Check out some of the videos here. My personal favorite is the one titled “Out of Reach” in which a mother tries to balance a child on her head so he can grab an item on a high shelf for her. The overall effect is more sad than comic. It is clearly frustrating to have to resort to gymnastic and child labor to try to buy something. Even when a male customer comes to her aid, he has to climb on a lower shelf in order to reach the item.
Now is it bad that the store has this video? If you were told that the store had surveillance cameras in order to prevent theft but when they happened across a mother going all Flying Wallenda to reach a printer cartridge they used that information to change how the store was set up, would you find that inappropriate? What if the whole point of the cameras was to track how shoppers reacted to the store? Would that change your opinion of the store?
According to the New York Times, the latter is, in fact, happening (In Bid to Sway Sales, Cameras Track Shoppers, Mar 19).
The most basic surveillance setup has been around for a few years. It uses video cameras in ceilings and sensors near fitting rooms to learn how many customers pass through the doors and where they tend to go. At the other extreme, some retailers are taping shoppers’ every movement and using specialized analysis to study the shoppers’ behavior. For example, after seeing scores of customers struggle to navigate a particular area, analysts might suggest that the retailer widen the aisle. The companies that install and analyze video for retailers say that they are sensitive to privacy issues but that the concerns are overblown. They say they are not using the technology to identify consumers but to give them easier and more enjoyable shopping experiences. And, they added, they have the sales results to prove it.
So what kind of results do they get? Cisco, one of the firms selling these systems, gives the example of Cabela’s, a seller of outdoor and hunting gear. They used cameras to determine how long it was taking to for sales staff members to approach shoppers. Staff members were taking too long to assist customers, but Cabela’s was able to reinforce the policy and with that see a sales improvement.
So does this example suggest that privacy advocates have a misplaced concern? Are these systems for spying on employees as opposed to customers? The real bugaboo is facial recognition. These systems apparently can currently recognize a face enough to classify a shopper as old or young. That, the consultants say, should allow retailers better understand the demographics of their clientele. What concerns privacy advocates is that at some point technology will allow the retailer to identify and track individual customers.
I am not sure if this concerned is misplaced or not. Presumably, the firm could use this information to make special offers based on what they know about the customer’s shopping patterns and behavior. They do that already with loyalty cards that customers willing fork over. Is this all that different? One issue might be that once you have a visual image, the firm could base offers on some factor like race that might be of questionable legality. I suspect that most sensible retailers would have offers that can depend on pretty much anything but race. The other major distinction is that customers opt into loyalty programs. A video based system is much more passive and there may be no easy way to opt out.
A bigger issue might be what is the legal status of the purchase history. Does the firm have to turn someone’s shopping history if the police come knocking? Again, there is a question of whether this is different from information gathered from loyalty cards. Customers willingly let supermarkets and warehouse clubs track there purchases. And that can be helpful to the government. The day after the New York Times article ran, the LA Times reported that the for the first time the Center for Disease control tracked the purchases of the victims of a salmonella outbreak via their loyalty cards (Supermarket cards help CDC trace salmonella outbreak, Mar 20).
The first case in the salmonella outbreak was reported last summer, and by November, CDC investigators were examining a multistate cluster of cases. Through interviews and questionnaires, investigators suspected some kind of Italian meat was the culprit, but people couldn’t remember what brand they bought, [CDC epidemiologist Casey Barton] Behravesh said. So the CDC asked supermarkets for certain buying information on seven victims in Washington state, focusing on suspect products rather than everything the customers had bought, Behravesh said. “We didn’t care about the brand of toilet paper people were buying,” she said.
Now it should be pointed out that the victims gave their permission and CDC investigators didn’t go through all of their purchases. Rather, they asked, say, Costco “What kind of salami did he buy?” This all sounds benign and it is hard to imagine that a poisoning victim would not be willing to share their information if it could keep others from falling ill. But again, this is an opt in system. You know when you go to Costco that you have to show your membership card and you probably shouldn’t be surprised that they save that data. But what if participation in these systems was forced? If the goal is to track all possible cases of food poisoning, for example, then we shouldn’t allow any anonymous purchases. But then do I have to turn over that data to my health insurance company? Do they have a right to know just how much ice cream my household consumes?
It is obviously a slippery slope. Yes, watching for operational failures in the store seems reasonable and even laudable. But how far to take that seems rather unclear.