Separating Better Data from Big Data: Where Analytics Is Headed

WCAI data analytics

mic Listen to the podcast:

Wharton's Eric Bradlow, Peter Fader and Raghuram Iyengar discuss what's next for customer analytics.

Ten years ago, the most forward-thinking companies were just starting to dive into the potential of data and analytics. Since then, brands have moved from using analytics to answer what customers are doing to exploring the how and why, and also to figure out what they will do in the future.

The Wharton Customer Analytics Initiative (WCAI) is celebrating its 10th anniversary this year and has seen every step of that evolution. Knowledge@Wharton recently sat down with Wharton marketing professors Eric Bradlow, Peter Fader and Raghuram Iyengar to discuss how the field has developed over time, and what they expect to be the key trends over the next decade. Bradlow and Fader are the founding directors of WCAI, and Bradlow and Iyengar are the current co-directors.

An edited transcript of the conversation follows.

Knowledge@Wharton: When WCAI launched ten years ago, what was the environment like in the field of analytics? What were some of the initiative’s key goals?

Peter Fader: Even though the initiative started 10 years ago, the idea started about a decade before that. I had been doing a lot of work primarily in the music industry serving as an expert witness for Napster and just seeing the dysfunction that was, and to a large extent still is, the music industry. I vowed as the new century came around that we’re going to bring some rigor and accountability and just Wharton smarts to not just music, but also movies and television and publishing and media entertainment in general.

It was a frustrating few years as we pounded on the doors of all the different music and entertainment firms trying to get them to participate until 10 years ago when our founding donor, Art Bilger, said basically, “If you keep the M (media) but drop the E (entertainment), I’ll give you a whole bunch of money.” So we basically said let’s drop a primary focus on entertainment and make it more about media. And that’s when Eric and I started the Wharton Interactive Media Initiative, or WIMI. And that’s what really got things going. Maybe Eric can pick up the story from there.

Eric Bradlow: It turned out that two to three years later, that became a limiting name as well, because while it was the right decision for Pete and I at the time — because again, interactive media and digital advertising was the most advanced area of analytics — as we started growing to industries like pharmaceuticals, telecom and others, they said, “We’re not in the interactive media business.”

And so … on January 11, 2011, the Wharton Interactive Media Initiative no longer existed and it became what we know now as the Wharton Customer Analytics Initiative. It was the right change because today we work with companies from all industry verticals, from health care, telecom, nonprofits, consumer packaged goods, etc., because everyone has the need for analytics and the use of data in that way.

Knowledge@Wharton: Every firm amasses a lot of data and is increasingly trying to collect and interpret that data. How have the companies’ views of that data and analytics technology changed?

Raghuram Iyengar: I think there has been a change in the sense that when Eric and Pete started this organization 10 years ago, there were many companies out there who were looking outside to see what are other academics were doing, what were the best practices. What has also happened is that there is a lot of stuff happening inside different companies. In some sense what you see now is that there are many companies who have their own analysts. They are thinking about what are the problems that they want to solve at the company. At the same time, there is a limited bandwidth – the in-house analysts can’t work on everything. So that’s where I believe that our organization comes in. We still act as matchmakers to talk about how those companies can, in some sense, leverage all the different academics or the students that we can match them with.

Bradlow: Yes. I would add two things to what Raghu said. One is the way Pete and I have always described WCAI from the beginning — that we’re the Ph.D. bench strength most companies wish they had. And the good news is a lot of companies still want that. But building on Raghu’s point, the set of companies that have very little analytics capabilities is shrinking — and that’s a good thing. It’s shrinking fast. Which is one of the reasons why my second point is that we have broadened the mission of WCAI. When Pete and I started WCAI, it was really about research; research was the centerpiece. Research is still the centerpiece, and we’re a research center. However, now that companies have gotten much more sophisticated in their use of analytics, we’ve started broadening what we do at WCAI to include, of course, students, alumni, executives, etc.

Knowledge@Wharton: Because the barriers to having analytics technology, the barriers to being able to analyze data are lowering and it’s becoming cheaper to acquire this technology, what are some of the non-tech “soft” skills that have become really important to do this right?

“The set of companies that have very little analytics capabilities is shrinking — and that’s a good thing.”–Eric Bradlow

Bradlow: In a recent article in Poets and Quants I actually talk just about that issue. As a matter of fact, Pete and I had this discussion a long time ago. I’m glad that [the School] has agreed and has kept it this way today…. You might say, “Why doesn’t Wharton have a masters in analytics program?” Pete, Raghu and I think it would be a big mistake for Wharton to go down that path.

What every company wants is someone who knows business, has the softer skills, but also knows analytics. If you want to think of it, we always talk about this Venn diagram. On the one side you have people with business knowledge. Well, there are a lot of graduating MBAs. There are a lot of people who know analytics. But people that sit at that intersection — that intersection really has not grown as fast as I thought it would….

Iyengar: Just building on Eric’s point I think this idea of what we call as translators, people who are well versed in the analytics part of it, that’s great. But they also know how to manage teams. They know how to manage people who are also doing the analytics. One of the things that we’ve done with WCAI is run a very successful executive education program where we talk about three things: tools, talent and metrics. And this idea of talent, how do you hire the right set of people? There is a difference between, let’s say, a data analyst, a data engineer and a data scientist, and a person who can in some sense talk to all of them.

Knowledge@Wharton: How are the questions that companies are trying to answer with data changing?

Fader: They’ve gotten more diverse. It used to be that we’d be appealing just to those geeks and nerds within the marketing organization. But we’re seeing that analytics is a great way to break down some of the barriers. And we’re seeing just genuine conversations happening between the marketing folks and the CFO office and research and development, talent management, and so on…. We’re also seeing a greater sophistication in the skills that they have. It used to be, “Can we do this for them?” And now it’s, “Can we help them do it better on their own?”

Bradlow: Pete’s being extraordinarily modest here, and let me say why. I think both his book Customer Centricity and his research on customer lifetime value have really been the part that has stuck the most over time. I think customer lifetime value, which Pete spends half of his research time thinking about, really is a unifying framework that brings together people in marketing, who are asking, “How do I spend to raise customer lifetime value?” But the CFO [also] gets it. You make money one customer at a time. You might have millions of them. Even after he moved on from being a formal co-director of WCAI, customer lifetime value, the work Pete has done in customer centricity, is always our starting line, because it’s the thing that resonates the most with people in different job functions.

“There is a difference between, let’s say, a data analyst, a data engineer and a data scientist, and a person who can in some sense talk to all of them.”–Raghuram Iyengar

Knowledge@Wharton: Raghu, what would you say are some of the biggest misunderstandings that companies have about data, that they have about analytics even today?

Iyengar: I’ll tell the story in two ways. One is that somehow just having a good model, having good analytics done, somehow will make it implementable. That’s wrong. You have to have a complete buy in. Everybody in the organization should understand what we are striving for. In that sense, building on Eric’s point about Pete’s work on customer centricity, everybody gets it. The CFO gets it. The CEO gets it. The CMO gets it. Once you have the top management with you I think it becomes easier. So that’s the implementation part of the solution.

The second thing is, in some sense, trying to understand if you start thinking about complexity, models can be complex, but you want to always start with a business problem. I think it’s becoming increasingly easy for many of us, when you have a conversation with companies, to start with a business problem, but that’s something we’ve seen all along, which is many times they get so involved in the problem itself or the solution that they don’t understand the problem.

Knowledge@Wharton: There’s another group here we haven’t mentioned yet, which is the customers. What is the most important thing for customers to know about the collection of data and how it is being analyzed?

Bradlow: That’s one of the issues that we get asked about all the time, in terms of privacy and security. It really depends on your perspective. I’ve never been clear from a researcher’s point of view, or even just as a human being, if there’s a right answer to this question. I can only speak for myself. I’m thankful that customer analytics exist because I like better recommendations when I go to websites. I like targeted e-mails. I like targeting ads. I understand the cost of that means my data is shared. And so I understand that. As the customer, I save time. I save money. And it actually broadens the types of things that I get exposed to, from my point of view. But I understand from a societal point of view [that there are concerns].

“It used to be that we’d be appealing just to those geeks and nerds within the marketing organization. But we’re seeing that analytics is a great way to break down some of the barriers.”–Peter Fader

Actually it reminds me of a story. Pete and I once gave a talk at a legal studies and ethics conference. We were both talking about separate streams of research that we were doing. [Conference attendees asked] what [would happen] if regulation comes around and [companies only have access to] more aggregated data — if companies actually couldn’t have access to that individual-level data? We’ve both done separate streams of research that said actually companies don’t lose as much as you would think.

The argument companies have is, “If we don’t keep the individual-level data we’ll lose lots of money.” It turns out it’s not obvious that’s true. It was probably the biggest standing ovation either of us ever received. And it wasn’t even at a marketing conference! So this is a touchy issue. For me, I like the fact that individualized data exists — and I’m talking about myself as a person [and] as a scholar.

Fader: …I’m just thinking about how this month, the GDPR [General Data Protection Regulation, a new policy that puts stricter requirements on how companies handle personal data] will go into effect in Europe. I worry a lot about that. I worry about the policy makers who … probably haven’t read our research, who don’t really understand what kind of data is useful and what is intrusive and what’s not. And I think there’s a bit of an overreach going on there. I think there are some concerns of that happening in the U.S.

“We need to sort out the kinds of policies and actions that companies have taken from the nature of the data and the analytics that can be performed with it. And that conversation has been all muddled.”–Peter Fader

I think that centers like WCAI are really helpful to bring some clarity, instead of just overreacting to this conversation, so that not only executives, but also policy makers, will attend and listen to the kinds of things that these folks have to say, in order to make better decisions, not just with the customers, but with the overall ecosystem around data as well.

Knowledge@Wharton: In the past couple of months there have been heightened concerns about how some companies are handling user data. What kind of conversations need to take place, either face-to-face, virtually or via marketing between the companies and the customers about how data is being used and what companies are doing to protect it?

Fader: It’s a great educational moment about what kinds of metrics really are predictive and what kinds of metrics are, at most, nice to know. I’ll just speak for myself here. I’m not speaking for WCAI in saying that a lot of the data associated with the whole Cambridge Analytica mess was kind of useless, and it’s not clear to me that they made a lot of effective decisions using it. That doesn’t mean that it’s right — there were definitely bad things done. But I think we need to sort out the kinds of policies and actions that companies have taken from the nature of the data and the analytics that can be performed with it. And that conversation has been all muddled.

Bradlow: I did a talk about this, which I called “Not Big Data, Better Data.”… People assume, “Well, they’re collecting all of this stuff. That just has to be bad.” As Pete pointed out, which I agree with, most of it is not actually predictive of anything. Therefore, they’re collecting, in some sense, a lot of big useless data. And actually one of the things I’ve been thinking about for a long time is what are the kinds of new data sources that companies will be able to collect that will actually be both valuable for the firm, but also possibly better for customers?

Knowledge@Wharton: What do you think some of those might be?

Iyengar: From a customer’s perspective they have to start thinking what are the benefits that they’re getting, tangible benefits? It could be like Eric just said, customized e-mails, customized ads and so on and so forth. But also what are the costs? In some sense, is [the data] being used as a collective? Is it being used with personalized information? How is it being used? What is it being used for? Who is the end user? Is this data being sold to other people? All of these are important conversations to have now, especially thinking about how in the future there might be less of this data. So how can we extract the most information at the least cost to consumers?

Knowledge@Wharton: To get into some of the developing technologies that might make a big difference in collecting and accessing data, the Internet of Things and connected appliances have made it possible for firms to collect data at or near the source. What type of implications does this have for firms and also for the user?

Iyengar: I think the one straightaway, of course, is this issue of privacy in some sense. That’s a given in that if you are [in a store] and you’re suddenly getting [a targeted] e-mail as soon as you leave, clearly there are a lot of implications from the point of view of privacy. But at the same time I think there is this issue of benefits and costs. Some consumers may actually be OK with that because they understand that this data is being collected. In other words there is heterogeneity. I think it’s important for marketers to understand that not every customer is the same. Even things like personalized marketing may work for some people and may not work for others. So thinking carefully about new ways of collecting data, but also understanding that not all customers can be treated alike, is very important.

“From a customer’s perspective, they have to start thinking what are the benefits that they’re getting, tangible benefits?… But also what are the costs?”–Raghuram Iyengar

Knowledge@Wharton: Another thing that’s going on right now is the emergence of AI and machine learning. But that also means that data and analytics are now being analyzed by robots in addition to humans – what does that mean for the development of this space?

Bradlow: There are lots of different what I would call “rote” tasks where computers, artificial intelligence machines and [other applications] can actually replace [humans]. In some sense at the end of the day, we need things that are scalable and automated. That’s true. But there is also still going to be an art to it. So I’m actually not one of those people that believe that all of the area of analytics is going to be replaced by artificial intelligence and computers making those decisions, because again, I think there’s still domain area knowledge that won’t be trivial. And in some sense it has to do with, if you want to get semi-technical for a second, there are always going to be interactions between the consumer and the context that you can’t measure that easily and directly. You can’t just measure it by, if you’d like a statistical term, a four-way interaction between these variables or some complicated decision tree that’s trying to model how the consumer is making a decision.

But wouldn’t it be great if in some sense all the decisions that can be made in a large scale automated way are made in that way? That leaves us humans, who have limited capacity, actually spending time on those more subtle and difficult decisions. That’s how I view it.

Knowledge@Wharton: What do you expect to be the major trends in customer analytics over the next decade?

Fader: Going back to our roots, the tagline for the initiative … was the idea of people doing things over time. That if we can watch who’s buying what or interacting with whom or with a website, can we make statements about what they are likely to do — what’s next? It was always about behavior, and still, to a large extent, it is. I think I can say for all three of us, most of our research is about watching what people have done and projecting what they’re going to do next.

“Imagine taking someone’s physical location and appending it to their transaction history: Now all of the sudden you have the ability to do targeted marketing but based on where you currently physically are. I think that’s going to be a big part of the future.”–Eric Bradlow

But the really cool stuff coming up is the stuff that isn’t necessarily behavior: Neuroscience. And when we were first setting up it was kind of Star Wars far out there, “this will never happen.” But, it is happening and it’s happening fast…. Once we can really get inside people’s heads and once we can really integrate what people are seeing and thinking and planning, it brings a whole new dimension to the kinds of data that we have and therefore the kinds of analytics that we’d want to use and the kinds of decisions that firms would make.

Iyengar: Just to complement what Pete said, I think there are two [additional] future directions. When you think about broadening the scope of analytics, one is neuroscience. Another, potentially, is looking at culture within organizations [and asking] can you quantify that? A third one is not just about the diversity of applications, but what can [an organization like] WCAI do for analytics? It is training the next generation of people to handle those kinds of problems, and we will have access to all this wonderful data that can help train the next generation of translators.

Bradlow: Pete mentioned our tagline, “people doing things over time,” which is still our tagline. I like to use a different quote though, which is if you think about the famous expression about what marketing is — it’s delivering the right product to the right person at the right time. And I’ve always said at least in the last 20-30 years we’ve been able to do two of those. We reflect heterogeneity, we’ve got that. We can make product recommendations because of your purchase history. But the right time we were unable really to do. In the next 10 years, spatial data and people’s geospatial position is going to be big. Whether that’s where you are located in a shopping mall or some other location, the same way that we track people online and people are re-targeted based on where they are online, we’re going to be able to do the same thing out in the physical world.… Imagine taking someone’s physical location and appending it to their transaction history: Now all of the sudden you have the ability to do targeted marketing but based on where you currently physically are. I think that’s going to be a big part of the future.

Citing Knowledge@Wharton

Close


For Personal use:

Please use the following citations to quote for personal use:

MLA

"Separating Better Data from Big Data: Where Analytics Is Headed." Knowledge@Wharton. The Wharton School, University of Pennsylvania, 10 May, 2018. Web. 12 November, 2018 <http://knowledge.wharton.upenn.edu/article/where-analytics-is-headed-next/>

APA

Separating Better Data from Big Data: Where Analytics Is Headed. Knowledge@Wharton (2018, May 10). Retrieved from http://knowledge.wharton.upenn.edu/article/where-analytics-is-headed-next/

Chicago

"Separating Better Data from Big Data: Where Analytics Is Headed" Knowledge@Wharton, May 10, 2018,
accessed November 12, 2018. http://knowledge.wharton.upenn.edu/article/where-analytics-is-headed-next/


For Educational/Business use:

Please contact us for repurposing articles, podcasts, or videos using our content licensing contact form.