The problem of giving consumers what they want

After about a year and a half of intermittent reading, I've finished the terrific Thinking, Fast and Slow. At the same time, I've been co-teaching a class on the future of television. I'm in the process of preparing a final lecture for the class, and I plan on drawing from Kahneman to talk about supply and demand in the world of TV (or whatever TV will become, i.e., some system for the distribution and consumption of video online).

The rhetoric of giving viewers what they want, when they want it is a big part of how new TV technologies and content are being sold. As we move from an environment of restricted choice to one of expanded choice, it would seem that consumers of media are more likely to get what they want. How could it be otherwise?

First, let's think about "what they want": how to define that, how to measure it. There have always been feedback mechanisms built into commercial systems, ways in which producers determine demand so as to know what and how much to supply. In the days of early radio and TV, producers heard from audiences via letters. Then came Nielsen, with ever-improving sampling techniques that more closely reflect what people are choosing, though the Nielsen system and the TV choice environment it worked with had limitations: there were only so many options from which to choose; there were some people (e.g., college students) who were difficult to track and so they didn't show up in Nielsen's picture of audience demand. What if we could eliminate these limitations? Wouldn't we have a more pure picture of audience demand to which we could suit our supply?

Technology improved, costs fell, and these trends helped eliminate flaws and overcome limitations in the feedback system. I'd like to raise two issues related to this improving image of audience demand in the TV marketplace. The first is that there are certain attributes of the TV system that aren't overcome by simply monitoring audience behavior more closely and/or removing restrictions on the viewers, giving them more "control" (which is the general progression of improvements in TV technology). The second is that a market in which supply and demand were more perfectly matched may not be desirable.

No matter how good audience behavior monitoring gets and no matter how cheap it is to implement, it will run up against privacy concerns of the audience. As choice expands, we may get a clearer picture of how people behave in any situation without any limitations, and its bound to reveal some ugly truths about individuals and groups behavior. As much as the audience wants to have its desires understood, it wants to be selective about what it shares about those desires, especially in a world in which desires are so closely associated with identity and potential.

Also, even if we could know exactly how people behaved in a choice environment with few restrictions, are we then permanently locked in to what they will desire? There's a wonderful scene in Mad Men where Don Draper reacts to the in-house psychologist who tells him what her focus group observation revealed about audience preference. Don's objection may be to the difference between what people say they want and how they behave, but I think the more fundamental objection is that people's future desires can't be predicted by their past desires. Advertisers and content producers are in the business of telling people what they will want, and as much as that sentiment rubs people the wrong way, it explains something about audience behavior no amount of data or market research can. Of course these are the words of a defensive ad exec asserting his value in an age where empiricism is creeping into an artistic realm. But I think there's some truth to it.

Don is wrong about not being able to predict future behavior from past behavior if he's talking about certain kinds of individual behavior. But the next trend, the next popular TV show, may not reveal itself no matter how hard you stare at people's current or past behavior. Psychologists may know that people pass on certain ideas to other people in certain ways under various circumstances, but the actual ideas they pass on can be set in motion by anyone with the budget to get in front of enough opinion leaders. Advertisers and show runners are just such people. They survive by developing new tastes, new markets for new stuff, and they can still do it in an expanded TV marketplace. So you have this force that will keep dropping ideas that are beyond whatever audiences currently want so as to do this, no matter how accurate the audience demand measurements get. 

Then there's a point made by Jonathan Franzen at a discussion during last year's New Yorker Festival. He discussed how the number of tweets or mentions on Twitter was now being used as a metric of how worthwhile an up-and-coming writer was to a publisher. That is, a publisher would sign a writer who had 100,000 followers or mentions on Twitter and not one who had 100, just because social media mentions and followers are pretty good indicators of present audience demand. This is another instance of the tightening of the feedback loop between creators and audiences. This could force writers to cater to the audience in ways that they did not before. But this idea makes Franzen and other content creators uncomfortable. Writers spend more time on self-promotion and homogenization of their work and less on developing their voices as creators. Better work is produced, so the thinking goes, when creators are not so beholden to current audience preferences (at least the ones that audiences are capable of articulating). The work is "better" not just from some elitist, subjective judgement of its worth but from a market standpoint: if an author or a show runner takes their cues from the Twittersphere, the product will be less pleasing to that very audience in the long run than if the author or show runner listened to their inner muses.

Finally, Kahneman's book brings to mind that the definition of what an individual wants (and the way that the individual ultimately acts based on their desires) depends on various characteristics of the choice environment, namely the timing of the choice, the number of choices, and the arrangement of available options. For example, whether or not there are thumbnails for similar videos on the side of the screen, as there are now on YouTube, may influence what people click on. Recommended videos on one's Netflix screen is another example, or songs that come up on Pandora. Users could've searched for whatever their hearts desired in the search box, and yet what they viewed was influenced by the arrangement of options, not solely the product of a pre-existing, internal set of preferences.

Two questions that popped into my head toward the end of Thinking, Fast and Slow related to Kahneman's conceptualization of two different kinds of thought processes: System 1 which is intuitive, automatic, fast, instinctual, emotional, and System 2 which is deliberative, slow, rational. Often times, these "systems" or ways of thinking reflect conflicting desires: System 1 wants to eat a burger while System 2 wants a salad. System 1 wants to watch an action flick while System 2 wants a documentary. So how we define consumer desire depends on whether we appeal to System 1 or 2. Here are my questions: What would a world look like that was entirely geared toward System 1, without restrictions? Are we now living in that world?

The inconsistency between what we say we want and how we act under various choice conditions is not infinitely large. Depending on what you compare it with, it could be considered insignificant. It doesn't make much sense to throw out the entire system of valuation because it is flawed. Better to detect the flaws and correct for the flaws in a way brings about positive individual and collective outcomes. The first step is one many people haven't taken: seeing the inconsistencies between what we say we want and what we choose in different circumstances. Then, think about the world or the life that we want and think about how to design a choice environment to bring those about. 

The relationship between supply and demand work in a marketplace of culture objects (e.g., TV shows) has a level of complexity that is not accounted for in the rhetoric that surrounds improving TV technologies such as Netflix and YouTube, both how they are sold and how they are celebrated by the press and by consumers. The relationship isn't infinitely complex, and work like Thinking, Fast and Slow lays out some rules for how people behave in certain choice environments.

Remote Controls

This moment keeps nagging at me, demanding that I think about it, write about it. First, I must acknowledgement the ways in which metaphors, or the likening of one moment in history to the present moment, can hinder understanding. By cherry-picking the ways in which the two moments are alike based on our preconceived notions of the fundamental nature of the present moment while ignoring all the ways that the two moments are not alike (or the ways in which the present moment is similar to another moment in history), we don't move any closer to understanding our current moment. But here I use the past moment not as a means of comparison or metaphor, but as a way of identifying how certain trends in media use got started. 

I'm speaking of the invention and popularization of the television remote control. The remote, along with the increase in the number of channels, marked a crucial lowering of the barrier to toggling among choice. It was possible to browse entertainment options before, but not quite as easy, and that shift toward easy browsing marked a change from comparing several options to one another to what I call entertainment foraging. Our experiences of using media in an impulsive manner and the attendant feelings of guilt grow out of this moment. The internet and mobile devices have merely extended the logic of the remote control to more moments and areas in our lives. Even when we stay on a single website like Facebook or Buzzfeed, we are often hunting or foraging for some unknown thing. We tend to think of media use as content consumption or connection with an other, as individual experiences: skyping with a friend, watching a video, spending time on Facebook. But I'm interested in the moments in between, the time spent looking for something, the time spent choosing, the proliferation of what you might call "choice points". It's the glue that holds together the other moments, but it takes up a lot of time, perhaps as much time as the moments themselves.

When I started thinking about media choice, I thought that change from the traditional media choice environment to the new media choice environment was the change from deliberative choice (System 2, in Kahneman's terms) to impulsive choice (System 1). But eventually I came to believe that even if the options are few, when its a matter of how you spend your leisure time, the stakes are very low, and so you make a quick choice. There isn't much at stake, so why deliberate? Even when the choices were few, we probably still chose impulsively or ritualistically, without much careful consideration. So perhaps our media choices were always usually impulsive, but they were impulsive with many borders or restrictions, different borders and restrictions than the ones we have now. The options from which we chose leisure media experiences were limited by bandwidth and shelf space. The times at which we chose such experiences were limited by synced schedules and clear demarcations between work and leisure times and places. Without the borders, without the restrictions, the options have changed. When the options change (and this is highly counter-intuitive, but supported by a ton of empirical evidence), our choice patterns change. Increasingly, our impulsive choices, collectively or individually, feedback into the system that generates the option menus. Our options, and our selections, are dictated by the impulsive self with less interference from the outside world. This doesn't bode well for our long term self, our abilities to achieve long term goals.

What can we do about it? What are we doing about it? There are new technologies that form a middle layer between media applications that offer us options and our impulsive choosing selves. I call these software applications, like Freedom or Self Control, choice prostheses. Are they effective? That depends. In some ways, use of choice prostheses resembles dieting, and most diets do not work in the long term. In other ways, they resemble choice architecture or nudges, which are more effective in changing behavior in the long term. This is the next step in my research on media choice: to better understand how choice prostheses work and how they might best be used to change our choices for the better.