A - My responsibility is to manage all research that is being conducted within my area on a product line basis. I am responsible for all the quality, the communication, the planning, and the actions needed. I guess I am a coordinator of all the different activities of communicating to my brand group what the data mean. I am a sounding board to them. They may come back and look at some data and say, "We think it means this. Why is it that the data show that?" They are asking me to explain the data in certain ways so that they can understand them. They don't want me to explain it to them in research jargon.
I go through the information to see what story it is going to tell me, to decide if it's significant and how significant. I first look at the criterion measures. What measures will give me the most information, to show me, for example, that product A is better than product B, or commercial A tests better than other commercials? I then look at support questions to see if I can understand why A did better than B, and I try to read the story, rather than just report the numbers. I try to "translate" the numbers into a story, rather than write a statistical report.
The most difficult part of my job is finding out what management really wants to know. You can talk for hours and still not find out what they really want to know. They may think it is crystal clear, and then they come in and they tell you what they want to do, and you sit there and say "Why?"
Q - How does a typical project proceed?
A - I don't think there is such a thing as a typical project, but I'll try to give you an idea. For example, marketing personnel would come to us if they want to do a product test or if they want to come up with some claims that they can make in advertising. We call that a claim substantiation use test. Once they tell us the claims that they want to make, we would design a study representing the user group they are interested in. We would then contact the supplier. Normally, we would contact three and get competitive bids, and then determine when the study is going to start. In the meantime, the R&D people are making sure that the product to be used on the test is representative of what is available. They package the product, label it, and get it ready for shipment for the supplier according to our instructions.
We have what we call an approved supplier list. There are literally thousands of companies that call themselves marketing research suppliers. We only deal with about twenty-five to thirty that we consider the very best, most reliable, honest, and dependable companies available. Integrity, confidentiality, and security are all key in this business. We also consider how easy they are to work with, the kind of service they give, and how timely and accurate their data are. Cost is not really much of a factor, because the caliber of these companies is equal, so the prices are going to vary only $500 or $1,000.
Now at the time you award the study to a supplier, you sit down with that supplier. You have already designed the study, because you have gone out for bids on it, but then you work out the questionnaire. It all depends upon the product categories you are working on. There are some standard questionnaires, and there are times when you want to ask questions you haven't asked before. In those cases, we pretest those questions before we go out and use them on a large scale. Whenever we do a study for the first time, we pretest it extensively to make sure we are getting the information we want. We did one study recently that was a large-scale national probability study. We were doing 2,500 interviews, which is not all that common, and we wanted to make sure that every one of those questions worked. So we spent three days at the supplier's office listening to interviews being conducted and changing things as they went along to make sure that the questions were understandable. That's not typical, but it's the kind of thing you do when you are spending $100,000 for a study. You want to be absolutely certain that the results you get reflect your objectives and that it's money well spent.
While we are designing a questionnaire, we are also trying to keep in mind how we are going to tabulate the data. So, you are setting up ahead of time for criteria, for statistical tests, for cross tabulations, or just basic straight tabulations. You are designing the size and composition of the sample in order to be able to look at specific groups. That's where experience comes in. After a while you know how things are going to look. You pretty much know how many women out of 300 will fall into each category on a tracking study. You know, based on your share data from Nielsen, the share of your product in relationship to the share of the competitors' products, and you can factor that down to the number of users there would be in a representative sample. So you build sample sizes based on the number of different tabulation breaks you may want to look at later on in order to analyze the data. So you plan ahead, so that you think you know all the things you want to look at. Now, when you get into analyzing the data, you may want to ask for additional data, because you may see a couple of pieces of information that may seem to fit, but you need a third or a fourth piece of information to give you enough conclusive evidence to make a statement.
