How Data Science Can Help Marketers Manage Unprecedented Disruption

November 5, 2020

Listen to the interview:


The COVID-19 pandemic has caused major disruptions for businesses, spanning nearly every sector and encompassing everything from consumer buying habits and supply chain concerns to human resources and team management. As companies seek to adjust and predict what “the new normal,” will look like, data science could help them navigate the choppy waters ahead. The Marketing Science Institute’s November 17 annual analytic symposium will explore how the use of analytics and artificial intelligence will figure into a post-pandemic world.

MSI’s partners for this year’s conference are Wharton Customer Analytics and AI for Business at The Wharton School of the University of Pennsylvania. Recently, Barbara Kahn, executive director of MSI and a Wharton marketing professor, sat down with Kartik Hosanagar, Wharton professor of operations, information and decisions and director of AI for Business, and Raghuram Iyengar, Wharton marketing professor and director of Wharton Customer Analytics. They discussed how AI and analytics can be used for governance and risk mitigation, and how algorithms can be used to manage disruption in unpredictable times.

Listen to the podcast using the player above. An edited transcript of the conversation follows. 

Barbara Kahn: Kartik, you are  one of the key speakers at the symposium. Can you give us  an overview of the key topic areas that the event will cover? 

Kartik Hosanagar: Overall the event is focused on the use of analytics and AI in a post-pandemic world. We’re looking at three different topic areas for the event. The first one is looking at risks associated with the application of analytics in marketing and related areas, and looking at how we mitigate those risks, and what are some governance frameworks that analytics or marketing professionals should think about. 

Another area is around diversity and inclusion as it applies in an analytics-driven world. When we’re using historical data to make decisions, it’s easy to pick up some of the biases that have existed in the past, and to institutionalize those biases in the form of algorithms that are automating decisions. And so that’s a really interesting area that we’re looking at.

The last one is around data mindset, and that one’s focused on looking at the disruptions that have happened over the past six to nine months, and the role of analytics and AI in helping managers and businesses navigate these disruptions. 

Kahn: Raghu, why should marketers care about all of these topics? 

Raghuram Iyengar: AI is not just cutting across thinking about how to hire people, but it’s also about making marketing more efficient, and perhaps thinking about how AI and machine learning are changing the customer journey. When you talk to marketers, they’re always thinking about where customers are, what they are looking for, and how to get them to close out buying the product. AI plays a role in every part of that journey. 

The disruption of the last six to nine months has fundamentally changed how customers behave. How are they shopping? Think about the retail store, for instance. Few people are going in there. So how has AI and all the digital transformation that is going on changed the way customers are interacting with companies, and what can companies do proactively to help customers navigate through this pandemic?

Kahn: Kartik, can you build on what Raghu was talking about and talk about some of the opportunities that you see from your perspective? 

Hosanagar: First of all, at this point everyone recognizes that we’re living in an information or a data age. There’s so much information available about consumers. I think that’s changing consumer expectations, and how they want to interact with companies. Raghu’s example of retail is a great one to look at, where companies are seeing customers across so many different channels. Sometimes you might see them at the store, but increasingly you see them online across many different devices.

How do you give a unified, and a consistent experience to the customer across all these channels? How do we personalize the customer’s journey? Because I think as consumers, we’re all now in a world where we expect one-to-one marketing. We don’t want to see the same messages as our peers, we don’t want to see the same products as our peers. The classic example is Netflix. You log into Netflix, even people in my own household do not want to see the same titles that other family members see when we log in. We all want to see a completely different interface, a completely different experience. 

That is the classic challenge and opportunity for companies. The successful companies will be those that recognize this, embrace data, embrace analytics and machine learning, and present consumers an experience where it feels like, ‘Hey my company is talking to you on a one-on-one basis, and it’s about you and me and not the millions of other customers we have.’ That is  the game changer, and the opportunity for marketers.

Kahn: Raghu, can you tell us about some of the challenges or caveats in this area? 

Iyengar: Let’s take the examples that Kartik mentioned, which are very interesting and important examples. Consider  Netflix or Spotify, for instance, which is well known for having some amazing data scientists. If you think from a customer perspective, things that we were watching, things that we were listening to six or nine months ago, where we were listening to them, where we were watching them, that is perhaps different now. A lot more people are staying at home, and their habits have changed. 

Now the question for many of these companies is, have these habits changed fundamentally? How sticky are these habits? Are they going to revive as we recover from the pandemic? Quite frequently what you find is that machine learning models obviously learn over time, but they rely a lot on historical data. As companies think about all the treasure trove data that they had, they need to consider how much of it can be carried over and how much of it needs to be recalibrated, so to speak, as customers change their behavior. 

Hosanagar: There is another kind of risk here, which is that of biases that might creep in based on what you learn from data. One example I’ll provide is, right now Facebook is facing a lawsuit from the U.S. Department of Housing and Urban Development focused on algorithms that allow marketers and others to target their messages based on cultural affinity, religious affinity, and so on. That particular lawsuit relates to housing-related ads, and the ability to target them based on zip codes and ethnic and cultural affiliations. And so that’s an example where it could easily be misused. 

Another example comes from companies that are trying to use chatbots to interact with customers. You are automating customer service, which is great, and you get instant customer service as a consumer, but that can also backfire as happened when one of Microsoft’s chatbots turned racist and sexist in conversations with people online. I think the risks are manifold, and clearly there is a big governance question that arises as a result. 

Kahn: Raghu, do you have anything to add? 

Iyengar: I think some of the work that Kartik has done, which he will showcase in the conference, will have a lot of things that people might take away. Some things that they have to think about are, obviously not every company is as big as Microsoft, and so what can they do in their own way? There might be some small, low-hanging fruit that they can think about. 

But at the same time, one of the takeaways that companies should be having is that AI and machine learning should be part of the bigger strategy as well. That means not just thinking about one or two little things that they can do, but thinking about how it integrates into bigger decisions they have to make.

By using you agree to our use of cookies as identifiers and for other features of the site as described in our Privacy Policy.