With marketers’ rush to social media, we often see strategies that are adopted across categories without much attention to how well they are aligned with the way consumers make decisions in a certain domain. Inviting fans to “join the conversation” can be perfect in some categories, but the potential impact of engaging consumers in a conversation about paper towels is limited at best.

Things work differently in different categories. The shifts we describe in this book will not happen evenly across the board. There are areas where they are in full swing, then there are areas where they are progressing very slowly, and there are domains they will most likely never reach. We don’t expect these trends to apply in the same way to cars and paper towels, to well-connected and to less connected consumers, and to decisions made with or without time pressure. In this chapter we introduce a framework—the Influence Mix—that lays the foundation for this discussion. Simply put, this framework should help marketers determine the relevance of the trends we’ve described to their particular situation. From brands losing their role as proxies for quality to the declining effectiveness of persuasion techniques—the extent to which these trends apply to a particular firm depends on its customers’ Influence Mix.

We start with a simple idea: A person’s decision to buy is affected by a mix of three related sources:

For example, when you buy a new cell phone, you’re influenced by your prior attitudes, habits, and pre-stored information (P), by your friends, reviewers, and experts (O), and by the cell phone marketer (M). Let’s briefly discuss the main characteristics of each of these sources.

Reliance on information sources is, of course, a function of their costs and benefits, and in the past decade we’ve seen a significant rise in benefits and a simultaneous decline in costs of information from O, which has drastically increased the relative contribution of this source.2 The main winner is O and the main loser is M because in most cases O is a direct, dominating substitute for M. Marketers’ self-interest makes them inherently suspect, whereas O (despite its limitations) offers much richer and more credible information about products.

What about P? Why doesn’t P take over the mix more often? To understand the answer, a few more words are due about its vagueness. There’s a great scene in the Marx Brothers movie Duck Soup in which Chico tries to pass himself off as Groucho. He places a fake mustache under his nose and sticks a cigar in his mouth, but one of the characters, Mrs. Gloria Teasdale, is skeptical because she just saw the real Groucho leave the room. To persuade her that he indeed is Groucho, Chico delivers a killer argument: “Well, who you gonna believe, me or your own eyes?”3 As it turns out, this line is not as absurd as it might sound. There are experiments that show that sometimes we are willing to rely so much on other people, that we are willing to push aside not only our preferences, but the evidence in front of us.4 And plenty of research shows that people’s beliefs and preferences are often ill-defined, so information from O may overshadow prior beliefs and perceptions.

We can think of the impact of O in terms of the classic economic distinction between “search goods” and “experience goods.”5 The quality of search goods and their attributes, such as the beauty of flowers, can be inspected and determined before purchase, whereas experience goods and their attributes, such as cars, movies, and new types of food, can be assessed accurately only after gaining personal experience.6 Now, if reliable O provides accurate information about quality, then more and more products become effectively like search goods, for which quality can be assessed before purchase. In fact, in many (but certainly not all) cases, the information available before purchase is a better predictor of an individual’s long-term consumption experience than that person’s own initial personal experience with the product. For example, a car may have an unexpected problem during the first week of ownership, but its reliability ratings, for example, in J. D. Power and Consumer Reports, are likely to be better predictors of the consumer’s long-term experience with that car.

If you’re looking for a summary of how influence is transforming in many categories, here it is: Since P is vague and unstable, other sources usually capture “decision share” from it. In the past, M gladly played this role, but with the rise in availability of O (which is perceived as more informative and reliable), O is taking over.

Think about how you feel about cilantro, licorice candy, or jazz. Your preferences for these things are probably stable and very clear to you. Could someone simply trick you into liking or disliking these things? Probably not. You either like jazz or you don’t. The existence of stable preferences sounds pretty obvious, but for the past few decades, researchers in the area of behavioral decision theory have raised serious doubts about the existence of any such stable preferences (except for repeat, habitual decisions and brand loyalty). Instead, the common belief in the field has been that preferences are typically constructed when decisions are made, and then they usually dissipate. Yet in some cases, stable preferences play a bigger role than they get credit for.

Incidentally, the roots of theses stable preferences may surprise you. The results of a study that Itamar recently conducted with Aner Sela from the University of Florida suggest that people’s preferences for jazz, chocolate, sci-fi movies, compromises, and (maybe) hybrid cars appear to have a large heritable component. No, people don’t have a Prius gene, but the study suggests that a certain combination of heritable traits leads a large group of consumers to behave in a similar way. What does this mean? It means that part of your liking (or disliking) of dark chocolate doesn’t come from your experience with chocolate and from M and O influence. Rather, it may be in your DNA. (Notice the words “part” and “may.” Nobody is arguing that chocolate preferences are totally genetic. And it will take years before we fully understand the exact meaning of these results.)

When you walk into the supermarket, many of your routine or habitual purchases are dominated by P. Getting milk, bread, and eggs into your shopping cart is a matter of habit and is affected mainly by your prior preferences, leaving little room for M or O influence. Of course, at some point P may have been shaped by M and O. Your long-term preferences for chocolate are also influenced by chocolate inputs in the environment—both the actual chocolate and the abstract concept of chocolate. Everything people around you said about chocolate. All those Hershey’s or Cadbury commercials you’ve seen over the years. So advertising (M) and your friends (O) can, over time, help shape your preferences (P). These external forces can also affect whether your hidden preference tendencies will see the light of day. For instance, while many video game players had the potential to enjoy a motion-sensitive remote, without the creative idea of the Wii inventor that preference would not have come to light.

Our main point is this: With all due respect to O, marketers should understand its proper role in each category. While people may, once in a while, talk about paper towels, their choice is not significantly influenced by their peers in this domain. Even in categories that people discuss more often, the chatter doesn’t always mean that O dominates the decision. Here’s an example: Around the turn of the millennium, everybody was talking about Krispy Kreme doughnuts. Nicole Kidman reportedly referred to them as “God’s gift to doughnut lovers” and, according to Fast Company, Willard Scott of NBC revealed that he worships in “the church of Krispy Kreme.” At first, this may seem to imply that the Influence Mix was dominated by O, but think about the way consumers make decisions in this domain. While it’s true that people tell their friends about a new doughnut place (and we’ve seen a few doughnut shops with more than a thousand reviews on Yelp), by and large people don’t decide on getting a doughnut by reading reviews, listening to Nicole Kidman, or conducting extensive research. Having lots of people buzz about your product can certainly be a good thing, but it doesn’t automatically mean that O is a major component in consumers’ decisions. When it comes to doughnuts, decisions are influenced by things like prior preferences, store location, and promotional activities, which means that managers like those at Krispy Kreme should not be blinded by the buzz about their brand (which is bound to die out, as it did around 2004 for Krispy Kreme) and should continue to invest in marketing activities such as coupons and other promotions.7

When O or P don’t play a major role in a purchase decision, it means that room is left for M influence. Think about buying a toothbrush, for example. Standing at the drugstore in front of an aisle of toothbrushes, a shopper is likely to be influenced by whatever is at her eye level, the packaging of the product, and by brand. So makers of toothbrushes are likely to operate under different rules than carmakers, for example. Yet too often we hear sweeping generalizations about “marketing” regardless of the product category.

The critical questions for you to ask are: To what extent do my customers currently depend on O and on specific O types or ingredients in making their purchase decision, and to what extent might they depend on O in the future? Since we’re not dealing wit...