Skip to 0 minutes and 7 secondsLISA HARRIS: Hello, I am here today with Reuben, who's going to be talking about his research into personal data. So Reuben, what exactly is personal data?
Skip to 0 minutes and 17 secondsREUBEN BINNS: That's a tricky one because the regulators aren't sure exactly what it is, but the way I think of it is data that we reveal about ourselves as we use computers. Things like behaviour, things you click on, words you put into a search engine, but also more traditionally it's seen as things like your personal address, your name, your email address, your telephone number, things like that.
Skip to 0 minutes and 45 secondsLISA HARRIS: OK so all of this data that we're inputting online, what opportunities does this offer to marketers?
Skip to 0 minutes and 52 secondsREUBEN BINNS: Well so some marketers might be interested in knowing what kinds of things you might want to buy, and so that's something that they guess based on things that you do online. So it might be that you click on a certain sports page, and they decide to market certain sports products to you. So it depends on your interest but they'll use that to try and target messages to you, whether that's through adverts that appear on web pages or whether it's within a platform like Amazon, where they use your previous purchases to recommend new products to you.
Skip to 1 minute and 28 secondsLISA HARRIS: OK and as consumers then should we be worried about all this?
Skip to 1 minute and 32 secondsREUBEN BINNS: Well I think some people are worried about things like spam, things like intrusive messaging. I think that's certainly one thing that we should be worried about, but more long term there are other worries. So some people are worried that it's unfair if some people get shown certain products at certain prices based on their personal data, while other people might get a worse deal. Other people are more worried about their autonomy. So the idea that if your movements are being tracked and the environments that you see online are tailored to fit that, to fit the models that the companies have about you, people are worried that would mean that their autonomy might be undermined.
Skip to 2 minutes and 12 secondsThe things that they get offered, the options that they see, might be shaped by their previous behaviour in a way that means they can't choose things freely.
Skip to 2 minutes and 20 secondsLISA HARRIS: So are there any new services being developed then to help consumers navigate through these minefields?
Skip to 2 minutes and 26 secondsREUBEN BINNS: There are a range of new services that are coming out that basically try to make people more in control of their own data, and more in control of how information about them is revealed to organisations. And so some of them have models where you can say what information you want to reveal to certain trusted partners. So you might decide that you don't want to reveal your information to one company, because you don't like the products they offer or you don't like the way they operate, but you're happy to share it with another organisation. And they'll be able to develop a trusted relationship with you, and so that's one model.
Skip to 3 minutes and 1 secondPeople are also looking at models where the individual might be able to earn money from their data. In the same way that currently intermediaries are making money from individual consumers data. The consumers in this case would be able to get a cut of that money. And so that's one option, but at the moment people are unsure whether that's a genuinely empowering solution, or whether it might just be a way of encouraging people to share more data. And the same problems that I mentioned before, to do with discrimination or undermining autonomy, they might still be there.
Skip to 3 minutes and 33 secondsSo there's a question about how empowering these services really are, but a lot of them are making a lot of effort to try and empower individuals.
Skip to 3 minutes and 42 secondsLISA HARRIS: OK and are there any particular insights that are coming out of your research that's adding value in this area?
Skip to 3 minutes and 49 secondsREUBEN BINNS: So one thing I did was I did a study of consumers, how they responded to product recommendations, depending on whether they had control over the process or not. So in one condition, people had to reveal their behaviour about what they bought in the past, and that was used to target them in a traditional way that they do with current behavioural advertising. And the other circumstance, in the other condition, people were able to reveal what interests they wanted to reveal, and it turned out that the people who had more control of the process in that way tended to prefer the recommendations that they received.
Skip to 4 minutes and 26 secondsThey reported that they were more likely to actually buy the products that were recommend to them because they have more control over the process.
Skip to 4 minutes and 34 secondsLISA HARRIS: OK, thank you very much.
Controlling access to your personal data
In this video, PhD student Reuben Binns, talks about how his research looks into new ways for people to understand and control how their personal data is used.
Reuben has recently been awarded his PhD and he now works as a Researcher at the Computer Science Department of Oxford University on the SOCIAM Project.
This is becoming an increasingly important issue for digital marketing, due to the growth of what is called ‘behavioural targeting’, ‘targeted advertising’ or ‘web profiling’.
These are techniques used by online publishers and advertisers to tailor marketing material specifically to individuals based on data gathered about how they behave online. For instance, if you visit web pages or make search queries about a new pair of shoes, you may find advertisements pop up elsewhere on the web for that same product.
This all happens in the background; most of us are not aware it is happening and have no control over it. However, a new model is emerging which aims to give consumers more transparency and power over how they are marketed to.
Behavioural targeting may sound fairly harmless, but some of the marketing companies who provide it are beginning to sense danger.
The ‘uncanny valley’
To understand their concerns, we can make an analogy with a hypothesis called the ‘uncanny valley’ (which comes from the theory of robotics).
It says that we prefer robots the more accurately they resemble humans, but only up to a point; if they are too realistic, they become creepy (Mori, 1970).
Could the same be true for targeted digital marketing? As we have heard, we are segmented and profiled on the basis of personal data that we reveal through our online and offline activity - and there is evidence that a growing number of consumers find this a little bit creepy.
Some brands are beginning to worry that if their personalisation becomes too accurate, there could be a consumer backlash.
The answer may lie in making individuals the authors of their own consumer profiles.
Rather than building a picture of an individual’s preferences based on surreptitiously collected data, this approach starts with your self-declared interests and preferred brands.
This is the approach taken by nFluence, who won this year’s Data Strategy Awards mobile category.
Other emerging profile-building services include Handshake, DataCoup and YesProfile, which enable individuals to earn money from revealing their self-authored profiles to marketers.
Even if behavioural targeting is based on accurate data and statistics, it may never be able to predict the idiosyncrasies of real individuals, their unique preferences, intentions and circumstances.
Self-authored profiles also have the potential to reflect the individual’s view of who they are - or who they would like to be - rather than simply what their past behaviour might suggest.
In the end, this may come down to a question about the ultimate purpose of digital marketing:
One view is that it exists to help businesses sell whatever behavioural data suggests a user might buy.
Another view is that it is there to help individuals find products and content they want and need.
Each view suggests a different approach to digital marketing - one based on targeting people, the other on simply asking them.
Some key questions for this new approach to marketing are:
Do individuals know themselves well enough to compete with existing behavioural analytics? Might we introduce new forms of inaccuracy into the process?
Do consumers want to engage with these processes at all? What is the incentive to participate?
Would you prefer to ‘own’ your own marketing profile? Do you know your own preferences and needs better than your browsing history would suggest? Or could behavioural profiles be a better guide?
For further background information, this article in Marketing Week describes some of the emerging services which put individuals in charge of their own data.
After reading this article, what are your thoughts?
Mori, Masahiro (1970). ‘The uncanny valley’ Energy 7.4 p33-35.
© University of Southampton 2016