Reverse location targeting.. scary!

I attended a talk by Prof. Sandy Pentland at the Yahoo! Mission College campus about “Honest Signals” – a term for the unconscious responses we all have, which have been discussed by many people including the oft-cited Blink by Malcom Gladwell.

The difference with Prof. Pentland’s work was that he was essentially taking the theory that we are creatures of habit, and using bulk data to predict behavior. For example, in one study he tracked the location of students over a period of time, and was then able to very accurately predict which students were friends, what their preferences were, and based on past behavior as well as the past behaviors of others what they were likely to do in the future.

Of course, has had “people who liked X also liked Y” for almost as long as it has existed – but the difference here is that a much larger amount of data can legally be gathered without your knowledge or consent. If a city (or marketing agency, carrier or government) tracked cell-phone identifiers and built an aggregate pattern of location they would be able to quite accurately state that people who go to X then Y usually then go to Z – remember – this may have nothing to do with where you live or where you work. These social factors, the most often used marketing metrics, can be completely ignored for this kind of targeting – which may be frighteningly accurate and timely.

So, can we opt out? Prof. Pentland is asking companies to sign up to a new kind of data protection charter – but really this is just for identifiable data. For anonymous ethereal information ownership is irrelevant. Anyone can see it and anyone can use it.

This is not targeting users based on their location – it’s using location to target users by common behavior.

People who go to these bars at these times often end up in hospital. What does this mean for your insurance? What if your employer owns, and monitors your phone? It’s a new world once this kind of data can be collected and processed at scale. And it’s coming. Either buy more terrabytes of storage, or throw away your cellphone and wrap your head in tinfoil.


2 responses to this post.

  1. Alex – you should check out SenseNetworks, they use implicit (via operator) and explicit consent to location info to create very rich user profiles.

    The benefit and the challenge with mobile is that (contrary to web and PC) a LOT of data is gathered implicitly. Plus we are giving away more and more personal info in exchange for convenience and services (Facebook, etc). Plus we are in the midst of a culture change; the younger (15-20) generation is almost indifferent to privacy.

    Scary if you ask me. In a world of opt-in-by-default we need an explicit opt-out blanket policy and regulation.



  2. Yes – Prof. Pentland was using Sense Networks as an example – but their taxi-system instrumentation rather than their mobile integration.
    Your point about the younger generation becoming indifferent is a good one too.

    Most mobile users think their phones know where they are at all times – and when you ask for permission to “use their location” they often assume that means at all times, at street level, even if the phone’s turned off! Therefore, I used to be very concerned about asking for location unless we absolutely needed it, but after a tour of the location services on display at the mobile world congress this year we had a meeting with our policy team and the response was to just go for it.

    The thing that worries me is the use of the anonymous data – as this cannot be undone. Let’s see what happens when it starts affecting our insurance premiums..


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: