13 August 2013

Take Nudging Seriously! Think Twice Before Starting to Nudge

The past five years have seen a boom in media coverage and application in practice of insights from behavioral sciences, particularly from behavioral economics. Books such as Nudge: Improving Decisions About Health, Wealth and Happiness by Richard H Thaler and Cass R Sunstein, Predictably Irrational: The Hidden Forces that Shape Our Decisions by Dan Ariely and Thinking, Fast and Slow by Daniel Kahneman have contributed significantly to the popularization of behavioral sciences and their application in practice.

The UK government has established a “Nudge unit” which has the task of applying insights from behavioral sciences into public policies. Apparently the US government is on the path of creating a similar unit.

Businesses, particularly marketing departments, are stepping up the application of this knowledge with the aim of increasing sales and profits. NGOs use the findings in behavioral sciences to increase the amounts collected in fundraising endeavors.

The truth is that applying knowledge from behavioral sciences or nudging is highly attractive and appealing. One reason for this is that nudging promises large outcomes with small inputs (changes, investments etc.). The second and most important reason is that Nudging is simply cool.

Let’s focus a bit on what makes nudging so cool. My best (educated) guess is that coolness is the result of the surprise we experience when learning about findings in behavioral sciences. When we learn that simply changing the default option in a form leads to huge differences in behaviors such as enrollment in different programs we are amazed. This is because this finding goes against the popular intuitions we hold. When we learn that diffusing a certain smell in a shop leads to increase in sales and in the customers’ evaluations of the service and the merchandise we experience the same high level of surprise; again because we don’t consider that what we smell should make any difference in our decisions and evaluations.

Knowledge from behavioral sciences is appealing and surprising because we couldn’t think it was possible that a small change in the context to make such a big difference in the behavior (outcome).

Knowledge from behavioral and decision sciences is so cool that we all want to use it, we want to be in the shoes of those researchers and after all, who wouldn’t want to increase sales, donations or make their job easier by applying some cool tricks?

Nudging Seems Easy

Because small changes in the environment can lead to huge changes in the behavior, nudging seams easy. Although you couldn’t think that changing the structure of a form can have huge impact in the number of people enrolling in a program, now that you know, it makes all the sense in the world. Moreover, it seems so easy to do… after all, how hard can it be to change a form?

It goes similar with using scent in a shop. Once you know that diffusing the scent of chocolate in a book store leads to an increase in sales of romantic books, it seems so easy to just place some chocolates near the counter or buy some aerosols with chocolate smell and place them around your shop.

After reading one of the books I mentioned earlier or after viewing some interesting videos with behavioral economics specialists we are aware of many “tricks” or tools that can be used and there is a huge temptation to start applying them. However, the question is which one? Should we simply pick up the one we found the most interesting? Should we just do anything that comes to mind?

The main reason why Nudging seems easy is because individuals are exposed to information on what works. People who want to start nudging often discard the context in which it worked – such as book store – and most of the times the wannabe nudger knows close to zero on the psychological mechanisms that underlie the alteration of behavior. Simply put, one knows that something works, but doesn’t really understand why and often discards the differences between the contexts in which the finding was made and the one in which (s)he has to apply the finding.

Nudging seems easy ONLY POST FACTUM  

Once you know it works it looks easy, but it is damn hard to think of what would work before the fact and before testing. Most researchers publish scientific papers on what works. Books and mass-media give attention to what works. But behind all these things that work, there are many more things that didn’t work. A researcher can’t publish in a respectable scientific journal a paper which includes five failed experiments, even if she spent half a year on a project that proved to give insignificant results.

Remember that the surprise we experience when hearing about the effectiveness of behavioral interventions (nudges) comes from the fact that we couldn’t think it was possible! Now that you want to design such nudges, you have to think of something that previously was unthinkable. That’s not so easy, is it?

If you want, things are similar with the difference between explaining and predicting. Explaining why something happened is infinitely easier than predicting the same event. Think for a few seconds about the recession and banking crisis of 2008-2009. There were countless people who showed up on TV or in the newspapers explaining in detail how and why it happened. At the same time, very few of these people were able to predict the same events.

Let me give you an example from my own backyard. In October I’ll give a training session on choice architecture and during the first day the ten participants will learn about the established psychological effects on choice such as the compromise effect, loss aversion, the status-quo bias etc. They will also see illustrations of these effects put in practice. In the second day, during the first hour the participants will see three case studies on how these effects are used. Probably the most appealing case study will be the one on how Linked In structured its offer for paid accounts.

I’m sure that these case studies will be very interesting and things will make all the sense in the world. However, after this first hour, participants will get to work on, first, analyzing a choice set used by an on-line service provider. This I expect to go quite smooth since they will only have to recognize the effects described in the first day and presented at work in the first hour… But the fun begins only when the participants will have to construct from scratch choice environments aimed at increasing donations for a charity, increasing sales of high margin products for a company that sells through a catalog and to help a museum increase its ticket sales for tours outside the main exhibition.

Applying choice architecture in an effective manner is far more difficult than seeing it already applied. It will be a bit tough, but so is the work of creating behavioral interventions (nudges) from scratch.

How to make the creation of nudges easier?

I have to be honest with you and say that there aren’t five easy steps to success. Creating behavioral interventions or nudges is hard and requires a lot of knowledge, work and testing.  

What will help for sure is to understand how and why these nudges work. Behind the examples you see published in academic journals or in mass media there are some rules. Learn and get an in depth understanding of these rules.  My advice is to ask at least twice: why?

If you know that playing classical music in a restaurant leads to an increase in sales in value, but not in volume – people go more for the premium products, ask WHY?

The answer is that mental constructs associated with high-class become more salient. After getting this answer, you should ask again WHY? And then you should get the answer that people associate classical music with high-class and that the music acts as prime (priming tool).

This is only one illustration of getting an in depth understanding on how nudges work.

Don’t go head on

Even if nudging seems easy and effective, you should think hard and look for the right intervention to use. The kind of in depth knowledge I’ve illustrated above is needed to create effective interventions. If you go head on and apply something that you find cool, there is a big chance that it will not work. 

Behavioral interventions are science, not magic!

After you have designed one or two carefully selected nudges, you should test if they work and my advice is to run experiments. This is not easy and requires accuracy and patience. Don’t make comparisons between before and after. There can be many reasons for why before was different than after and some of them are unrelated to the nudge.

Should you not apply knowledge from behavioral sciences?

I believe that businesses, governments and NGOs should apply knowledge from behavioral sciences. However, my main argument is that in order to create effective interventions one needs to work professionally. There is a high risk of “nudge” becoming just another buzz word used by people who don’t really understand its meaning.

Work with real professionals who understand in depth how behavioral interventions work. Stay away from people who promise fast miracles. Work with specialists who can give you pertinent answers to at least two “Why” questions. Work with those who do / advise experimentation in the sense of running controlled randomized experiments to measure the effectiveness of the intervention.

Take Nudging Seriously!

No comments: