Data mining: Finding an arsenal in a bunch of dry bones

In the classic Stanley Kubrick film 2001: A Space Odyssey, based on Arthur C. Clarke’s book of the same name, prehistoric Man has a good day at the office when he realizes how to use an animal femur as a weapon. The implication is that this moment of invention bestowed upon our ancestors a competitive advantage.

What I love most about my job is being a part of a team that has these same “Aha!” moments. Quite often we make newer, better tools that enable us to go out and kill our suppers. How cool is that? 

Even cooler is how the lines between disciplines blur, and we have invention mash-ups. Because we are involved in direct marketing and research as well as interactive projects, we get to invent in those areas as well.

Next week I’ll write about just such an invention involving geomapping. But today it’s analytics, one of the more fertile areas for innovation — and improved ROI — in business today.

I interrupt this story to recommend a free webinar on gaining a competitive advantage through analytics, put on by the American Marketing Association and Aquent (thank you, James Gardner, for the head’s up!). It takes place on October 31 and covers topics every marketer should be familiar with.

3D BarcodeThe analytics “invention” I describe below came from a conversation I was having several years ago with Mike Czerwinski (a co-worker in a prior life), who is one of the smartest people I’ve ever met. Too bad he squandered all that brainpower by getting a PhD in mathematical physics. I joke of course. His scientific background and ceaseless curiosity help him mine data in ways that are sometimes quite unexpected.

Over a beer, Mike and I were contemplating the barcodes on the back of Wisconsin driver’s licenses. They contain all the written information on the front of the card — far more than name and mailing address. The information contains height, weight — even eye and hair color. As Mike might put it, many data points.

This led to the discussion of nightclubs. It’s a very competitive industry, where your profitability hinges on the caprices of a community’s young and social. Significantly, Mike and I had both just read Malcolm Gladwell’s The Tipping Point.

I can’t speak for Mike, but I certainly know that I was looking with a fresh set of eyes, seeing everywhere the social epidemics that Gladwell describes. More importantly, I was ever-wondering how to find Mavens and Connectors. These, according to Gladwell, are the few who influence the many. They can be found at the head of most consumer stampedes, and are to be courted and coveted by marketers — if they can be identified.

How to identify them in a nightclub? If you could see these vectors of influence, and offer them free stuff (drinks, a party room for a special occasion, etc.), they would reward you by drawing their many friends with them back to your club again and again. And if you owned a number of nightclubs instead of one, you could find Connectors or Mavens who frequent one of your clubs and coax them and their loose posses to another, on a night when that club needs the business.

Most importantly, you watch the frequency of their visits. Your club could be packed, but if these influencers (as they are often called in word-of-mouth advertising) have stopped attending, your business is about to plummet. Their departure can be a leading indicator, which can help you make decisions about advertising, staffing, purchasing — every part of your business.

But how to find these valued customers? The answer, as it occurred to us, was the scanners that nightclubs use to document, upon entry, the legality of their patrons. Because being caught with under-aged drinkers can shut a club down, entrants are often handing over their licenses to an attendant who runs the card through a video scanner. I suspect that the typical scanner just takes a picture. It doesn’t actually capture the data. But it could. Patrons give up this piece of privacy for the privilege of partying at a club, and in so doing hand over an opportunity that Mike and I suspected was being overlooked, like a pile of dry bones.

What if the database of patrons was analyzed daily? We wouldn’t be looking for the most frequent patrons (necessarily), but for those loose groupings of people — those clusters — that show up frequently in the line of ID cards. Each cluster would be like a solar system in a nightly star map of patrons. And each planet in these solar systems would invariably orbit the same recurring Connectors and Mavens.

And guess what? You’d have these people’s mailing addresses. Once you identify these influencers (or suspected influencers — you can afford to be wrong occasionally), you’d mail them offers that really matter. A mailing to a few hundred people could potentially move a thousand or more partiers to your establishment when you need them the most.

If you know of anyone applying this idea, please let me know. It’s a concept that still needs to be tested, but I think it could have tremendous potential. Click to view a larger version of this mapIn the meantime, here is a social map that charts the frequency that bands play at certain nightclubs. The clubs are colored according to type (blue=cover, red=original). Lines between clubs represent the number of bands that played at both clubs over the 18-month time period. The size of the clubs’ dots indicates sharing of bands. To get the full effect, and to read the whole story, check out this report with a time-lapse animation. Although the report suffers from lack of clarity, it does have some interesting observations.

Review the graphic and you’ll notice that the cover clubs don’t book the same bands as each other (the blue nodes are smaller) while the original clubs do share the same talent (the greater sharing of bands over the period increases the size of the red nodes).

I think you can infer from this that cover bands, which tend to play genres of music (e.g., Disco, Grateful Dead / Jam, Modern Country), tend toward a few clubs that fit this style of music, while a greater “genre” is new, original music. This material needs to be played in more of a circuit of clubs, ala vaudeville. Back to Mike’s and my invention, I think it would only work in the red clubs, because sharing of bands also means sharing of patrons, who follow the bands. Or, of course, the idea would work for nightclubs featuring no live music at all.

Like much in academic studies, I see little immediate business value to this Boston College study, which in some ways states the obvious. But it makes for an interesting view of the nightclub scene. And who knows … it might fuel your own beery inventions.

A message to Mike: We really need to go out and get a beer sometime. I miss the shop talk and rambling inventions of the four of us — you, Kevin, me and Guinness. Call me.

5 thoughts on “Data mining: Finding an arsenal in a bunch of dry bones”

  1. Hey, Jeff, here’s a link to an older discussion on hacking driver’s license bar codes:

    http://www.instructables.com/id/E0DU1K3WR8ET9K5QQJ/

    I’m too old for clubbing, but if I found out a club–or a liquor store, or a grocery store where I bought liquor– was storing and using my data (does WI include SS# in its license bar codes?) without my OK, I’d be unhappy. Handing my license over to prove my age does not, in my mind, give the reader my permission to permanently harvest the information.

  2. Hi, MJ —

    I suspect most people would be concerned if they learned that a merchant was harvesting their data, but it still happens all the time, usually in exchange for something of value. For instance, we allow a grocery store to associate our food buying habits to our name and address, all in exchange for discounts.

    If a nightclub was to use ID information to offer free things to those whom they consider prime users, I think the same transaction is taking place. But the consent would have to be more explicit than simply handing over an ID to show you’re old enough to drink alcohol. There should be a sign posted, and perhaps a sample of mailed offers that select patrons receive.

    As for information available, I believe that the DMV doesn’t place SS#s anywhere on driver’s licenses, so that info isn’t encoded on the back of the card either. Still: Height, weigh, AGE — hot stuff! Assuming the data is accurate (dubious), that’s a lot to hand over to a bored bouncer. Point well taken, MJ.

    P.S. Thanks for the link!

  3. Thanks for your comment. I think only part of the intrusiveness of marketers today (or filthiness, to use your term) is that technology makes it possible and frankly quite interesting. Another reason for this type of targeted messaging is a sincere effort to give people more of what they want. (Want, mind you, not need — a distinction that is the only thing that gives me pause about my job). When people get something extra from a known merchant, or find a new merchant that satisfies a need (well, want), they would tell you they’re happier for it.

    Sometimes when I’m trying to put something in perspective, I’ll imagine what a marketer would have done 100 years ago. For instance, did the owner of a general store in a small town in the year 1906 behave in a similar way as we do now, and was this merchant eventually driven out of town for those same marketing practices?

    Example: When Judy came in for her annual bolt of gingham, and Sam the grocer gave her an extra length of a new fabric for free, because she’s a “good customer” but also not coincidentally because she is part of an active sewing circle and could steer more business to the store, was anyone harmed or upset? I see the idea I described above as being no different. It’s just that with the mass of people we sell to in 2006, automation has to notice the best customers and figure out the few who can influence the many.

    Am I off base with this opinion?

  4. The ability to identify AND attract Mavens and Connectors is perhaps one of the greatest values that agencies can and should provide customers.

    Identifying them, I believe, is relatively easy if you spend the time to get to know the audience.

    Using your example, the maven-esque Lisa’s List will tell you most of what you need to know about the nightclub scene in Milwaukee…with and without music. Readers of *info would be your connectors.

    I’m fairly confident that it would be easy to find examples like these in almost every US metro if one were to do some earnest investigating.

    Thus, like most audiences that form a meaningful market, there are many marketers already capitalizing on them.

    The challenge then is to either see that audience in a different way (which these new technologies help us do) OR/AND make messages more attractive than those that have beaten us to the punch-line.

    If that’s what an agency is providing its customers, and in turn their customers, then there should generally be little or no cause for pause.

    As for tattoos on eyelids…that’s creative.

Comments are closed.