Interview with J. Walker Smith on Algorithms.


Adotas spoke with J. Walker Smith, Chief Knowledge Officer at Kantar Consulting to discuss the impact of algorithms.

Which aspect of retail is AI impacting more, marketing or shopping?

When discussing marketing or shopping we are really talking about “algorithms,” not “AI.” By algorithms, we mean a set of structured and fixed decision rules, which are embedded inside marketing automation software or consumer apps. AI refers to learning systems that update and (hopefully) improve the decision rules through experience and other inputs. There is a blurry boundary between the two, but I’d like to focus on algorithms.

The biggest impact is the ongoing erasure of the distinction between marketing and shopping. What differentiated the two in the past has been lost in the digital era. If I had to pick one of the two, I would say marketing. However, I say this with the following nuance: The biggest commercial impact is seen in shopping, but the biggest functional impact is seen in marketing. Marketing can now be quantitatively gauged and evaluated. Obviously, this wasn’t always true. Shopping is still about conversion, just as always. Marketing is now about triggering action, but it used to be only about awareness and perceptions.

How have algorithms changed the way consumers shop?

Think of this in terms of the new “algorithmically enabled consumer.” People are using algorithms to make smarter decisions. Algorithms can search for the best price relative to a given set of ingoing preferences, whether stated or derived. They can also match particular items (or people) with various kinds of consumer criteria (demographics, past purchasing, psychological, etc.) to enable a “best fit.”

Algorithms are often predictive (e.g., Netflix), thus improving fit, saving time and providing value. Consumers want algorithms to enable them to make better decisions — both a better product and a better shopping experience. This use of algorithms (mostly through apps) is only going to increase, as consumers are going to rely more and more on algorithms to do the legwork, which is exactly the kind of convenience that consumers have always wanted more of.

With a shift towards algorithms, how does value change?

Value improves in two ways. First, in terms of perfect fit: Consumers want personalization, but not if it comes with more demands on their time, their cognitive capacity, their pocketbooks, or their stress and worry. This also means a shift of value, because, when every item is highly personalized, the signaling value of premium products will be eroded. Put simply, what I have won’t signal any status to you because you’ll have a product personalized for you and, thus, won’t care about mine. Furthermore, personalization will increasingly be expected, rather than a point of differentiation, so value will have to be found in other places than the promise of customization. Increasingly, I think the point of differentiation will be the quality of the experience, and algorithms will drive that.

The second way value improves is in terms of time: As algorithms make shopping more efficient and less time-consuming, consumers will get time back in their lives. This will create new opportunities for brands to deliver new and innovative things for people to do with this reclaimed chunk of time.

What are some of the ways marketers can pivot to this new algorithmic reality?

The biggest pivot will have to be a better understanding of consumer decision-making that is algorithmically enabled. Even with the digital transformation of the marketplace, most marketing models treat consumers as if they are making decisions wholly on their own. In fact, consumers are using algorithms to navigate the decision process, and marketers will have to learn these new realities.

In particular, to the extent that algorithms are predictive, marketers will have to learn how to influence algorithms because these predictions – not consumer decision-making – will drive purchasing. More specifically, this means influencing the profiles of consumers that predictive algorithms rely on for matching against product options. This is not something marketers have done much of, so it will require pivoting to a new set of tasks, priorities, and metrics.

Is it necessary for marketers to adapt to the new algorithmic reality in order to succeed?

Without question, yes. Algorithms will drive decision-making, and marketers have always had to adapt to whatever influences decision-making.  In that sense, algorithms are a new form of the same old challenge. However, algorithms are more than that, because they don’t just sway decisions, they actually commandeer them. Marketers will have to learn how to advertise to algorithms in order to succeed.


About J. Walker Smith

J. Walker Smith is Chief Knowledge Officer for Brand & Marketing with Kantar Consulting. Walker has been described by Fortune as “one of America’s leading analysts on consumer trends,” and he is the co-author of four highly regarded books, a columnist, a blogger, an avid tweeter, and a former public radio commentator. He is a 2012 inductee into the N.C. Advertising Hall of Fame. In 2014 and 2015, he won the WPP Atticus Award for Strategy and in 2013 for Consumer Insights. Walker holds a doctorate from the University of North Carolina at Chapel Hill. You can follow Smith on Twitter and LinkedIn.


Please enter your comment!
Please enter your name here