The Role of Customer Experience for Machine Learning

Jim Sterne is the founder of the Marketing Analytics Summit (former eMetrics Summit) and co-founder and Board Chair of the Digital Analytics Association. An internationally known speaker and consultant, he is the author of numerous books, including Artificial Intelligence for Marketing, 101 Things You Should Know About Marketing Optimization Analysis, Social Media Metrics, and The Devil's Data Dictionary. He has spent more than 35 years selling and marketing technical products and has devoted all his attention to the Internet as a marketing medium since 1994.

D.D.: We talk a lot about customer experience these days. How can machine learning support a customer-driven strategy?

J.S.: That’s a big question. First of all, what is customer experience? I’ll break that into two points, one is the customer journey, and the other is customer satisfaction.

1. Customer Journey

J.S.: Customer Journey is a make-believe construct that marketing consultants have grabbed onto in order to try to understand what goes on. Because everybody goes through the 4 stages of the marketing funnel: awareness, desire, research, and acquisition. No, they don’t, they just don’t. And we have no idea how somebody actually found out about our product. Because every time you ask the drop-down question “How did you find out about us”, 95% of the people pick ‘other’. Simply because we don’t know.

So when you’re going to ask your entire customer base how they came to choose your product, they don’t know and they’re just going to answer your question because they think you want to hear from them. And you’re going to take that data that is not valid, and you’re going to invest in a customer journey that doesn’t reflect any individual because everyone took a different path.

2. Customer Satisfaction

J.S.: And that is why customer satisfaction is so important.

The important question is: why did you get in touch with us - whether it was email, phone call, downloading a white paper? Why are you here? What are you trying to accomplish? And how good were we at helping you accomplish that task? And maybe that task is “I was looking for your phone number” or “I went to your website yesterday to remind myself what time zone are you in.” Is that a customer journey? No, but it’s a valid use.

Customer Experience for Machine Learning

J.S.: Customer experience is about opinion and attitude. Do my customers think I’m doing the right job for them, at this point in time? So, if I have a dozen reasons why somebody gets in touch and I ask them which one of them was it and “how satisfied are you?”, the first thing I find out is where people were unhappy. I find out that my website navigation doesn’t work or I find out that I have hidden an important piece of information too many layers deep that I should just put on the homepage.

Let’s go to enterprise-level where I have thousands of people coming to my website at any point in time, doing hundreds of tasks, and I am collecting a ton of data. Now I can start doing some analysis.

If I take all that behavioral information and apply all that machine learning, what I get is patterns of behavior. That’s not going to be a customer journey, but it’s going to be an indication of what kinds of people are doing certain types of things. When I connect that behavior with customer satisfaction I am able to find out if people who are either happy or unhappy when they are trying to accomplish a specific thing.

Trying to imagine what’s going on in the minds of my customers is a bit of fallacy. Because I imagined that the product I am creating is going to solve this kind of problem, and I put myself in their shoes, because I’ve lived in this problem-information space for so long, I have a clear idea of what I’m doing. And all my customers are prospects, they have no idea, they might not even know they have a problem for me to solve.

J.S.: I sound skeptical, but I think that this idea of customer-centricity is so fundamental. Customer experience is something that I think you should learn about in all the ways they are. Do surveys for satisfaction, look at behavior, do text analysis on all the communications that come into your company – whether it’s email, or chat, or phone calls that you transcribe. Collect all that data and immerse yourself in it to know what’s going on. Interview individuals, ask them open-ended questions, and listen to their tone of voice. Add all of that together and you might have a clue of what’s going on in your customers' minds.

If you have enough data, you can start to find patterns. The people to whom we sent these messages, responded in these ways. And then segment those: out of those people, these are different. And this is where machine learning is so valuable for customer experience: it can cluster, segment, and look at different behaviors.