I mean, today you probably will have some opportunities to take stuff and put it in your backpack, and nobody would know. You could take some silverw… - Dan Ariely

The basic principle with self-herding is that people don't know what they want very often. Figuring out our preferences is in fact very hard. If I ask you, "What is the best place to eat?" you have many choices. These are hard questions with lots of possible answers. They way we often answer these is by asking, "What have I done before? What I did before must have been a good decision, after all I wouldn't have made it if it wasn't a great decision, let me repeat it." We can consult our preferences or we can consult our memory. It turns out it's often easier to consult our memory.<p>What does it mean to marketers? To the extent that you can get people to behave one way there's a good likelihood that people will keep on doing that later. If you can remind people how they behaved, there's a good chance that they will keep on behaving that way. So the logic suggests that the efforts of marketing should be particularly concentrated when people make first decisions about the product, like people in college who are making first decisions for themselves. It could include product introductions. It could also include things that happen when there's a shift in the economy. The recession caused many people to reconsider their habits.

With technological changes, there have also been changes to irrationality. We have been bombarded with more information, so making decisions is tougher than ever. Not only that, but we have a strong chance of regretting our decisions. Before the world was digital, you could make a decision, and you would probably never be able to see that your decision was not a great one. Now, if you buy something online, the day after it’s a click away to see that the prices have changed and regret can set in. So there’s a combination of information overload and a high potential for regret.

Imagine I give you a choice: Do you want to go for a weekend to Rome, all expenses paid -- hotel, transportation, food, a continental breakfast, everything -- or a weekend in Paris? Now, weekend in Paris, weekend in Rome -- these are different things. They have different food, different culture, different art. Imagine I added a choice to the set that nobody wanted. Imagine I said, "A weekend in Rome, a weekend in Paris, or having your car stolen?" It's a funny idea, because why would having your car stolen, in this set, influence anything? But what if the option to have your car stolen was not exactly like this? What if it was a trip to Rome, all expenses paid, transportation, breakfast, but it doesn't include coffee in the morning? If you want coffee, you have to pay for it yourself, it's two euros 50.<p>Now in some ways, given that you can have Rome with coffee, why would you possibly want Rome without coffee? It's like having your car stolen. It's an inferior option. But guess what happened? The moment you add Rome without coffee, Rome with coffee becomes more popular, and people choose it. The fact that you have Rome without coffee makes Rome with coffee look superior, and not just to Rome without coffee -- even superior to Paris.