Some time ago Maurits was asked to write something for the popular blog on interaction design; Johnny Holland. And I didn’t want to keep this from you, especially not if you are from the UX field.
Researchers and practitioners in fields ranging from marketing, to psychology, to medicine have been interested in how people persuade or influence other people.
Recently, a number of the persuasion principles (the ways in which people influence other people) that where used in face-to-face influence attempts are starting to make their way into the design of online services. Even though most of these persuasion principles are effective on average there are clear indications that we should personalize our use of these persuasion principles. Based on my PhD work I will describe why these strategies should be adapted to individuals using persuasion profiles, and I will try to address some of the issues that are relevant to UX designers.
So, what are those social influence strategies?
Influence strategies are basically messages that are used to increase compliance, which are peripheral to the actual request. Thus, if you try getting someone to stop smoking, you can add to that request that “nobody really smokes anymore”. This latter message does not directly relate to the actual request, but does make it more likely for people to comply with the request. Cialdini identifies six principles of persuasion:
- Reciprocity – people feel obligated to return a favor, thus when a persuasive request is made by a person the receiver is in debt to, the receiver is more inclined to adhere to the request R. Cialdini (2004);
- Scarcity – when something is scarce, people will value it more. Announcing that a product or service is scarce will favor the evaluation and increase the chance of purchase;
- Authority - when a request or statement is made by a legitimate authority, people are more inclined to comply or find the information credible;
- Commitment and consistency – people do as they said they would. People try to be consistent with previous or reported behavior, resolving cognitive dissonance by changing their attitudes or behaviors to achieve consistency. If a persuasive request aligns with previous behavior people are more inclined to comply;
- Consensus – people do as other people do. When a persuasive request is made people are more inclined to comply when they are aware that others have complied as well;
- Liking - we say “yes” to people we like. When a request is made by someone we like, we are more inclined to act accordingly.
Why should we adapt to individual users?
If you think about it, each of the principles of persuasion that is mentioned above is used in one way or another on a lot of e-commerce websites. However, these principles are also used to help people stop smoking, or help people lose weight. In my PhD I have worked on these latter types of applications, and there the motivation differs from the one in marketing: We do not just want to increase compliance – and thus for example decrease smoking – on average, but rather we want to motivate individual people. We (researchers at the TU Eindhoven and at Stanford University) setup a number of studies to examine whether or not people respond consistently and similarly to the principles presented above. Overall, we found that while these principles work well on average, they often fail for individuals. Some people never listen to authorities, or try to do exactly the opposite of what others are doing. However, we also found that people are generally consistent: If you do not respond to authority arguments now, you are likely to also not respond to them tomorrow. Next, we tested whether if it is possible to use all the principles at the same time (so: “Please stop smoking! Nobody smokes anymore and your general practitioner recommends you to stop. Etc. etc.”).
Comparison of heterogeneity in the effects of influence strategies with the average effects of those strategies. The solid black vertical lines are the estimated average effects of each strategy, as compared with the control message. The black curves are the estimated normal distribution of strategy effects for the population, while the gray curves are the density of the estimates of the strategy effects for this sample.
We found is that that choosing a specific principle is actually more effective than showing all of the principles at once.
What are “Persuasion Profiles”?
These two results combined (people are different but consistent, and combining all the principles does not work) lead us to start playing with the idea of persuasion profiles: Collections of the estimates of the success of different strategies for individuals. Thus, your persuasion profile could say that the Authority principle works fairly well for you, while using the Scarcity principle pretty much always fails. We can build your persuasion profile by observing your responses to the principles mentioned above.
Estimated influence strategy effects for each participant in Model C, as compared to the control messages. Note that for some individuals, the estimated effect of one strategy is negative, while the estimated effect of another is positive.
These persuasion profiles can then off-course be used to select which principle to present next: So for you, given your profile, we would say that your general practitioner recommends you to stop smoking (which probably is true) while for some of your friends we use a different principle.
Does it Actually Work?
The idea of persuasion profiling is appealing, although also scary when used with the wrong intends. But, we still needed to show that it actually worked. To do so we setup a series of evaluations in which we use persuasion principles to influence people, measured their response, and then created a profile. The next time we tried to influence the same person, we selected a principle based on the profile, measured the response, and then updated the profile. We tried this in several settings: We put a big screen in the hallway of an office building and put up messages to encourage people to take the stairs. The messages used different persuasion principles. Due to the Bluetooth-key of their mobile phones we were able to track individuals, measure whether or not they took the stairs, and present them with a personalized message. What we found that more people took the stairs when we personalized the messages using the persuasion profile, than when we just randomly picked one of the messages. We showed in a similar way that persuasion profiles can increase the responses to emails, and also can drive sales on an e-commerce website.
The Bottom Line
I hope to have introduced the idea of Persuasion Profiles, and motivated why they can be effective in influencing people. However, just like other personalization efforts persuasion profiles raise both ethical and practical issues. For UX designers of applications that intend to influence the behavior of their users it means creating different types of content for each of the persuasion strategies. It also means that each user will have a different experience, and perhaps a totally different motivation to use your service. I believe the long-term consequences of these types of personalization are not well understood both by designers as well as by researchers.
However, I do believe persuasion profiling is coming up, and it can be for better or for worse. It can teach us about our own decision-making by showing us our profile – so then you know you always listen to authority arguments.
But, it can also potentially sway us into doing things we do not necessarily want to do…