Behavioural data


Brands like Waitrose are using behavioural data to target users who visit their websites to consume content — such as that affiliated with chef Heston Blumenthal, below — but don’t shop there.


Behaviourally-targeted ads often get the brunt of the blame for the “uncanny valley” effect when it comes to online advertising. There’s something unsettling about browsing for a holiday cottage, say, and then being bombarded in sidebars with holiday cottage websites. Often behavioural ads don’t get described as using personal data because they have no recourse to your name, age, or sex; however, it undeniably feels like a kind of personal interaction, and can create ill-will to the extent of alienation. Farhad Manjoo, writing for Slate in 2012, said,

“Today’s web ads don’t know enough about you to avoid pitching you stuff that you’d never, ever buy. They do know just enough about you, though, to clue you in on the fact that they’re watching everything you do.”

So is part of the problem with behavioural advertising that it’s not personalised enough? By simply tapping into what people are searching for and taking into account no other information it’s bound to be a hit and miss affair, traditionally assuming you want the thing you’ve been browsing. But looking rarely translates simply to buying.

As we become more adept at translating online behavioural data, the more successful brands can be in this area. Recently British supermarket brand Waitrose has used behavioural targeting in its digital strategy, aiming at those consumers who visited the website but did not shop there. It looked particularly at browsing history, paying attention to the days and times that people visited the site, and then sent out staggered incentives designed to appeal specifically to that market. The next step for Waitrose now is to gather more data to be more effective at pinpointing when to send out its offers — and it plans to do that by pooling all data across its parent company, the John Lewis Partnership. So you might buy a fridge from a John Lewis department store and then receive deals from Waitrose enticing you to put something in it.

There is also a realisation that behavioural data doesn’t merely constitute what you browse online. Brands have generally become aware of the importance of social media in promoting and controlling their messages, but have been less able to make use of the opportunity to treat Twitter, say, as a tool for customer insight. Using Twitter to predict trends has become news for politics, the stock market, and even criminal activity. But as the tools that can analyse the meaning of a tweet become more sophisticated, there’s a greater chance to use that information accurately in marketing. The key behind this development is sentiment analysis — as natural language processing gets better and better, so too does the understanding of nuanced speech that we find on social media. Until a computer could detect sarcasm, for instance, it had no chance of accurately gauging public opinion via social media. (Media research company Dollywagon gives a good example of how a computer does make sense of sarcasm on its website here.)

Social media can provide a vast and rich seam of behavioural data, and this is beginning to be utilised effectively, particularly in the entertainment industry. Satellite TV Provider Sky Italia has been working with IBM Big Data and Analytics to monitor the social media communications of its customers, and then using the information to tailor their services and offer the right incentives to maintain loyalty. In an age when a lot of people use their second screens to chat online about what TV programme they’re watching, social media is heaven-sent for behavioural insights.

The alternative to collecting online data from social media is to make your own data, and this is an approach that the motor industry has been embracing. One in five new cars are already in constant connection with the internet, and provide information to brands about the habits of their drivers. This is a step up from the type of behavioural insights that can be gleaned from what people are willing to reveal online — Richard Weston, writing for Business 2 Community, notes that,

“For the first time, we’ll be able to see what consumers actually did, meaning that we’ll no longer have to rely on claimed behaviour.”

Such tailored data will be of huge benefit for many reasons, from saving lives by predicting driving styles and possible incidents, to even lowering the insurance premiums of those who drive safely, if the data becomes available on a larger scale.

There’s a lot of concern and fear about the holding of behavioural data to overcome; 2013 research by Guardian Media Network/GfK showed that 69 per cent of consumers found the way brands hold information about them to be “creepy”. Colin Strong, Managing Director of GfK NOP Business and Technology, says on RW Connect,

“This has to be a serious concern for advertisers who often assume more personalisation can only ever be a good thing. The early signs from our research clearly indicate that this is not necessarily the case.”

But it looks as if, as our understanding of behavioural data improves, we are reaching a time where brands will be able to finely tune their marketing so that it no longer causes that uncomfortable feeling in consumers. By targeting different sources of data and translating it into an intelligent response, that uncanny valley feeling might just become a hiccup on the road to effective behaviour-based advertising, rather than a dead end.


This piece was written by Aliya Whiteley and originally published on brandperfect.org.

Show your support

Clapping shows how much you appreciated Brand Perfect’s story.