12-30-2008, 01:20 PM
Life Controlled by Data Bases ....
Too Smart By Half |
July 22, 2008
6:34 AM CST
I’ve talked about supermarket loyalty cards before—they’re the mechanism by which supermarkets collect purchase information from their customers, to find out what is being bought by whom.
(At this point, please drop all the Big Brother notions. Nobody is interested in, or has the time to bother with the fact that you personally bought batteries and aluminum foil last week.)
The goal of collecting this data is to do a better job of keeping merchandise in stock, and to do a better job of promoting the chain’s desirability to you, over that of their competitors.
You are encouraged to use your card by granting you access to discounted merchandise—yeah, it’s a bribe, all discounts are—so that shoppers can be grouped into various subsets which make it easier for the chain to take specific marketing and merchandising options over other, less effective or less profitable ones.
Most promotions used to be crapshoots. Retailers had absolutely no idea who was buying anything; they only knew what had sold. Hence, most promotions used to feature duh! items like milk, bread, Heinz ketchup, Tide detergent and Hellmans mayo, but after a while, they became less and less effective because, well, everyone was promoting the same items.
There had to be a better way to pick promotional items, and loyalty cards promised that better way.
So, if you see that a group of people has regularly bought flour, baking powder and sugar from your store, it’s a reasonable assumption that they bake stuff, and so you can decide to promote “baking” items like cornstarch, cake icing or castor sugar to that group, with the twin goals of selling more product, and making them more loyal to your company. Crude, but a lot more effective than the old ways.
The problem with loyalty marketing in a high-volume, high-transaction business like retail grocery is that the amount of data generated each day and week is enormous, and the computer horsepower required to make sense of the data, turning the data into information, and thence into action, is considerable.
A lot of the work can be done by people who know what they’re doing, of course—but there are probably no more than fifty people in the world who understand this game to its fullest, and have the experience to know which roads not to follow, and which actions not to take.
So a lot of the work has been turned over to the statisticians, to companies like SAS and Dunhumby, and these (very smart) people build enormously-smart systems which crunch the data and spit out information at the end. These systems are, unsurprisingly, horribly expensive, and they reside in only a few locations.
The problem is that because the systems do all the heavy lifting, the actual followup work is often done by inexperienced people—people who may be smart, but who have no smarts. These people are hired because they understand the systems, but not necessarily the business. They are considerably cheaper to employ than people who know the business, and orders of magnitude cheaper to hire than people who, like The Mrs. and myself, understand both the business and the systems.
So you end up getting a result like this one:
Tesco had screwed up, royally. How had this happened?
When Lynn Newby logged on to her online Tesco account she was horrified to discover her ClubCard had been used to purchase a pack of 12 condoms.
Believing the only logical explanation to be that her partner was having an affair, devastated Lynn confronted Andy Allott and prepared to cancel a dream holiday to Mexico and end their four-year relationship. She accused Andy of having a secret fling and refused to believe his denial, claiming the evidence was on screen in black and white.
The box of 12 Mates condoms had shown up at the top of a list of Lynn’s recent purchases made with the supermarket chain’s loyalty card.
But shortly afterwards a friend from work who Lynn had told about the problem rang to say she [too] had found a Â£7.10 pack of condoms on her Tesco shopping list and so had a member of the supermarket’s staff.
After being contacted by Lynn, Tesco promised to investigate the matter and she eventually received a letter of apology from Chief Executive Sir Terry Leahy, and a cheque for Â£100.
Note the words ”new member of staff” and ”input a promotion incorrectly” (a grammatical howler, but let me not go there).
Sir Terry told Lynn he could appreciate how distressing it had been for her and an investigation had revealed that the problem was caused by a new member of staff in the marketing department who input a promotion incorrectly.
He wrote: “This error was completely unacceptable,” adding that staff were being given further training as Sir Terry and staff are being given further training “ as a matter of urgency”.
Tesco has probably the best supermarket loyalty program in the world, and maybe the best program of its kind in any industry. I’ve seen presentations about it, have been massively impressed, and I read their book on the topic (Scoring Points).
But when I read their book, I saw several occasions where they had committed unbelievable blunders, simply because their data had seduced them into taking action which anyone with even a smattering of experience in the field would have told them to forget about. One such mistake cost them over a quarter of a million pounds—two years’ salary for someone who knew what they were doing—and this latest condom boo-boo is just indicative of where they’re going wrong.
I could have helped them avoid this, because the concept of “intimate items”, and the inadvisability of promoting (or even tracking) them was the topic of a speech I gave to a group of industry leaders, back in 1998.
This is not an advertisement for my services. I quit the industry many years ago, and although I still keep in touch with trends in the business, nothing much has changed since then, and I have better things to do with my life anyway.
What interests me about all this is the substitution of data for knowledge and in some cases (as the above disaster), for pure commonsense. It’s a trend I see happening all over, and not just in the supermarket industry.
Remember the fuss when Ted Kennedy was mistakenly put on the TSA’s Suspicions Persons Watch List? (I know, cynics may suggest it was no mistake at all, but we can leave that discussion for another time.)
This was a result of a decision driven purely by data, with no oversight by a knowledgeable person. Even a backup system to catch such anomalies would have helped—but you have to know that a backup system is necessary before you build one, and clearly, no one had thought of that.
As we individuals increasingly become just a collection of digits in The Great Big Database, expect more and more of these mistakes to occur.
Tesco’s cockup caused them to lose goodwill, probably a few customers, and most likely a ton of lost data, as people refuse to use their Tesco shopping cards. In the grand scheme of things, they got off lightly, simply because supermarket shopping is not that critical an activity in people’s lives.
As our lives become increasingly controlled by large databases (think: the aforementioned TSA, insurance companies and government, to name but a few), future outcomes are not likely to be as benign.
So if a simple supermarket db problems, almost cause a woman to dump her BF what is going to happen when someone other then Teddy K gets on a No FLY LIST by mistake ....
[ Reply w/Quote ]