Marketplace is community-funded public service journalism. Give in any amount that works for you – what matters is that you give today.
Making data-driven decisions has seemingly never been easier. We’ve got pulse surveys, performance analytics, reviews, anecdotes on social media — all just a click away.
And yet … all these inputs aren’t really helping us make better decisions.
That’s according to a new study from the software company Oracle, which surveyed workers and business leaders around the world.
Marketplace’s Meghan McCarty Carino spoke with Seth Stephens-Davidowitz, a data scientist and author who partnered with Oracle for the study. He said many business leaders would like to just outsource much of their decision-making to a robot.
The following is an edited transcript of their conversation.
Seth Stephens-Davidowitz: People feel like they’re drowning in data, businesses feel like they’re drowning in data. And sometimes instead of helping them make decisions, they’re feeling paralyzed by all this information and all this data.
Meghan McCarty Carino: How has our reliance on data in this context sort of mounted over time?
Stephens-Davidowitz: So the other thing we found in the survey is that most people, most business leaders, believe in the power of data. The vast majority think that data could help them make better decisions, so that’s one of the reasons they’re collecting so much data. They’re looking at so much data, more and more in the past few years. But it feels like too much to many people, to many businesses.
McCarty Carino: Yeah, tell me more about how this is affecting decision-making by individuals and organizations.
Stephens-Davidowitz: Well, a bunch of people in our survey say that it’s leading to anxiety, to unnecessary spending, to lower confidence, but generally what we refer to as decision dilemma. Eighty-eight percent of people said they’re suffering from decision dilemmas, not knowing what decision to make, because a lot of the information seems to be irrelevant or conflicting.
McCarty Carino: And so what is the impulse that many in the survey had, confronted with that kind of data indecision?
Stephens-Davidowitz: Yeah, the impulse was to throw away the data. One of the more surprising things we found in the data is that 70% of business leaders say they prefer for a robot to make their decisions. So I think a lot of people, because they feel like they don’t know how to use all this data to make decisions, they see, you know, ChatGPT and other [artificial intelligence] tools. They’re like, well, maybe they can just do this, they can make sense of all this data, because we sure don’t know how to. I don’t think we’re at the point where, you know, the average decision we have to make, we could just type it in ChatGPT and it will tell us what to do. You know, these tools are prone to hallucinations, making up facts. But I understand why people are excited by that prospect because it seems like a solution to these decision dilemmas people are facing.
McCarty Carino: Are there any use cases for that kind of technology to make decisions for us that you think are particularly promising?
Stephens-Davidowitz: I think ChatGPT right now and some of these tools are better at giving us more ideas. But as this survey said, people feel like [they] aren’t always helped by that. So let’s say you’re deciding a marketing plan. ChatGPT can give you, you know, 10 potential taglines for your marketing plan. But maybe that’s just more data, and people will drown in all the ChatGPT data.
McCarty Carino: I think even outside the business context, that was sort of the purview of the survey — everyone is feeling this in their lives. You know, when I go grocery shopping, I’m like, I have to look up 20 reviews of the brownie mixes to figure out which one is the best. I mean, what is your advice for people navigating this kind of data synthesis in our daily lives with so much data available to us these days?
Stephens-Davidowitz: Yeah, I think part of it is to know whether it’s a fundamentally important decision and you have to collect more and more information. You know, if it’s just a brownie recipe, maybe you don’t need to collect 20 sources of data, if it, especially if it’s going to drive you totally insane.
McCarty Carino: There’s kind of an opportunity cost there.
Stephens-Davidowitz: An opportunity cost, that’s a good term. But I think more generally, be OK with conflicting information. The world is confusing. I think we expect the world absent data sometimes gives us a false precision, you know. It’s very easy to pretend that everything points in the same direction, and that’s not the way the world works. And data maybe in some ways reflects the world more accurately, where there is sometimes going to be different information and conflicting information. That’s fine. And if 10 data points point in this direction, and three data points point in the opposite direction, go with the one the majority points. Don’t freak out that not all of them point in the same direction.
It seems like too much information isn’t just a consumer or business problem, but also a cosmic one.
Earlier this year, MIT Technology Review reported that the powerful James Webb Space Telescope, which is roaming the cosmos and gathering information on different stars, planets and galaxies, is actually sending back too much data, to the surprise of NASA scientists.
One compared the stream of data on galaxies and star formations to a “firehose” of information that has led to an avalanche of new scientific papers going public as preprints.
But when it comes to understanding the vastness of space, maybe even too much data is never enough.
Every day, the “Marketplace Tech” team demystifies the digital economy with stories that explore more than just Big Tech. We’re committed to covering topics that matter to you and the world around us, diving deep into how technology intersects with climate change, inequity, and disinformation.
As part of a nonprofit newsroom, we’re counting on listeners like you to keep this public service paywall-free and available to all.