The old saying goes that variety is the spice of life. A new study co-authored by Wharton marketing professor Jonah Berger looks at how this plays into consumer decision making. He and his co-authors analyzed millions of supermarket purchases and examined the impact of circadian rhythms to see if the amount of “spice” customers desire is influenced by what time of day they go shopping.
Berger recently talked with Knowledge at Wharton about his paper, “Does Time of Day Affect Variety-Seeking?” which was co-authored with Kelley Gullo, a Ph.D. candidate at Duke University’s Fuqua School of Business and Duke marketing professors Bryan Bollinger and Jordan Etkin.
An edited transcript of the conversation follows.
Knowledge at Wharton: What is the key question that you were trying to answer with this research?
Jonah Berger: You teed it up perfectly: Does the amount of variety we choose depend on the time of day? So, whether we’re shopping at the grocery store and picking yogurts, whether we’re online and we’re buying products, why do we pick more or less variety depending on when we shop? And this question actually came up in a very interesting situation.
A few years ago I was working with Taco Bell, and they were thinking about introducing breakfast. They had served breakfast in the past, but they were reintroducing breakfast. And the question was, “What should the breakfast offering be?” When you think of Taco Bell, you tend to think about really interesting products. You think about the Doritos Locos Taco, which is basically a taco wrapped in a big Dorito. They sort of do edgy, unusual, different things, sort of a rich flavor palette.
When I was talking to them, I was expecting them to say, “Our breakfast is going to be completely different from what everyone else is going to be offering.” But when they started to detail the offering, what breakfast would actually be, it sounded a lot like what we think about an Egg McMuffin, except in a slightly different packaging.
“One of the big drivers of variety is, do I want stimulation or not?”
Knowledge at Wharton: A taco McMuffin?
Berger: Yeah, basically. It was tacos. It was eggs and cheese and potatoes, and maybe a little ham, in a tortilla instead of in a bun. But it was very similar to what the other companies were offering. I said, “Why are you offering the same thing? Taco Bell usually does things differently.” They said, “People don’t like different things in the morning.” I said, “Well, that’s really interesting. First, is that true? And if so, why might it be true?”
And, you know, when you think about what you eat for breakfast, most of us eat the same thing every day for breakfast. We eat cereal every day, or we eat yogurt every day. We eat eggs every day. But we’re pretty happy eating the same thing Monday, Tuesday, Wednesday, Thursday, Friday, for breakfast every day. Yet most of us would hate eating the same thing for lunch every day, or hate eating the same thing for dinner every day. And so is that just a weird example, or might that say something deeper about our preferences for variety?
Knowledge at Wharton: You tested this using circadian rhythms, which I think we all have heard a lot about in its relation to sleep. Can you explain exactly what those are and how it might apply to shopping habits?
Berger: Basically, the biological basis of behavior is what we’re talking about. In addition to normal preferences and likes and dislikes, there are various circadian rhythms — basically a 24-hour cycle in our body of different hormones and other features. Those changes affect how we behave and what we might like and what we might dislike. There are hormones that help us wake up, there are hormones that help us go to sleep, and the changes in these various factors might also have impacts on the choices and judgments we make, even though they weren’t designed to do that.
“There are hormones that help us wake up, there are hormones that help us go to sleep, and the changes in these various factors might also have impacts on the choices and judgments we make.”
Knowledge at Wharton: How were you able to look at this in a real-world setting?
Berger: We started, actually, with a couple experiments. We surveyed people and asked them what they ate for breakfast, and what they ate for lunch and dinner. We asked were the things they chose for breakfast similar to each other, versus lunch and dinner being different? We followed people and asked them to make non-food-related choices. Some people were asked in the morning, some people were asked in the middle of the day, and some people were asked in the evening.
But I think the coolest thing we did is, we actually looked at over a million shopper cards at the supermarket. We looked at a whole bunch of people swiping their card at different times of day. So, whether you were at your Safeway or your Kroger, your Harris Teeter, when you use that shopper card, sometimes you might shop in the morning. Sometimes you might shop later in the day. Do those different times of day lead you to choose more or less variety?
To be very concrete, for example, imagine you’re choosing yogurts. When you shop in the morning, if you’re buying three yogurts, are you buying three vanillas, or are you buying one vanilla and one strawberry and one blueberry? Versus when you shop in the afternoon, might you be more likely to purchase more variety? And so we analyzed these millions of purchases on these shopper cards, and controlling for a lot of other features, we found that the variety chosen did vary across the course of the day. In the morning, we choose less variety. When we’re buying yogurts, we buy more of the same flavor. When we’re buying fruit, we buy more of the same fruit…. Later in the day, we’re more comfortable buying more variety. So in the morning, we want similarity. In the afternoon, we want more variety. Similar to the evening, as well.
Knowledge at Wharton: There are some people — and I’m actually thinking specifically of my husband here — who could eat the same thing every day, every meal. They wouldn’t care; in fact, they would be thrilled. Are these results outside and above people who are just routine people who like to have the same thing day after day?
Berger: In every research there are the averages, and then there’s heterogeneity. There are certainly some people who like the same thing in everything. [Your husband] may have a low preference for variety. Different people have more or less preference for variety. Some people love variety in everything that they do. Other people like less variety in things they do. There’s a preference for variety that varies across people. I would bet, though, even with your spouse, who may not like variety in general, I bet he still chooses less variety in the morning….
I think the intuition behind this, by the way, is actually pretty simple. Think about when you wake up in the morning. Most of us don’t want a lot of stimulation. We want to do something similar to what we’ve done before, because we’re not cognitively awake. We’re not activated. We’re not ready for action. We need our coffee, or something along those lines. In the afternoon, we’ve already gotten there. We’re already awake. We’re happy to deal with more stimulation. And so one of the big drivers of variety is, do I want stimulation or not? In the morning, we tend not to want stimulation so much, driven by our circadian rhythm. In the afternoon, we’re more interested in stimulation, so we choose more variety as a result.
Knowledge at Wharton: In the fall and the winter there’s a lot less daylight. Or there are some people who are morning people, and others who consider themselves night owls. Did these factors have an impact?
Berger: First, some of the listeners might be data savvy and they might say, “Well, you have correlational results in that supermarket study. How do you know it’s causal? Right? Maybe different types of people shop at different times of day. Maybe people have different needs at different times of day. How do you know it’s really a causal impact of time of day?”
We did some experiments in the lab to try to deal with that. We also used different times of year, with the idea being that if part of the reason people choose more or less variety — it’s not driven just by time of day, it’s driven by sunlight. Sunlight helps kick off these circadian rhythms. There has been some research done with people in caves, for example, where they force you to live in the dark for a period of time. Your circadian rhythm is less likely to activate when it’s not cued by the sun. One thing that kicks off that circadian rhythm cycle is the sun in the morning.
At different times of year, there’s more sun in the morning versus less sun in the morning. And so we used that as an exogenous shock to test our results. What we found is consistent with the idea of stimulation. When there’s less sunlight in the morning, the variety preferences change. And so that was one way to test the underlying mechanism based on time of year.
“In the morning, we want similarity. In the afternoon, we want more variety.”
Another you mentioned was morning types versus evening types…. The circadian rhythms of those two groups are different. Their willingness for stimulation — if you’re a morning person, you’re actually OK with stimulation in the morning. If you’re an evening person, you really don’t like stimulation in the morning. And sure enough, we found those individual differences in preferences for stimulation at different times of day moderated this effect. So the folks that liked more stimulation in the morning chose differently than the folks that didn’t, consistent with the idea that the underlying driver is stimulation.
Knowledge at Wharton: Why is it important for consumers to understand this about themselves?
Berger: Whenever I do research I think of two audiences. I think about us as individuals. What does this tell us about ourselves? And as marketers, why should we care about this? As business owners, why should we care about this?
For consumers, starting there, I think it gives us deeper insight into our own behavior, and also when we might incorrectly predict our own preferences. When we shop in the morning versus when we shop later in the day, we may prefer different things, which may have an impact on what we consume later. So in some studies we ask people not what you want today, but what do you want a week from now, or two weeks from now? The time of day they chose impacted that decision, even if they weren’t choosing for that time of day in a week or two.
So if you’re thinking about what to do on a vacation, for example: If you’re choosing in the morning, you might want more similar activities on that vacation. If you’re choosing in the afternoon, you want more diverse activities, more dissimilar activities. When you get there, that may be more or less consistent with what you actually like. And so I think it’s important to be aware of these — not necessarily biases, but these subtle factors that shape our decisions, so we understand what’s shaping them.
Interestingly, in the supermarket data we didn’t see an effect in two categories — baby food and dog food. And so that suggests, at least – we’re not sure for now –that when people are choosing for others, this bias might be weakened a little bit. When we’re choosing for ourselves … we choose based on, in part, time of day. When we’re choosing for our puppy or our cat or our child, we’re less likely to think their preferences might vary, we’re less likely to rely on time of day as a cue to what we might choose.
Knowledge at Wharton: What about marketers? How can they use this research? The holiday shopping season, believe it or not, is coming up pretty soon. How could we use it for that?
Berger: If I’m a marketer and I have a certain type of product, this suggests something about when I might want to advertise that product…. If I sell something that’s highly stimulating, that has a bunch of variety, I might want to focus more on [advertising in] the afternoon or the evening, rather than the morning. And if I’m going after consumers in the morning, I don’t want to focus a lot of attention to the amount of variety my product has, because people aren’t going to be interested in variety in the morning.
“If I sell something that’s highly stimulating, that has a bunch of variety, I might want to focus more on [advertising in] the afternoon or the evening.”
At the same time, if I’m thinking about, “Well, how do I talk about my product or service?” I might want to talk about it in the morning, when I advertise in the morning, in ways that are more similar. In the afternoon, I want to highlight ways that it’s different. In the morning if I have a service, for example, a new software package, in the morning I want to focus on a couple of things that it does. In the afternoon, I want to focus on … a more diverse set of things that it does. And so not just when to advertise my product, potentially, but what to [focus on] when I advertise that product in the morning versus the afternoon.
Or even from a retailer’s perspective, if I’m Amazon, for example, what products do I want to feature on the site? Well, in the morning, I might want to feature a wide variety. We focused on variety, but I could imagine the same effects being true for novelty, just in general. How willing am I to try something new, period? Variety’s really about the assortment of things I’m trying. Novelty’s about, do I want something new, period? This suggests I’m not going to be willing to try new stuff in the morning. And so I really don’t want to give attention to products or services that are new in the morning. I might want to focus on the afternoon and the evening, where people are already a little bit more stimulated.
Knowledge at Wharton: Does this have applications beyond retail? For example, if you’re in health care and you’re offering patients options of something, or situation where you’d be offering somebody a different type of options.
Berger: Imagine I’m a health care provider, and I’m trying to get people to exercise. Should I give them a number of similar activities they could do? So, you could do jumping jacks and you could do walking and you could do running. Three cardio activities that are somewhat similar. Or you could lift weights, and you could row, and you could run. Three activities that seem more different. We haven’t looked at this yet, but our data at least suggests that, at different times of day I might want to focus on one type of exercise plan versus the other. If it’s in the morning, I want to focus on more similar activities, because that will seem more doable to you as a patient…. Whereas in the afternoon, more varied activities might sound a little bit better….
“Interestingly, in the supermarket data we didn’t see an effect in two categories — baby food and dog food.”
Knowledge at Wharton: I was even wondering if there were some sort of applications here for pop culture, for television, for movies?
Berger: It’s interesting you mentioned this. I hadn’t thought about it. But you could think about, if I’m a radio station, what do I play in the morning versus the afternoon? In the morning, I may want to play songs that are similar to one another. In the afternoon, I want to play songs that are more differentiated. I think we talked a couple months ago about some research we had done with songs, where we looked at how similar or different a song is from other songs that are out there. And so yes, songs are all different from one another. But some are more different, some sound more similar. In terms of what mix we want to play on the radio, it might have implications for what mix we play in the morning versus the afternoon.
Knowledge at Wharton: What are some other future lines for this research?
Berger: You know, this research was very interesting to me. I’ve never looked at the biological basis of behavior before. I’ve done some work on variety. But one thing we’ve thought a lot about is, well, what are other domains where this might play out? Our perception of risk at different times of day? We’ve talked about our willingness to try new things versus things we’ve tried before.
One thing we’ve thought about also is choosing for ourselves versus others. We’ve talked a little bit about this already. But, you know, might it vary depending on who we’re choosing for? What if we’re trying to choose for another person, and they’re a morning type, versus we’re an evening type? Might that lead to different preferences and make it harder to find resolution when choosing for others?
Knowledge at Wharton: Or gift buying.
Berger: Gift buying, certainly. Gift buying is always a challenge. It’s always an interesting area to think about. But if we’re buying things based on what we think others will like, and they have different preferences, we may be a little bit misallocated. And so we need to think a little bit more about what they’re going to like. So, that is definitely a different area of research that we’ve started thinking about in this direction.