Persuasion Science on the Web

I recently made the time to listen to the NPR podcast entitled The Science of Getting A 'Yes' that interviews Robert Cialdini about his new book Yes! 50 Scientifically Proven Ways to Be Persuasive (2008).

The subject is very compelling, especially in the context of controlled experiments on the web as a motivator for testing assumptions about customers. I found the anecdotes from related scientific studies very interesting. I captured some notes:

• Consider communicating what might be lost by not adopting the product as opposed to what might be gained (prospect theory). For example, the potential harm of not breast feeding rather than the benefits a breastfed child may receive (from a caller to the show).
• Potential danger of negative social proof. For example, signs to not steal petrified wood from a forest increased theft, announcements not to cheat on taxes increases cheating (perceived that peers are acting in a particular way promoting that behavior). An interesting counter example was that of the reuse of hotel towels, where signs that promoted the number of people that reused towels were far more effective at encouraging that behaviour than generic slogans about helping the environment.
  
• An unexpected and personalised gift. For example, a waiter received more tips by giving one mint per diner with the cheque. More tips again with two mints per diner, and more tips again with one mint and a second round of mints provided 'just for those customers'.
• When asking for something use 'because' an suggest a reasonable reason for whatever is being requested. People generally want to reciprocate, they just need a reason.
  

The book has an official website, that provides podcasts and extracts. I stumbled across a more complete list of the 50 take-away points of the book, which although are bland, are still provocative for web-based experimentation. There is little doubt that this is a rich source for marketing optimizations, although the shear number of points makes it hard to internalize. I think a good strategy is to take a given product and go through the list assessing the product in the context of each point (where appropriate).

Cialdini also has another book entitled Influence Science and Practice (2008, 5th edition) that has a similar theme. The wikipedia page for the book seems to summarise the core message to a series of effects referred to as the 'weapons of influence', which are: reciprocation, commitment and consistency, social proof, liking, authority, and scarcity.

This form of marketing (exploiting sociology research) does feel somewhat insidious. For example the excerpt provided with the NPR interview provides an example of the success of a telemarketing campaign that was attributed to changing "Operators are waiting, please call now" to "If operators are busy, please call again", a clever use of social proof. I guess persuasion by definition is a manipulation of others, and the ethical use (let alone optimization) of these tools appears subtle. It is just word play on the web after all, although I suspect such approaches could be (are?) a gateway to more devious social engineering.