Marketing on Steroids

Posted at November 16, 2009

Guest post by Bob Angell

Believe enough time has passed where I can begin to discuss some of the Marketing concepts that were part of some significant Intellectual Property and Capital (IP/C) produced while I was at IBM.

Many years ago, most commerce took place at the neighborhood store(s) and most of the owners knew who you were, your family and the current economic situation you might be in. They knew your likes, dislikes and what you may or may not purchase. Fast forward 40 years and that is what every retailer would like to understand.

Given today’s technology, it is now possible for anyone to essentially profile who you are, your proclivity for certain activities in a variety of situations and whether or not you are “wallet” worthy for that experience.

Loyalty cards in retail settings attempted to provide some of this data by having you (the consumer) provide the demographics piece (to obtain the Loyalty card). When you used the card week-after-week at the retailer, the items purchased would provide them with an idea of who you are and what your spending habits might be. For most retailers the loyalty card is a waste of time and money because it is not capturing quality data.

For example, there is a retail store not far from my house where I make infrequent purchases that uses a loyalty card. Believe that when I signed up for the card I used some fictitious name (like Donald Duck) and used a bogus address, etc. I still get discounts for using the card (which I do), but the store gets squat on the who I am as a customer. In fact, if they are mining the data, it does not accurately reflect ANY market segment – its just garbage. Remember the GIGO principle – Garbage In, Garbage Out.

While working in IBM’s Data Mining Group and Retail Divisions, it became very clear to me that there had to be a better way for the Retailer who was serious about capturing good quality data. Retailers need to be able to identify who you are (not by EXACT person [actual 1:1 marketing simply does NOT normally work, and it is WAY TOO expensive]) generally and what segment of the general population you might fit into. Knowing this information, they [retailers] can maximize their portion of your wallet (wallet share).

When IBM’s Data Mining group would work with large retailers (mainly helping to maximize marketing campaigns) the goal was to mine down to the individual. This did not work very well and were satisfied with a very detailed group customer segment.

We developed and came up with the ability to identify the potential market segment you might fit into using sensors and data mining. Using the input, it would be possible to place you into a marketing segment.

For example, you might arrive in the parking lot at a big-box retailer in a brand new car with children in tow. There are several ways for one to view this situation as a casual observer. They are listed as follows:

The first and probable one might be that this is YOUR brand new car and these are your children.
Second, this might be your brand new car and these are YOUR grandchildren
Third, you might be a car thief and just kidnapped these children.

Whatever the reality of the situation, one needs a bit more context from other sources.

With the perfection of video, audio, biometrics, olfactory and other outside stimulus, one can begin to infer and very quickly determine the reality of the situation. Video provides the ability to do facial, cart and license plate recognition; Audio allows one to understand whether voices are natural or stressed, Biometrics can determine whether someone is nervous, sick, sad, happy or potentially receptive to a Marketing message; Olfactory stimulus can determine change in chemical composition(s) such that it might alert one to a good or bad outcome (cologne or gun-powder residue) on an individual. Also, with video input, you can begin to figure out the age of the individual(s) in question and begin to infer a few more snippets of data. Is the ability to infer perfect? No! Is it much better than the current situation? Yes it is!

By the way, these Marketing Messages can be delivered and processed in real-time. This is really important especially if you want to deliver and provide real-time marketing to your customers using Dynamic Digital Media or other similar methods. You can even infer that the customer might not have the “wallet” worthiness you would like and create a disincentive for them to shop elsewhere. With some of my work, we were even able to determine whether or not someone would be receptive to a marketing message.

I have used Retail as an example, but these methodologies can be used to cover a wide range of activities from keeping bad-guys at bay for National Security to providing real-time information to Physicians while delivering healthcare.

Therefore, here are 5 basic Marketing lessons learned as follows:

Marketing is MORE of a Science than it is an Art.
Marketing is NOT effective in MANY situations – you need to learn where, when and how it is to be deployed for maximum efficacy.
Successful Marketing is about creating an urgent need out of someone’s want. Knowing the person’s receptiveness to a given situation one can infer whether they are able to make this translation.
Understanding the person’s capacity, ability and desire for this Marketing conversion is also very significant. If someone is struggling financially and does not have the means to pay for an item, all the marketing in the world simply does and will not work.
Marketing to the masses is all about SMARTLY converting these individuals to act on their wants.