Civicom: Your project success is our number one priority

A Conversation with Stan Sthanunathan of Unilever on His New Book

Stan Sthanunathan of Unilever sits down with Larry Friedman to talk his new book "AI for Marketing and Product Innovation".

Editor’s Intro: Stan Sthanunathan, Executive Vice-President of Unilever’s CMI organization, has been a real force for change in the Insights world for many years. He has just co-written a new book called “AI for Marketing and Product Innovation”, along with Dr. A.K Pradeep and Andrew Appel. The book is being released this month globally (and can be found here). To commemorate the book’s release, Stan graciously spent some time with me on the phone. What follows is a “lightly edited” version of our conversation.


Larry: Stan, thanks for taking the time to speak with me today. Before getting into the substance of your book, I’m wondering what your inspiration was in writing it?

Stan: Marketing has been undergoing a major transformation because we have a clear point of view that the world of technology is impacting the work of Insights. There is certainly a lot of hype around things like AI, with people talking about Watson and other platforms. There are also a lot of startups doing exciting things in this space that we felt could have a significant impact on the insights industry. At the same time, there was also a lack of clarity about what AI could really do. Is it something that was going to replace you and me, or is it something that is going to make us look smarter? So we wanted to set the record straight, what is AI all about where can it be applied, and what kinds of companies could help us on our journey. We knew that there actually was a lot the technology could do, so we said let’s document it all, and write a book applying what artificial intelligence can do for the insights industry.

 

Larry: Given that, what are the main points you try to convey in the book?

Stan: Let’s take a look at the typical work that we do in the market research profession, both qualitative and quantitative. Historically, the idea was that technology could only play a role in quantitative data, but we are coming to the conclusion that it can play a pretty significant role in qualitative as well. That is somewhat surprising application for AI. So what we talk about in the book is how AI could impact all the work we do in market research related to the 4 P’s of marketing, price, product, etc.

 

Larry: Your focus so far has been on market research. What impact do you see for marketing itself?

Stan: AI will impact both marketing and market research. I see a world where AI will generate insights, it will generate content, it will evaluate it, and it will tell me how well it is actually performing in the marketplace. AI will help tie all these activities together at a faster pace than we can do today.

 

Larry: There are certain words like artificial intelligence that are used differently by different people. Can you tell me what you mean by AI?

Stan: For me, artificial intelligence is when a computer is able to perform tasks that normally require human intelligence

 

Larry: Another term that’s used in your book is “Augmented Intelligence”. How does that differ from Artificial Intelligence?

Stan: If I had to put it into a simple formula, I would say that Augmented Intelligence equals Artificial Intelligence plus Human Intelligence.

 

Larry: How does the human intelligence get used?

Stan: We think about it this way. The general fear is that artificial intelligence will take over, and put us all out of jobs. I always ask people to describe their day to me, and how much time do they spend on different actions. When you dig deep into those activities, you find that people spend a lot of time doing process-driven work. If you think about that from the perspective of activities around data, people spend so much time on a process that they’re tired when it comes down to actually finding the meaning behind the data. Is that really the best use of human intelligence? This is something we should delegate to something that can do the process work 24/7 without getting tired. This way a human can spend his or her time using the pre-processed information to actually add value to what is there. There is a lot of value that comes from experience, and it can’t be utilized if people are fatigued.

 

Larry: How are these things like artificial intelligence or augmented intelligence changing the way marketing is done?

Stan: If you go back to how we always did segmentation work, you would find that we would do a segmentation study, tell marketing which segments were good targets and why, then they would tell the agency what kind of creative messages were necessary, and the agency would come up with the creative and find media they thought would target the segments. In today’s world, there are incredible opportunities for hyper targeting, so you don’t need to create one piece of content. You can create many to get at very specific targets. In the old days, it would take someone a week or ten days to come up with one specific ad and pre-test it. Now, with AI-powered suites, you can come up with any number of variants and do real-world testing to determine which ones are most sticky with incredible speed and efficiency. This way, you do content creation, content testing, and spending allocation all together in an integrated way.

 

Larry: Stan, what you’re saying really resonates with me. For a long time, I’ve felt that for market research to remain relevant, it can’t just stop at “insights,” that action has to be fully baked into what we do. So for segmentation, a researcher can’t say “here are the segments, here is what drives them”, and then hand it off for someone else to figure out how to create action. What used to take weeks needs to be compressed into hours and days.

Stan: That’s exactly right, we need the AI to allow people to move up the value chain, so that they can focus on the “so what” and the “now what”, and not exhaust themselves so much on working on the “what”.

 

Larry: So does that then mean from a client organization perspective, that the Insights function and the Marketing function get driven closer together than they are right now?

Stan: Yes, that is inevitable. As a matter of fact, in my view, in a couple of years from now, the lines between Insights and Marketing will be blurred, and it will be hard to figure out who is Marketing and who is Insights.

 

Larry: I think that is really important for anyone reading this.

Stan: I can see it happening already. The challenge is how to get people prepared for this new world. Insights now can come from anywhere in the organization, so the focus for an Insights function has to be on how to drive that into action.

 

Larry: How are people within CMI in Unilever reacting to that sort of message?

Stan: I was really pleased by the willingness of people to jump in and say “I’m willing to transform myself.” People really wanted to get ahead of the curve.

 

Larry: It’s great that you’re seeing that kind of positivity inside, but what are you seeing in the supplier world? How are they reacting to that world of artificial intelligence and augmented intelligence that you’re articulating?

Stan: Our big partners have also been willing to change once they’ve seen the transformation we’re doing. They also realize they have to adapt.

 

Larry: You mentioned qualitative before, how do the themes we’ve been discussing play out in that world?

Stan: One of the things people think of when they think about artificial intelligence is big data, the more data we can get, the more insights we can get. I’d like to put a different spin on that, that there is much value to be gained from artificial intelligence applied to get qualitative insights. We can learn a lot, for instance, about new product opportunities.

Let me give you a simple example. Marketing will sometimes say that we can’t find trends fast enough in local markets, and that is one of the advantages that local brands will have. So what will predict a trend? That led us to a point of view that the movies they watch, the TV shows they watch, the songs they listen to, the magazines they read, all subliminally influence the sub-conscious, and that influences the trend. Now we can’t listen to all the songs, or watch all the movies, but can we use AI to mine all that to anticipate trends? When we tried some experiments, it turned out to be a goldmine. It won’t necessarily tell you specific product ideas, but it will tell you areas that you should be looking at.

 

Larry: Has anything come to market yet?

Stan: Yes! Ben & Jerry’s Ice cream for breakfast. Ice cream isn’t traditionally seen as a breakfast food, but we spotted the trend, put it into scoop shops and it worked!

 

Larry: Ice cream for breakfast works for me. Is there anything else you want to say in conclusion?

Stan: AI can be whatever you make of it. If we want to become slaves to technology, then it will come true. If we leverage technology to make us smarter, then there is a huge opportunity ahead of us. It will take 10-15 years for AI to get to the levels of human intelligence. However, I will never underestimate the capacity of human intelligence to stay one step ahead. I think human ability to adapt to change is incredible, so let us not be despondent that technology will take our lives away.

 

Larry: Stan, let us end on that optimistic note. Thank you very much for taking the time to speak with me today, and good luck with the book.

Please share...

One response to “A Conversation with Stan Sthanunathan of Unilever on His New Book

  1. From a researcher’s perspective, AI might be about trend spotting and insights (without prediction) that a segment might produce better results if targeted with different creative. As an old researcher who crossed the chasm 8 years ago, frankly that all falls short for me. AI as it is unfolding is all about PREDICTION. Google predicts the next words I will type. Amazon predicts what I might like to purchase. MTA providers are using machine learning to PREDICT what will happen under different advertising activation alternatives. AI must be targeted to prediction to be valuable to marketers and will then, and only then, mimic the shopper’s brain, which is the quintessential prediction engine.

Join the conversation

Avatar

Stan Sthanunathan & Larry Friedman