rot13($real_name)

December 16, 2007 7:35 PM

I'm going to get a 'customer loyalty card' from Bloom's (aka upscale Food-Lion) supermarket. The terms and conditions don't mention anything about the validity of the information, and I know enough about security and datamining that I don't want to distribute that information. Do they have a policy, or is there any way I could get caught using a fake name/address/telephone no, etc.?
posted by tmcw to shopping (36 comments total)

I doubt they will bother disconnecting your card and denying you participation in the program- if anything, they will catch it as an indicator for the overall cleanliness of their data. Most CRM and loyalty programs expect a certain amount of noise and dirt in their data.

However, are you sure that you are behaving ethically? They have agreed to give you a discount in return for tracking your purchases against your demographic data; you are reneging on that part of the deal. If you are going to fake the data, I think you may have the ethical responsibility to at least fake it with demographically similar data to your location.
posted by jenkinsEar at 7:40 PM on December 16, 2007

At my supermarket, the customer loyalty card is also a check cashing card. (That is, to pay with a check, you show your card and that takes care of all your identification.) I signed up for mine eons ago so I can't remember the process, but I'm sure it had a lot to do with my real identification. Which just makes me think that they must check that stuff.
posted by iguanapolitico at 7:46 PM on December 16, 2007

Just tell them you want a card but don't have time to fill out the customer information portion at the moment, but you will take it home, fill it out, and bring it back with you next time you come in the store. Then don't. I did this at Kroger's due to the same concerns you express, and my card always works fine. YMMV.
posted by jtfowl0 at 7:47 PM on December 16, 2007

I did the same thing as jtfowl0 at CVS and have been able to use the card for years. Food Lion cashiers (and therefore Bloom, I assume) are also really nice about using their "guest/manager" card if you ask.
posted by Sweetie Darling at 7:49 PM on December 16, 2007

I've done the same thing that jtfowl0 mentions four times - at two pet stores and two grocery stores. All the cards work fine and I've never had any issues.
posted by Ostara at 7:52 PM on December 16, 2007

I lied on my Kroger card form but gave them the right zip code and basic demographic info. The card works fine. I also used www.justoneclubcard.com to print the Kroger bar code on the same card as other loyalty card bar codes, and that has worked fine, too. I just tell the cashier which bar code to scan.
posted by PatoPata at 7:53 PM on December 16, 2007

When Albertsons introduced cards, I told the checker I didn't want to give any info, and she said I didn't have to. I got, essentially, a "blank" card. I got the in-store discounts, but they didn't (obviously, couldn't) send me any special coupons or anything.
posted by sageleaf at 7:55 PM on December 16, 2007

Sign up using fake data, I've yet to use real info for any affinity card and I've got a keychain full of them. The cashiers never blink even if I pay with a credit card that doesn't match the name on their records.

That, and it's rather amusing when the cashiers thank me, Ms. McShoppington, for my patronage.
posted by jamaro at 7:58 PM on December 16, 2007

I have a degree in food industry (retail grocery) mangement.

Out of the top twenty grocery chains with loyalty programs in the US at the time of my graduation (2004), only Kroger/Signature/Safeway/Fred Meyer/etc. was doing anything with data mining, and it was usually to drive same-store sales up by direct-mailing extra coupons to the homes of loyalty card holders.

Costco and Sams Club also do data mining with their membership cards, but that's more for purposes of deciding what to stock at a particular store.

Otherwise, you really have no need to worry about them doing evil things with your demographic data. My career is in IT; I got the degree because I thought that kind of data mining is interesting. But no one's doing it. The margins in the grocery industry are SLIM -- a max of 3% for the entire store after you count logistics and corporate overhead. That's why the little neighborhood grocery stores haven't been surviving, and why we don't have magical RFID grocery scanners that charge you as you walk out the door without scanning. Anyway, the margins are far too slim for the major grocery chains to actually make full use of this data -- but it's worth it to have the damned loyalty cards because everyone else has them and it lets them make an extra margin on anyone who doesn't have one. Plus, the aggregate data is useful in identifying marketing trends for the store.

Where they are doing demographic analysis, they're using aggregate data to determine things like how far people travel to the store, how much people in different neighborhoods tend to buy per transaction, and they mine and retain specific SKUs that they can get people to pay them to watch -- this is where the individualized quarterly Kroger mailers come in. There's no secret underground lair where evil capitalists plot to make you buy four extra boxes of cereal this month, honest. Most of the data isn't retained except in aggregate per-store or per-customer totals. It's literally terabytes of data per day that needs to be processed, and very few [i]research universities[/i] have that kind of raw data processing power ... much less a grocery store chain with a 3% profit margin.

It's really quite benign, where it's used at all. There's way more benefit to letting them mail you promotions than there is risk in telling them where you actually live. I say this not only as a paranoid IT guy, but also as a paranoid libertarian-leaning centrist who loves 1984 and V for Vendetta and hates big brother. Just drink the kool-aid, it won't kill you.
posted by SpecialK at 7:58 PM on December 16, 2007 [6 favorites has favorites]

I also did the same thing as jtfowl0 at my Food Lion, which later became a Bloom. It will be no problem.
posted by bristolcat at 7:59 PM on December 16, 2007

When you pay with a credit card, they then have your demographic information - or at least a unique identifier and name. It doesn't matter too much what you gave them on the slip of paper that the cashier probably threw out after you left.
posted by kcm at 8:00 PM on December 16, 2007

SpecialK - There is no reason to think they will not do something with the data in the future, and absolutely no reason to use real information in the first place.

So even discounting any paranoia, what is the benefit to using real information? What is the harm in using fake information?

Check cashing was mentioned by iguanapolitico. Does anybody actually do that anymore? If so, why? Is there any reason to use checks for in-person payments anymore?
posted by bh at 8:05 PM on December 16, 2007

So even discounting any paranoia, what is the benefit to using real information?

What Kroger's said years ago was that if you lose your keys with the card attached to the ring, and someone finds them and turns them in at Kroger's, they'll use the information on file to mail the keys back to you.

I didn't find that compelling enough to give them real info.
posted by dilettante at 8:18 PM on December 16, 2007

bh: It's not that they are going to do something with it in the future -- it's that they can't afford to keep it around to do something with for very long. Even if you reduce the information down to the bare minimum -- a list of SKUs attached to a transaction number + time/date attached to a loyalty card number -- and compress the shit out of it three ways from sunday, you're still facing 20-100 Kb per transaction. A typical 18-hour a day Kroger in a typical neighborhood will do well north of 10,000 transactions per day. There's 2,500 Kroger Co. stores nationwide. That's 7.5 TB or so per day. Of raw data. Not processed into records in a database -- raw data on drives.

Just to give you an idea of how much storage that is, my university has a very nice IBM Power5 supercomputer purchased a couple years ago for an ungodly sum. [b]It has 7.4 TB of storage attached to it.[/b]. Of course, storage is rather cheap these days, you can get a nice Sun 'Thumper' X4500 with 48TB of raw storage -- Can't remember what that is in usable TB under ZFS, but it's something like 35TB -- for 'only' $45k. To get something like Bizgres, which is a MPP version of Postgres, they're selling appliances that can work with about 100TB of data for $250,000.

Keep in mind -- 3% margin in a cutthroat market. 7.5 TB of data per day.

And that's just raw purchasing data -- much of the work to process stocking info is done locally in a store level database so the store knows what it has on hand, and then it's squirted up to logistics central to be processed into restocking orders. Purchasing and logistics is by far the heaviest user of IT-centric data in the grocery industry. Marketing is lagging far behind and is mostly relying on the pre-chewed data for logistics in order to know what to push on.

And no, no reason to use real info in the first place, but I saved something like $20 on my last grocery trip (I spend $150/mo on groceries) by using the coupons Kroger sent to me. That's reason enough for me to give them my real address.
posted by SpecialK at 8:23 PM on December 16, 2007 [3 favorites has favorites]

It's just a phone number. My Kroger card is tied to a phone number I haven't had for 10 years. I've never had anything else checked against it and have no idea whether it's in my name or my ex-girlfriend's.
posted by rhizome at 8:24 PM on December 16, 2007

Never mind what I said about the check cashing info, then. :) Although, aren't you guys always caught behind check-writers?? I have NO idea why anybody still uses checks for anything other than what must be mailed, but they sure do at grocery stores. Shudder.

And thanks for all that cool info, SpecialK.
posted by iguanapolitico at 8:50 PM on December 16, 2007

SpecialK: That's interesting, but I think I'll keep my Kroger I.D.

as far as Kroger is concerned, My name is George Orwell, and I live at 1984 Privacy Lane in Zip 46229.

It's never caused me any problems, and I've filled out various cards with names like "Herman Hoglebogle" and "Ogg Mxyzptlk". No one has ever questioned what I put on one of those forms.

It was fun the one time I was addressed, seriously, by a clerk as "Mr Orwell".

It doesn't seem to matter. I do it for the sport of it. YMMV.
posted by pjern at 8:52 PM on December 16, 2007

SpecialK: I take issue with your numbers. An SKU takes what, 40 bits? That gets you a number up to a trillion. Same for time/date, store number, customer number. With an average of 20 items, that gets you 800 bits = 100 bytes. 2500 stores x 10000 transactions per day x 100 bytes/transaction = 2.5 GB/day. I could do that at home.
posted by alexei at 8:53 PM on December 16, 2007

The interesting aspect of affiliate programs is that they cost money - not only to advertise the program, sign people up, print cards and collect data, but they have to also offer some sort of discount. Who pays for the program? Basically the people who don't use the affiliate program.

If you want to provide fake info, go for it, no one will "catch" you. But you're still giving away your privacy. They may not know your name or address, but they can measure how that an article about pesticides on bananas suddenly made you switch to organic bananas. Or how you didn't go for organic at 30 cents more per pound, but you bit at 20 cents. It's a bit creepy.
posted by kamelhoecker at 9:01 PM on December 16, 2007

Yeah, can't agree with the numbers thing, that amount of data ain't nothing. But yeah I have half a dozen 'member' cards with fake info and never a problem.
posted by zengargoyle at 9:04 PM on December 16, 2007

a list of SKUs attached to a transaction number + time/date attached to a loyalty card number -- and compress the shit out of it three ways from sunday, you're still facing 20-100 Kb per transaction. ... Not processed into records in a database -- raw data on drives.

I don't buy that math or reasoning. Why would the retailer store it as raw data on drives? The POS machine obviously is retrieving SKU and associated data (price, description, etc.) given a UPC. Whether that information comes from a networked server or locally is irrelevant, it still ends up with a record associated with the SKU. I think it's safe to assume that SKU, or its associated row key, is some large integer. How large? Well, let's give it 8 bytes. That means we can store over 18 quintillion products in the inventory system, that's 18 with 18 zeros. All in only 8 bytes. So, each product a person buys takes 8 bytes. Let's say on average the normal person buys 50 items on a trip to the grocery store. That's 50 * 8bytes or 400bytes. Still less than a kilobyte. We'll also need to store customer id and date of transaction. Let's say two more 8-byte integers, for a grand total of less than 2 Kb. Count in all the transactions and all the stores, and you're looking at a total of about 29Gb per day.

Obviously, most people don't buy 50 items, most transactions are probably only for 10 or so items, so this is a probably a high estimate. But I think most large retail stores can afford to store on the order of 10terabytes a year. Terabyte storage is available from your local Best Buy, so it's not unlikely for a business to have that kind of storage.

And as for margins and value. Well, that data has intrinsic value. Advertisers, or "business partners" will pay good money for that data. And if things are as bad as you say in the grocery business, they'll have even more incentive to sell that data.
posted by formless at 9:15 PM on December 16, 2007

alexi - I didn't bother sitting down to calculate it out to the nth degree -- I'm horrible with numbers, even if I do work with computers. These days, I do more of the strategic thinking involved in developing high-availability OLTP and OLAP type systems... I've got engineers to help me with the numbers. Generally the formatting I've seen for this kind of data was (imho) kinda wasteful. It had the store number, transaction number, time/date, and sku all on one line -- about 120 chars long; some of the data was alphanumeric, so it was in a text file as a string instead of a long integer.

Stupid or not, that's the format I've seen common IBM POS systems put out.

That equals, by my numbers, about 240 bytes per line. Average 10k transactions, stick with your 20 items each, so 200k lines at 240 bytes - 45MB is 109 GB. So you're right, my numbers were pretty far off. Mea culpa -- the number I got in my 2003 tour of Fred Meyers DP room in their Oregon HQ was just short of 1TB per day of raw data from all of the Fred Meyer stores, which are only a handful in the northwest. That included everything from timesheets to material intake & inventory to pricing changes done at the store level...

My broader point, which I was attempting to get across, is that the grocery chains don't maintain that level of detail on the transactions that they do in the stores because the database tables would quickly grow to a size that would be unmanageable even with indexes. They don't retain much raw data because of the size of the raw data. What they generally *do* do is load it into big data warehouses and crunch it down in some preprogrammed pattern -- and all of the work that was being done at the time of my graduation was to reconcile purchases with specific promotions, to identify areas that needed to be promoted, and to restock stores with items they needed. They didn't, at the time I graduated, have the extra data processing power to handle the kind of big brother data processing that everyone thinks they do... there's no return on their investment.
posted by SpecialK at 9:21 PM on December 16, 2007

They can also do interesting data mining entirely within the system. Demographic data is nice, but not needed to track trends over time, or correlation between products, or similar things.
posted by smackfu at 9:24 PM on December 16, 2007

Oh, and as far as selling the sales data -- AC Nielsen already has the market sewn up and exclusives with both manufacturers and sales channels. They pay an ungodly amount of money for it, and you need an ungodly amount of money to access it, but all they do is provide summarized data. There's some young upstarts in the market -- Rentrak Corp, one of my former employers, is one of them -- but they are having a hard time breaking into the market because AC Nielsen is such a powerhouse.
posted by SpecialK at 9:39 PM on December 16, 2007

One thing which I haven't seen mentioned here is using your phone number in lieu of scanning your card. Customers at Fred Meyer can do that if they provide their phone number and check a little box on the application, but if you don't give your correct phone number then you must have your card. (FM's policy is to NOT have "guest" cards.)

At FM there's only a discount on fuel purchases, though they do mail coupons if you spend enough money. So, at least at FM (which is owned by Kroger), to get the most benefit, you should provide a correct address and phone number.
posted by philomathoholic at 10:17 PM on December 16, 2007

This is turning into an interesting discussion. SpecialK, your stories are interesting. But I can't help but think all this data represents a wasted opportunity for supermarket chains, and it is a matter of time before a company that is willing and able to innovate on the back-end - like, I don't know, Wal-mart - is able to properly use this data in all the ways we are imagining.

Did you ever notice that if you use your loyalty card at Safeway (or at least the ones in San Francisco), your receipt is about 15 times longer, because they print out custom coupons for you? I wasn't certain but I could have sworn they were tailored towards me, since they were for items I tended to buy. For a while I was glad that my name and address weren't linked to the card. Then I realized I always pay with a credit card and it would be fairly trivial to grab at least my name and link it to the safeway card. So I decided I would rather not worry about it.
posted by PercussivePaul at 11:05 PM on December 16, 2007

I have a friend who does NOT mind spam or data-mining, and loooooves collecting "points" and saving on various things.

I use all her loyalty card numbers. She must look like some kind of super-consumer in the databases. She's happy because she gets extra-fast bonuses. I'm happy because my spending habits are duly cloaked and maybe I'm even messing up their data. Yay anarchy.

At the grocery store, I get "her" discount for my groceries, too. We all win.
posted by rokusan at 11:36 PM on December 16, 2007

I think this would be fraud, if that bothers you. Of course I doubt anyone would ever, ever be prosecuted for this.
posted by grouse at 1:08 AM on December 17, 2007

SpecialK, start another AskMe in a few years after you've been in the industry for a while. Or better yet, post something in the Blue about what companies will do to increase a pre-tax margin from 3.0% to 3.1%. The costs of the IT infrastructure pales in comparison to the management consultants they hire to figure out the most effective way to use the information they've gathered. And for what it's worth, why offer the loyalty card schemes in the first place if there weren't some significant payback for the discounts offered.

If you value anonymity and privacy or simply don't want your mailbox stuffed with junk, don't feed the machine. Personally, I eschew the cards as I've never felt that a cash (or not-in-store operated credit) transaction should come with a rider that I disclose personal information. If it came down to having food on the table, I would probably fill out the form with demographically similar information as others have suggested.
posted by michswiss at 4:13 AM on December 17, 2007

I have cards from three grocery stores. Two of the forms had a little box down at the bottom that you could tick for "I prefer not to give any information," and you still get the same card. The third card didn't have that, so I put in a made up name and address. I guess I am missing out on the special coupons they could be sending to me, but that doesn't bother me, and I get quite enough junk mail already.
posted by Forktine at 5:31 AM on December 17, 2007

But you're still giving away your privacy. They may not know your name or address, but they can measure how that an article about pesticides on bananas suddenly made you switch to organic bananas. Or how you didn't go for organic at 30 cents more per pound, but you bit at 20 cents. It's a bit creepy.

They can do this anyway, if you use the same credit card. They get at least your name and a unique number per card, which isn't any different than the loyalty programs.

They're probably not allowed to store the actual card number, but I'm sure they can generate a one-way hash function that will protect your number from being extracted, while still tracking you as closely as they like.
posted by Malor at 5:36 AM on December 17, 2007

SpecialK, start another AskMe in a few years after you've been in the industry for a while. Or better yet, post something in the Blue about what companies will do to increase a pre-tax margin from 3.0% to 3.1%. The costs of the IT infrastructure pales in comparison to the management consultants they hire to figure out the most effective way to use the information they've gathered. And for what it's worth, why offer the loyalty card schemes in the first place if there weren't some significant payback for the discounts offered.

Actually, I decided the grocery industry was kind of boring and got out of it.

The main reason right now for the loyalty card things is that everyone else has one, and they'r seen as a marketing thing. As far as I know, the industry really hasn't changed in that they are still relying on data from ACNielsen as opposed to mining the data they collect themselves.
posted by SpecialK at 7:48 AM on December 17, 2007

My actual Food Lion card still works. I got it in high school over a decade ago in another state when I had a different name and used an address and phone number that even my mother doesn't have any more. I don't think I had email yet as it would have been close to 15 years ago. False information is a non-issue

If they don't require any kind of update on the information or a periodic renewal of the card, then their demographic checking leaves something to be desired, anyway. I don't think that they'd know any more about me if I put in a false name and contact info on a new card.
posted by Cricket at 10:22 AM on December 17, 2007

I meant to say "that they'd know any less about me."
posted by Cricket at 10:24 AM on December 17, 2007

Some more information I've culled from the web.

We've got a grocery store in Michigan called Meijer. It's a decent place, good prices, nice selection of organic foods, etc.

They've also got coupon printers hooked up to all of the self-checkout lines. The coupons it prints are usually associated with your current purchase. No big deal there.

Turns out, those machines are from a company called Catalina Marketing. Catalina marketing has deals with about 18,000 supermarkets nationwide.

According to the above story, and this story in USA Today about data mining, Catalina does have individual-level data. It's anonymized, but as recent experiences with the AOL query logs and NetFlix data set show, that doesn't mean much. Bargains at a Price also mentions the DEA requesting and receiving data on shopper's purchases.

I do have to say, I love this page on Catalina's site: request more info on data mining. It's listed as one of their products :). Yes, I'd like to purchase a dozen data minings please, hold the privacy.
posted by formless at 1:48 AM on December 18, 2007

Many (most?) cashiers at Shaws and Safeway will swipe their own card or a guest card if you ask, sometimes without asking.
posted by devilsbrigade at 11:44 PM on December 22, 2007