Wharton@Work

July 2023 | Marketing

Big Data Isn’t Better Data: What’s Wrong with Analytics

Marketing professor Raghuram Iyengar, who leads Wharton’s Analytics for Strategic Growth: AI, Smart Data, and Customer Insights program, says although most organizations are collecting many kinds of data, using it to find and create growth opportunities — and to make better decisions — requires a targeted approach. He advises starting small, keeping an open mind, and matching your data strategy to real challenges your business is facing.

“Big data isn’t better data for a couple of reasons,” says Iyengar. “There’s a race going on now, where companies are trying to collect as much data as possible. But I like to think of data as the new oil. Striking it isn’t enough. For oil to be useful, you have to extract it and connect it to the end user with pipelines. It’s the same for data. Some companies have a lot of it, but they don't know how to extract it or how to build those connections. That's one reason.”

The second, which Iyengar says is also challenging, is that not every type of data is useful. “Sometimes organizations collect data for the sake of data — they’re drowning in it but don’t know what to do with it. It’s not tied to any of their KPIs [key performance indices]. A better approach is more targeted, collecting the types of information that can help target consumers or make your products better, for example.”

Who’s Winning the Data Race — and Why

The pandemic, like other massive shocks to the economy, helped to dramatically shift consumer behavior. With many stores closed, people had to quickly figure out how and where to get what they needed. While online shopping is a part of that story, it’s not the whole story. Marketers had to figure out how to reach their customers in new ways, both in terms of messaging and in the channels they used to relay those messages. The ones who succeeded, says Iyengar, were the companies that were already very proactive, with a test-and-learn culture. “They were never satisfied with knowing what they know right now, but worked to understand where things are heading. Those companies clearly understood that things were changing, and worked to get ahead or keep pace with those changes through constant experimentation.”

But another really interesting thing happened, he says. “Many of the companies that were already collecting a lot of data and continued to rely on their models to help them make decisions got hurt. Why? Their models unfortunately were not very helpful, because during the pandemic consumer behavior was very different, so the data collected were also very different. So all those models, all that automation that they may have been doing before, using it without thinking too much about what kind of data it was trained on, actually became a disadvantage.”

“The companies that were nimble, who were testing and learning, did much better,” he continues. “They were willing to say that even though they didn’t have all the answers, using data to reduce at least some uncertainty was better than none at all.”

Using Data to Improve Decisions

During the pandemic, Iyengar and fellow marketing professor Dave Reibstein conducted video interviews of C-suite executives to learn how they were handling the unprecedented situation. The CMO of Hershey talked with them about the demand for candy, which peaks twice during the year: over the summer and around Halloween. “The summer was difficult for them because it was just a few months after stores closed,” says Iyengar. “But as fall approached, they understood that they couldn’t rely on sales from the big retailers that consumers traditionally bought candy from. Instead, they experimented with their ad budget, and put a greater focus on smaller retailers.”

“Hershey was very nimble in trying to figure out how they could get inventory to the right places, and do the right type of advertising through the right communication channels to see which ones might be more appealing,” Iyengar says. “That really reflects a company that is purposeful, thinking carefully about how to use data to meet a specific end goal, during a situation that was clearly not normal.”

Agility and Preparedness Are Key

Iyengar says if there is a silver lining to the pandemic, it lies in the lessons we can learn from companies like Hershey. There will be other shocks to the economy in the future, and the clear winners will be those organizations that understand the need to be nimble. “When I was in India for a vacation, I spent time in the north which borders Pakistan on one side and China on the other. I saw many Indian soldiers, in very difficult conditions, constantly doing drills because while nothing may be happening at the moment, they want to be prepared. Routinely going through those drills makes it part of the culture, ingrained like muscle memory. Then when a crisis hits, they don’t have to figure out what to do.”

“Companies that have a culture of learning, constantly questioning how resilient they are, constantly questioning whether they can survive different types of threats, will emerge as the winners when the next issue comes around,” he explains. “As a leader, you may not know the answer, but you have to be open to what the data is telling you. Having that ability in the organization, starting at the top, to constantly learn, to question their intuition and operate with transparency, is what will make the difference.”

Marketers Need to Look Ahead

The pandemic accelerated marketing changes, including the proliferation of new channels and changing consumer preferences. With widespread brick-and-mortar closures, new priorities, objectives, and methods including digital capabilities had to be quickly created and adopted. Iyengar says as a result, the role of CMO is being reevaluated. “It’s not just about running ads anymore, or talking to Madison Avenue executives. CMOs need to be looking at digital and have a good understanding of what type of data to collect. They might also work with the chief governance officer to figure out what kinds of data should and should not be collected.”

To succeed, “you can’t think only about what works today,” he continues. “You need people who are thinking about the next TikTok video they should be producing, and at the same time be open to potential new developments. A learning mindset says, ‘I know what I know, but I also know there are, as Donald Rumsfeld famously said, unknown unknowns, and I need to be prepared for them.’ That means you need to constantly learn from data. Don’t be content with the level of knowledge that you have now — be open to new ways in which things might be disrupted in the future.”

The Digital Playing Field Is Wide Open

Another lesson comes from the large incumbent firms that didn’t leverage their data well during the pandemic. “Many of those who were at the top in terms of collecting and using their data at the end of 2019 actually saw their market share drop during COVID-19,” says Iyengar, “because they didn’t see and embrace the changes that were happening with their consumers. And there were companies at the bottom that moved toward the top. The incumbent advantage suddenly went away. Those top companies were happily complacent, resting on their laurels and not looking forward. Nimble companies with lower market share were able to use the situation to their advantage.”

“The fact that things didn't go well before doesn't mean that they will never go well,” he explains. “Firms who are now ramping up their data capabilities, and who stay open to and prepared for more seismic changes, can have an advantage over larger companies that have a harder time changing course.”

To do that, Iyengar advises starting small. “Think about the strategy and tactics that will move you forward. While you may have a great vision of digitally transforming the company, choose a couple of key decisions that you would like to make differently, and one or two tangible problems facing your company.”

“That way, you nibble away at small things first, understand how to go about it, and then replicate it for bigger things. Rather than going from zero to 100, you go from zero to 10. See what worked and what didn't. Then it becomes easier to move from 10 to 100. Don't get overwhelmed. Start small, experiment a bit, understand what's happening and what's not happening, and then move forward.”