• by Sean Hargrave , Staff Writer, April 11, 2016

Is it a bit too spooky or an amazing tool to help solve the puzzle of whether digital advertising truly drives in-store sales? That's the question on everyone's mind upon hearing that Facebook has made a deal with Dunnhumby to monitor cross-channel whether social media advertising results in in-store sales.

How the nuts and bolts of the deal will work has not yet been fully elucidated, but this appears to be a combination of combining in-store and online data to show whether Facebook advertising -- and presumably Instagram -- can have an impact on the bottom line when a sale is not completed online. 

Strangely enough, I was talking with Acxiom the data specialists the other day. They are involved in the deal in a role that would appear to be a third party that crunches through the information and ensures that Facebook and Dunnhumby don't mark their own proverbial homework. It wasn't about the Facebook deal, but was all about how they're working with retailers to combine offline and online data to get a better idea of the customer. I think a better definition than offline and online, is first-party and third-party data, because it's basically a case of what the advertisers knows combined with what a data company and a tech company know when a retailer's customer logs on to social media or Yahoo or Google. Some of the unknown data will be offline (address) and some will be online (log-in identity).

There's a lot you more additional information there -- it all has to be anonymised, but can still help brands target their media spend to specific groups and demographics. A retailer might not know who someone is online because they haven't logged in to their store account -- but could with third-party data know that they are the gourmet food-loving type and is interested in fine wine, for example, and that could influence whether they choose to target them through display or a promoted post. 

The trick is making it work the other way around so you tie up how many social campaigns are placed with activity at the cash register. The key thing here is advertising data will have been anonymised yet somebody in-store has to be recognised. Although neither party has stated publicly how the system will work, it is fair to assume that it will normally rely on the shopper identifying themselves through a loyalty card. Without that, there is no legal way of tracing somebody. If the person was choosing to pay through a loyalty app somehow, they could be identified, but other than that, you have to be talking about loyalty cards. So, when someone hands in their loyalty card this is where the data trail is presumably followed back and the retailer can establish whether they had recently suggested that consumer should buy a certain product, through a promoted post on social.

I'm no legal eagle, but I'm pretty convinced that a lot of people will get upset about their actions at the till telling the retailer who they are and what ads they've been shown. Deciding whether to place the Facebook ad or promoted post or not is very straightforward -- and you can see how it works, despite the person being anonymised. Working backwards from an in-store sale to advertising that has been based on cookies, rather than an individual, is less clear.

Also at the back of my mind is the tighter data protection rules that are on the verge of becoming EU law. These insist that brands get informed consent on their use of data and so retailers are going to have to treat very carefully.

As details of what Facebook and Dunnhumby are up to begin to emerge, I can only imagine there has to be a very clear line of informed consent. As I say, I'm no lawyer, but it can surely only work if someone is identified themselves on social media and consented to cross-channel identification through that store's loyalty card. It's hard to imagine how it can work unless each customer tracked is a fan of the retailer on Facebook, uses a loyalty card at the till and has ticked a privacy box to say they don't mind linking their social activity to the loyalty card. This could no doubt be presented as a shopping club with the best offers and point-collecting options.

So, surely this can only have a limited remit from the outset, and quite how it will grow beyond tracking committed customers to winning over rivals' shoppers is a mystery to me. If you wanted to track back someone from presenting their loyalty card at the cash register to having received social media advertising, then presumably that's possible if the Facebook account and the loyalty card email account are for the same email address.

Either way, you're surely limited to measuring the effect of advertising on people who already like you enough to have signed up for a loyalty card? They may have found a way of extending the remit but, legally, I really can't think of one.