Measuring ad effectiveness by linking to offline sales

By September 24, 2012 2 Comments

On Friday I wrote about how privacy advocates will welcome Facebook’s release of their Shared Activity plugin which makes it easier for users to control how their actions on Facebook-connected third party site show up in their Facebook feed. Today’s news points in the other direction. This morning I read of Facebook’s project with Datalogix to measure ad effectiveness by tracking whether people bought a product in a store after seeing a Facebook ad. They are using loyalty card data to make the link.

Predictably, privacy advocates are arguing that if Facebook is to go down this route they should only do it for consumers who explicitly opt-in.

These opt-in vs opt-out arguments are happening all the time now, and in some ways they are a bit of a charade. Very few people change their default settings and so rather than being about personal choice, deciding to make something opt-in is to effectively kill the project – i.e. to regulate so that something is opt-in is to regulate it out of existence, most times at least.

For me the advantages of linking online ads to offline sales far outweigh any risks. Better tracking leads to better targeting, allowing publishers to charge higher rates and show us fewer ads and/or offer us more stuff for free. Additionally, the ads we do see are more likely to be for something we want, and therefore less annoying.

There is a lot to be gained from this type of tracking, which is why Facebook’s advertising customers are pushing them down this route.

And I can’t see that there is much risk. Datalogix anonymises the data it buys from retailers so there is no way that Facebook can tell which of their users are actually buying what. For me, real privacy risks come when people can work out how to access my money or work out where I live, not from high level concerns that if the data is in existence something bad might happen. I’d love for someone to explain if there are any real concerns in this case.