What happens when a huge number of people share a single grocery store loyalty card?


Many years ago, one of my colleagues interviewed a candidate for a position in his group. The candidate worked with the data generated by a grocery store's customer loyalty cards. (Today, we would use the buzzword Big Data, but back then, it was just another thing.)

Most customer loyalty programs allow you to enter your phone number at the register to identify your account, in case you forget your card (or simply don't feel like carrying it around with you everywhere). Many Microsoft employees use the phone number of the main Microsoft switchboard. We treat it as a community account. Everybody gets their discount, and once in a while, somebody randomly wins a discount prize when the number of points reaches whatever target level is currently set by the program.

The candidate said that these massively shared accounts are easy to detect and filter out from their data analysis. What messes up their data analysis is when two people with different lifestyles swap cards. The system sees that somebody who used to buy yogurt and bulk brewer's yeast is now buying potato chips and frozen pizzas, and it can't figure out what is going on.

Comments (26)
  1. Brian_EE says:

    Maybe I’m missing something, but why would two people swap cards instead of just keeping their own card?

    1. Richard Wells says:

      Happens by accident frequently. Someone drops a card and then picks up the wrong one. Without names printed on the card, telling the difference is challenging.

      1. Brian_EE says:

        The grocery-store cards here (Western NY) have your name embossed on them, like a credit card does.

        1. Paul says:

          Are you talking about Buffalo Wegmans? :)

          1. Brian_EE says:

            Wegmans is a Rochester thing. We’re just nice enough to share it around other parts of the Eastern US. Kind of like how Buffalo shares its (generally) poor-performing major sports teams with us.

        2. Dave says:

          I know a guy whose name on his loyalty card is “‘); DROP TABLE customers: –“. Seriously.
          Oh, and his first name isn’t Bobby, that’s another guy.

          1. We call him Little Bobby Tables.

    2. Aged .Net Guy says:

      Roommates? SO’s?

  2. yadt84 says:

    Frustrating Big Data analysis isn’t a good enough reason?

    1. Peter Doubleday says:

      That’s a good point to raise. So I’ll try to answer it. (Ex-contractor to Bing here.)

      We’ll have to narrow down “big data analysis” to “collecting and analysing a large number of data points about a single individual,” which I suspect is the variety you propose to frustrate. A billion data points on cancer research? I’m assuming you’re OK with that.

      I think we’re looking at Big Data as a surveillance technique here. If you worry about these things, you’re probably more worried about Big Government at the top level of danger, or maybe Facebook/Bing/Google at the next level, but in all honesty I can’t see why you’d want to frustrate Safeway. (Frankly I’m not that worried about the top two levels, but this is a personal view.)

      If you frustrate Safeway Big Data analysis, all that happens from your immediate point of view is that you don’t get a 50¢ coupon off your next purchase of Folger’s, because Big Data said you don’t need to be encouraged to buy Folgers, you need to be encouraged to buy a pair of Jimmy Choos. (Or whatever.) It isn’t obvious how this helps you. Or, indeed, frustrates Safeway, although they might have to discount that pair of Jimmy Choos in order to sell it to somebody else.

      And from the general Safeway point of view, all that happens is that they restock things they won’t sell as fast as they thought they would, and understock things that people actually want to buy.

      Saying Boo to the Man might make you feel good, but I wouldn’t take it to absurd levels, really.

      1. Joshua says:

        I went out of my way with loyalty cards to “frustrate big data”. What I actually did was ensure a few swaps shortly after creating them and then kept using the same one for years and years. Actual objective: disconnect the card number from my name and phone number.

        1. user says:

          The first time I went to a Safeway, they gave me a card with a form attached but didn’t ask me to fill out the form. I’ve been using the card ever since with no personal info whatsoever.

          1. Jonathan says:

            >> The first time I went to a Safeway, they gave me a card with a form attached but didn’t ask me to fill out the form. I’ve been using the card ever since with no personal info whatsoever.

            Though the minute you use a credit card to pay for a checkout linked to that card I assume they correlate the event and tag your loyalty account with the name from that credit card.

      2. Adrian says:

        “A billion data points on cancer research? I’m assuming you’re OK with that.”

        I’m not sure whether I’d be OK with that. It depends a lot on how the data was collected, how it’s used, how it’s protected, whether it links cancer data to individuals or other PII. But I’ll focus on the points you do raise.

        “I think we’re looking at Big Data as a surveillance technique here. If you worry about these things, you’re probably more worried about Big Government at the top level of danger, or maybe Facebook/Bing/Google at the next level, but in all honesty I can’t see why you’d want to frustrate Safeway.”

        The third-party doctrine essentially makes the business records of corporations easily accessible to government, often without warrants or notification, as in the government collection of cellphone call metadata. If you are concerned about government surveillance, you should be worried about letting your grocer keep track of how many coffee filters and Sudafed you buy. Or whether your tastes lean toward Middle Eastern cuisine.

        Of course, this will be sold to us as a feature rather than a bug. For example, if your grocer knows you bought some bad eggs, they can notify you directly about the recall. But having the grocers do this kind of notification can be expensive and inefficient, so we’ll just have the grocers share your purchase records with the government and have the government do the notifications instead. Don’t you feel safer already?

        “And from the general Safeway point of view, all that happens is that they restock things they won’t sell as fast as they thought they would, and understock things that people actually want to buy.”

        I’m pretty sure inventory management can be done based on aggregate sales that don’t require linking specific purchases back to individuals. You can even analyze which types of products are bought at the same time without privacy implications. I see no need to enable mass surveillance, profiling, and price discrimination to ensure that there will be enough coconut-flavored gelato in the freezer aisle.

        Frustrating big data analysis at all levels is the duty of all citizens.

        1. Peter Doubleday says:

          One can imagine an environment in which “all citizens” have a duty of some sort. My perspective (and you are free to disagree) is that such a duty, as applied to nebulous things like “big data,” is only really asserted by various groups in the USA. Which is fine and all, but the USA is actually less in danger of Big Government intrusion than, say, China, Iran, or even my favorite foreign government (France) or my own (the UK).

          Now, given ten really important things to worry about, which one would you pick? Quick! If you answer “Big Data perverted to government usage,” then I wish you well, and make sure you look out for the black helicopters hovering overhead at midnight.

          Not that it’s relevant, because my main point was — you can feel free to check this out by actually reading what I said — there are categories here.

          Big Data a la Safeway is only a threat to paranoid nutters.

          1. Peter Doubleday says:

            Honesty compels me to admit, however, that your concerns have forced me to check out the threats you suggest. And, do you know what? You are right, and I am wrong.

            Apparently, Safeway has acquired twenty three helicopters. (I hope they used Big Data coupons to save money.) Devious fellows, these Safeway guys: not only do they claim that the helicopters are used purely for the purpose of ferrying executives around, but they’ve also colored them bright pink, just to confuse Big Data algorithms that assume all such menaces are colored black.

            Now, you have to forgive me. The IRS have just driven up outside my house in a couple of Humvees.

            Why they chose pink Humvees is a mystery to me. Perhaps this is a sign of Collusion With Safeway Big Data?

            I’ll get back to you on that one.

          2. Adrian says:

            Name calling (e.g., “paranoid nutter”) isn’t consistent with the Ground Rules Raymond instituted for comments on his blog. Also, naming specific brands (e.g., Safeway), especially in a negative light, is frowned upon here, which is why I didn’t include links for the events I referenced.

            I did not write about conspiracy theories. I referenced a few of the bad side-effects that actually have occurred as a result of massive collection of personal information and from drawing inferences from those database, both by private corporations and by the government in collaboration with those private corporations.

            The social norms of how acceptable big data collection and analysis has become have transformed substantially in an astoundingly short period of time, partially because the folks asking questions about the ethics of it are dismissed as “paranoid nutters.” Just a couple decades ago, more than half of all residential telephone customers in my state paid a monthly fee to be kept out of the phone book. (Clearly they were all nutters!) And the phone companies had to get your explicit permission to look at your calling history (also known as their business records) in order to see if they had a plan that could save you money. And Congress passed a bill to protect our video rental histories. And a bookstore famously went to court to resist a federal subpoena requiring the purchase record of a single customer.

            I’m not talking about black helicopters. I’m talking about monitoring of private communications by the government that bypass the usual checks and balances–possible we gave the government the right to claim business records without a proper warrant and judicial review. I’m talking about destructive drug-enforcement raids of private homes based on nothing more than high electricity bills and the time of day your put your trash bins at the curb for collection. I’m also talking about corporations and the government (often unintentionally) discriminating against groups of people under the guise of statistical algorithms that rely on purchase histories and credit scores among other inputs.

            Yes, this is in my top 10 list of things to worry about. If it’s not in yours, fine–we can have different priorities. But calling me as a “paranoid nutter” is uncalled for and inappropriate for this forum.

      3. They shouldn’t be reliant on loyalty card data to predict stock, that prediction should be tied to the registers/barcode/inventory system. That won’t help if management screws up though, a local store in a national chain (that has a loyalty card and coupons) here constantly run out of stock of heir own branded things and have been doing so for a few years now.

    2. yay you says:

      Realistically, isn’t the only outcome of “frustrating” big data that you just get more poorly targeted ads? Yay you. The volume of ads you receive probably will remain unaffected. Nor are you likely to make the analysis become systematically ineffective, what with a probably small number of bad data lost as noise in the ocean of big data.

    3. morlamweb says:

      Along the lines of “frustrating Big Data” (how an amorphous concept can become frustreated is another matter), I remember a time when I was regularly asked for me Zip code at retail checkout lanes. This wasn’t to verify my credit card info, as I still sometimes see when I (very infrequently) stop for gas. This was an attempt on the part of the stores to gather demographic information. If the checkout clerk didn’t take my polite “no, thanks”, then I answered “90210”. Either they got the hint and hit the “decline” button, or they dutifully entered that Zip code into the system; either way, they’re not getting my Zip code. It was completely unnecessary to complete the transaction. And no, I didn’t have the loyalty cards at the time.

      Nowadays, the only time that I find myself in the checkout lane is for the weekly grocery shopping trip, and I do have the store’s loyalty card. I’m sure that I gave them some information when I signed up for it, but I don’t remember what they asked for on the form. I’m far less concerned with that stuff now.

  3. Ndragonawa says:

    You can use Jenny’s phone number if you don’t have a loyalty card:
    “your local area code” – 867 – 5309

    From what I have seen, there is a good chance that someone used that number when setting up their account as a prank.

    1. Brian_EE says:

      OT trivia – Legend has it that back in the say when that song was popular, (716) 867-5309 belonged to the daughter of the Buffalo, NY police chief. She received a lot of prank phone calls. (see http://hudsonreporter.com/view/full_story/14791420/article-867-5309-won%E2%80%99t-lead-to-Jenny-After-30-years–famous-phone-number-still-generates-calls–profit-)

      Somewhat topic relevant – these days, Tommy Tutone works as a software developer. At the time of a VH1 show (I forget which one) he was featured in, he was doing UI development for educational software.

      1. Rick C says:

        For many years, there was a “project” to call Jenny’s number in every area code and list the respondees.

    2. Joe says:

      I tried this at CVS. Man, the coupon roll that gets printed out is gigantic.

  4. Rob K says:

    Once upon a time, there was a certain retailer who wanted to allow non-unique 7 digit alternate IDs, since they operated in many, many area codes. The system would return the entire set of matches to the point-of-sale so the proper account could be selected from them, Seems like a good idea, right?

    But many people don’t like to give their real phone number– many, many, many people. What’s an easy number to remember?

    Jenny, I’ve got your number!
    I need to make you mine!
    Jenny, don’t change your number,
    867-5309!

    Bad things happened when thousands of matching records were returned to the POS.

  5. One thing that annoys me with some loyalty cards is that they store up the discount into a store chain account. I’d rather have that 5% discount there and then as I get ringed through the register.

Comments are closed.

Skip to main content