The shopper hadn’t decided what they wanted yet, but the supermarket’s data on what they’ve eaten for the past five years means the retailer knows before the consumer does.

When the shopper gets home, an Amazon package is waiting. They had only ordered the package yesterday, but Amazon’s data on consumption patterns mean the package was dispatched three days ago – it knew before the consumer when they would buy.

These are the more extreme examples of personalised service and advertising, and it’s as yet unclear whether this will become the norm in retail. But everyone from Shop Direct to Tesco and John Lewis is putting personalisation at the heart of their business models.

Shop Direct chief information officer Andy Wolfe said earlier this year: “Personalisation is big on our agenda. We’re investing heavily in data science and analytics and will use that to drive what you see on the site.”

Retailers are careful about how they approach personalisation because nobody wants to annoy customers. They are understandably loathe to push it too far, and every retailer working on personalisation or loyalty projects is quick to point out the measures it is taking to protect shopper privacy and ensure no-one is made to feel uncomfortable.

How personal is too personal?

But as competitive pressures lead more retailers to push the envelope and come up with new ways of serving customers, knowing exactly when to stop will be valuable.

It’s a question every retailer will need to ask – how personal is too personal? No-one wants to be the creepy retailer that makes shoppers feel uncomfortable, so at what point should personalisation stop?

Guy Champniss, associate professor of marketing at Henley Business School, is researching this question, alongside Colin Strong, managing director at data company GfK and Dr Kiriaki Koutmeridou of City University in London. They are using the concept of the ‘uncanny valley’.

The uncanny valley is an idea that was developed in 1970 by Japanese robotics professor Masahiro Mori as a result of robotic creations becoming more like humans.

Mori discovered that humans relate to and welcome robots that look a little bit human, or that have human characteristics. But at a certain point, robots can look too human – they become too close for comfort, and people start to find them unnerving. Mori referred to the moment when the feeling of affection turns to revulsion as the uncanny valley – so-called because people cannot quite put their finger on what is so wrong with it.

Champniss says: “We started to wonder whether the same disastrous effect could be lying in wait for those brands that increasingly personalise their advertising in their attempts to get closer to their consumers. We were interested in exploring whether there is an uncanny valley of the brands.”

Psychological effects

Additionally, Champniss and his team wanted to examine the psychological effects on consumers of knowing their movements were predictable and behaviours mapped out. He says: “We thought personalised marketing was being presented as a win-win for everyone.

But anecdotally we heard of people getting unnerved or annoyed… it’s really quite personal information and shoppers can’t always work out where it had come from and whether they have given permission.”

He used a marketing testing platform to show a random selection of people in different scenarios – some showed more traditional marketing, such as television adverts during popular TV shows, while others showed newer marketing methods. These included initiatives that used eye-tracking technology that registers where a shopper looks on an interactive shopping wall. If she looks at decadent products such as luxury ice creams or bubble baths, the system might assume she is tired and wants a rest. That data can be sold to a holiday company, and 24 hours later billboards might show her ads for holidays. If the data suggests someone has an ailment, that information could be sold to a health insurance company.

The initiatives shown were on a sliding scale, becoming progressively more targeted and personal and people were asked how the schemes made them feel.

“Hyper-personalised advertising takes away part of our personality”

Guy Champniss, Henley Business School

Champniss found that there was indeed a drop-off point between finding personalised advertising helpful and finding it disturbing. He adds that this point tends to come somewhere between targeted advertising, by which he means advertising based on online behaviour such as browsing and buying data, and ‘traded’ advertising where data has been collected and traded between companies. One real-world example of that is Google’s use of data gleaned from people’s emails to advertise products to them – he uses the example of someone emailing his sister to discuss moving his mother to a nursing home, only to be targeted by ads for nursing homes the next day.

“Up to a point, people seem to like personalisation. But very suddenly it tips over the edge and people are really anti-advertising – there’s a strong sense of revulsion,” says Champniss. He adds such revulsion is not felt towards the product being advertised, but towards the brand doing the advertising.

Revulsion isn’t the only feeling that highly targeted advertising engenders. It also leads to a type of helplessness and a drop in creative spontaneity.

“You see creativity dropping because it’s a realisation that our previous behaviour could let someone know what we are going to do before we do it ourselves. When you realise you can’t change something, you stop trying to.”

Champniss adds that while some shoppers will fight against a system that attempts to predict their behaviour and interests, others will conform.

“Hyper-personalised advertising takes away part of our personality,” he says.

The point at which marketing becomes too personalised is likely to vary between categories and even between brands. Brands that are more trusted can probably push the boundaries more, although they have further to fall if things then go wrong.

Champniss’s team is now working with brands and loyalty companies on the specific point consumers’ reactions will turn from welcoming personalised offers and services, to feeling annoyed, confused or helpless.

“We are looking at whether there’s a different effect in different sectors,” he says. “My sense is that there may be. If it’s a brand you really feel comfortable with, it will engender greater trust. For big, monolithic brands there will be associations with Big Brother.”

He adds that brands can do their own work to discover the crucial turning point. “There could be an early warning system you could put in place – if you know how your customers view your brand, you might get a better idea of how much tolerance and leeway you have. But almost every advertiser could run against this problem at some point, and we don’t know what it would take to get customers back.”

Personalisation has become a crucial part of retailers’ response to the changing trends in shopping – as consumers enjoy more choice and low prices, it’s a key part of ensuring your brand stands out and maintaining relationships with consumers. But caution is clearly needed – data sharing and predictive analytics could spook customers more than they help them.

Uncanny valley

The uncanny valley is an idea that was conceived in 1970 by Japanese robotics professor Masahiro Mori, who worked in an environment where robotics research was advancing quickly, and robots were beginning to look more human.

Mori discovered that robots with slightly human characteristics, but which are still visibly robots in some way, are welcomed by humans – C-3PO in Star Wars is a good example.

But at a certain point, robots can look too human – they become too close for comfort, and people start to find them unnerving. Mori referred to the moment when the feeling of affection turns to revulsion as the uncanny valley – so-called because people cannot quite put their finger on what is so wrong with it. Researchers say a similar thing could happen with personalised marketing.