Nudges are Cheap in the Same Way Medicines are Cheap
In the majority of presentations on applied behavioral science, it is strongly emphasized the fact that Nudges are low-cost (cheap) interventions/ changes that have a disproportionally large impact on the desired outcome.
I believe that this is, mostly, an exaggerated claim. Nudges are cheap in the same way medicines are cheap. Manufacturing one pill doesn’t cost much, and the pill, usually, is highly effective in achieving the desired outcome.
In the case of Nudges, behavioral interventions such as changing the default option, adding social comparisons, and rewording of a message are or, at least, seem cheap. But, there are at least three big categories of not-so-obvious barriers to adopting behavioral interventions.
First, there is the issue of research costs. In the case of medicines, the problem is what chemicals to put in a pill. Manufacturing a pill is cheap, but the research needed to know which chemicals solve a certain problem is expensive, very expensive. It goes similar with behavioral interventions. Rewording a message or introducing social proof are, usually, cheap, but the question of which intervention will be effective can only be answered by applying thorough research. And that doesn’t come cheap.
Feasibility of applying behavioral interventions
Second, there is the issue of feasibility of applying behavioral interventions. Changing the default option, adding an element of social proof on a package or rewording a message in a letter seem like they are easy to implement. However, more often than we like to admit, these are very difficult to implement in real life. Consider the following examples.
The by-default enrollment of employees in 401(k) plans was very controversial and the matter was settled by a decision of the US Supreme Court of Justice. That can’t be easy.
Changing a message on a packaging of consumer goods implies a complex process of having new designs, making changes on the assembly line and dealing with the hurdles of the necessary logistics to make the changes.
Rewording a message in letters from public authorities requires to go through several levels of bureaucracy. After all, what and how a public authority communicates is, rightful so, heavily regulated.
The willingness of decision-makers
Third, there is the issue of decision-makers’ willingness to accept the deployment of behavioral interventions (Nudges). I’m not talking about the unhealthy reluctance that some people exhibit instinctually. Rather, I’m talking about the healthy caution-driven reasoning.
Does it make economic (business) sense?
is a perfectly legitimate question.
Promoters of behavioral interventions emphasize the low-cost – high-impact nature of Nudges, but, as I outlined earlier, there are legitimate questions about this nature. In practice, Nudges aren’t as cheap, both in terms of research costs and hassles of implementation, as they first seem to be.
One could easily argue that even these overall costs are offset by the large impact (benefit) that behavioral interventions bring.
I believe that the large benefit part, too, needs to be questioned. Take the example of the already famous study done by the Behavioral Insights Team in the UK on which message leads to the highest increase in enrollment in the organ donation registry (i.e. registering as an organ donor).
The most effective message lead to an increase in enrollment, compared to the control / status quo, of 0.9 percentage points. Enrollment rates went from 2.3% to 3.2%. It’s not bad, but it isn’t astonishing either.
Moreover, there seems to be no statistically significant difference between six out of the seven messages used. From a skeptical point of view, the main learning from this study is that doing anything different than what was done at the time, albeit informed by behavioral science, can’t do any worse than the status quo. (I know that one message looks like it performed worse than the control, but that is only significant at the 95% confidence interval, which is irrelevant when the sample size is larger than one million).
Can similar results be obtained by doing something simpler or more established?
is another legitimate question that decision-makers might ask.
For example, if the goal is to increase enrollment rate by one percentage point, then other more established solutions such as advertising, incentives etc. might do the job.
The more established solutions have the big advantage that they are easier to be adopted / to be accepted. In organizations there isn’t one single decision-maker, thus the proposal to deploy a behavioral interventions project needs to be accepted by several people. More established solutions to the issue in question will incur less resistance than the deployment of a Nudge project.
When a Nudge expert says that she can do research in the hope that a low cost intervention can be created, she needs to consider the overall costs, including the hassles of implementation, and the willingness of the client to accept it. Moreover, she needs to know that her proposal competes with other measures that decision-makers are more comfortable adopting.
What can be done about this not-so-rosy situation?
I don’t have a definite answer, but I guess that medicine (pharmacology) can be an inspiration. When a new drug is developed, the costs and hassle are huge. However, subsequent to the drug being released on the market, doctors who prescribe it don’t need to do thorough randomized control trials to assess its effectiveness. They need to see if the patient’s condition improves or not.
I guess that, at some point in the future, there will be a well-researched set of best practices in applied behavioral science that will be used with reduced (minimal?) monitoring on the effectiveness.
Moreover, it is/ will be important to incorporate these best practices from the early stages of any project – product or service design, marketing communication, policy etc., as this will reduce (eliminate?) the hassles of changing what already exists.