Author Jeff Zabin explained data capture and targeting techniques to iMedia Brand Summit attendees.

Jeff Zabin is well known for his customer data capture and marketing techniques. Currently, Jeff is the Director of Marketing at Fair Isaac Corporation, a company whose analytics help brands acquire customers more efficiently. Jeff is also the co-author of two books: The Seven Steps to Nirvana, and Precision Marketing.

Zabin addressed the iMedia Brand Summit in Florida in February. Read the first part of that presentation here, and part two here. Part three addressed context and marketing smarter. Here's the final segment:

Jeff Zabin: I want to talk a little bit about integrating customer information. There are a lot of James Browns -- are there any James Browns in the audience? Not one, huh? Well, there are a lot of James Browns, but there is only one godfather of soul. And, the question is, you know, how do you capture the godfather? How do you make sure you don’t mix him up with all of the other James Browns in the world? And, the way you do that is by using a unique. And, that gets tagged to this James Brown, and that way you don’t deal with all of the house-holding problems and all of the other issues that come up.

Integrating customer information -- once you know who your customers are and you have that unique identifier in place, you can begin to marry the data and the information that you have about them with third-party sources. And, you can also begin to capture more information about them. We work with lots and lots of companies, lots and lots of data providers are out there. I always think that it is always good to capture information through primary sources. Create a trusted relationship with your customers. Create a reciprocal value exchange so you have a quid quo pro in place. You give me information; I will reward you with discounts and coupons and content. I think that is a better way to capture customer information. But, ultimately what you want to do is create a multi-dimensional profile of your customers with relevant information that you can then use in your marketing campaigns.

I always say, in terms of integrating your customer, your distribution channel data, don’t be like the drunk who has lost his house keys. (I think “drunk” might not be politically correct. I should have probably said, “Don’t be like the substance abuser who lost his house keys.”) But, there’s an old joke, and the joke is that: There is a drunk under a streetlamp; he’s looking for his house keys. A passerby comes up to him, he says, “What are you looking for?” And the drunk says, “I am looking for my house keys.” And then, the passerby says, “Well, where did you lose them?” And, the drunk says, “Oh, oh, I lost them over there.” And, he says, “Oh, so why are you looking for them over here?” And he says, “Well, the light is better.” So, you know, that’s a good joke that sort of applies in the context of making sure that you capture your customer information from all of your different sources. 

You know, some consumer branded goods companies have a 50,000 watt spotlight shining into some of their channels’ grocery stores, for example, but they might not have any idea what is going on in restaurants or in concession stands. So, make sure that you capture information consistently across all your channels, as much as possible. 

And, I also say: don’t send merchandise catalogues to dead people -- it is just a bad idea. That is part of the value of customer data integration and making sure you are constantly updating your data in a dynamic manner, because you know, people get married, they change their names, they get divorced, they retire, and believe it or not, people die. Some of the people who you think are your best customers -- and they may have been -- may have been six feet under for five years now and you are still sending them promotions to enter a contest to win a year’s supply of laundry detergent. So, make sure you keep your databases up to date. You can constantly enhance your data base with, again, all of these different types of data. So, understanding the propensity to respond to your offer; understanding their channel preferences; understanding all of the demographic, and psychographic, and behavioral, and attitudinal information that may be relevant, that may ultimately influence your campaigns.

And, then, the next step is really advanced analytics -- doing predictive modeling. I checked the weather report before coming down to Florida to make sure I really wanted to come. And, you know, I saw that it was going to be in the 70s and, lo and behold, it is in the 70s. So, what is predictive modeling? It is really looking at historic data to foretell future events, or future behaviors. That is all predictive modeling is about. So, which of our customers are at risk? Which of our customers provide most of the value that drives our brands? Which customers should receive which promotions, and which treatments, and at which time, and through which channels? That is all about predictive modeling. 

I have this notion of blueprinting the ideal customer. I think, ultimately, what you can do is, once you understand what the DNA or the genetic makeup of your best customers is, how can you go out into the rest of the world and find people just like them? I think that is the value of having all of this information in your customer database and understanding which customers are your best customers.

I am going to talk quickly about a couple of interesting things we are doing. Next generation -- our market basket analysis -- this is a technology we just launched, actually last month at the National Retail Federation show in New York. What this does is, it is sort of an analytic visualization tool, and it allows the retailer to understand “What are the frequent item combinations my customers are purchasing?” both at the individual level and also in the aggregate.

Some combinations are obvious. You know people who buy flashlights are likely to buy batteries. People who buy shaving cream are probably going to buy razors. But, there are a lot of non-obvious combinations, and it is interesting. Using real customer data, we found that people who buy blenders have a higher than likely propensity to also buy Salsa music. So, what is the reason? What is it? Margaritas! Exactly! So you have that bridge product in the middle that allows you to understand that. By the way, they also buy vacuum cleaners to clean up after the party!

So, it is interesting once you can begin to map out this information. And, you know, if a CPG company can plug into that and understand, “Okay, what other products should I be bundling? -- people who buy my product also buy what other product? What is a temporal outlook? Over what time frame?” So, that is tapping into retail merchant transaction data. 

You can also tap into bank data -- direct deposit data. This is another tool we have, and you can begin to understand, looking at transaction data … you think you have 100 percent share of loyalty, share of wallet, with your customers. But, you look into this data and actually see, “Oh, 80 percent of them are shopping at my competitor.” So, you can begin to understand the geographic distribution. Next generation cross-selling is really about using decision tables, and using optimization tools, so you can define the goals and constraints of your marketing programs, and you can begin to understand what is the best, where are the best, dynamic trade-offs. 

I am going to skip through “Generating Buzz,” because there is a lot of talk about this, using blogs and these kinds of tools. I mean, the point is, who do you trust? Well, you trust recommendations and word of mouth more than you are going to trust anything that a brand company has to tell you. This is a tool that was developed by Intelliseek, which is very interesting. It allows you to actually track the buzz using hundreds of thousands of blogs. And, it uses latent semantic indexing and other methodology to really understand what the positive versus the negative buzz is that is out there.

What is the future of precision marketing? Is it going to be like what we saw in the Minority Report, with Tom Cruise, playing the character of John Anderton, walking into a Gap and biometric retinal scanners are scanning his eyeballs and serving up very context sensitive messages? Maybe. There are some privacy issues involved. I think it is going to be more like my grandmother and my grandfather who had a corner store in Council Bluff, Iowa in the '50s and '60s. My grandmother, she knew who her customers were. They would walk into the store. She could serve up customized offers to them. She could do cross-sell and up-sell, and she had all of that in her customer database, which was her own memory.

So, I just want to close by saying that I was a very cute child. (Laughter.) And, I want to thank you for the privilege of speaking today; and, thanks to the organizers of the iMedia Brand Summit. (Applause.)

Rick Parkhill: Thanks, Jeff. Good job.

Zabin: Thank you, thank you.