Use Your Data to Convert On the Fly: Persuasion Profiling Q& A with Science Rockstars

Posted by | Comments


During our last webinar, Science Rockstars Chief Science Officer Maurits Kaptein introduced an interesting technology – the PersuasionAPI – that promises to increase the likelihood of visitor conversion by serving them web pages adapted to their specific “persuasion profile.” Persuasion profiling is an intriguing concept: it presents a new level in personalization which takes and deliberately uses Big Data to study how visitors respond to different persuasion strategies and adjust content accordingly. 

Read the abridged transcript of the webinar Q&A to learn about the theory and principles that guide Maurits and his colleagues’ work at Science Rockstars:

1. Why do Robert Cialdini’s persuasion principles work, and why is using all of them at the same time less effective?

One of the prevailing theories about why these principles work is that they are mental shortcuts that, we've learned, often lead to the right decision. For concensus, for example, we've learned that it saves us the time and the trouble when we follow what other people do. 

Shorcuts are not meant to be consciously processed. If you put a lot of shortcuts, people may start processing them more actively, which may not always be beneficial for marketers.

2. Any suggested works on persuasion that can be used as reference? 

Robert Cialdini's “Influence: The Psychology of Persuasion” (Maurits’s collaboration with Dan Eackles- a  collection of academic papers on persuasion profiling) 

3.  Can you dynamically change something based on the session (the pages the visitors see)? And how do you decide which page to show? 

For most of  Science Rockstars’ clients,  when they run their page, they make it go to Science Rockstars’ services, then Science Rockstars sends back advice of which version to show based on the customer’s persuasion profile. 

So you need to have different versions of pages with variations for each of the principles available. Mostly on e-commerce websites, we're not talking about a full redesign of pages. We're talking about buttons that are placed on top of products, little banners, a limited set of icon images; Science Rockstars determines  which of the icon images gets placed where. 

4. How do personas, which are typically used by ad agencies, relate to persuasion profiles?

Use of traditional personas is good for visualizing your target audience. They’re a great way to help people create content-- to get everybody in the right focus and think along the same way, especially in a way that's tailored toward a specific segment of your customers. In persuasion profiling, we also generate content.  So in a way, there's an overlap between the use of personas for content creation, and the idea of using persuasion principles is one of the driving force.

Science Rockstars is doing things slightly differently. A customer is not categorized just based on his/her previous purchase. Science Rockstars is learning continually because of the recognition that estimates are uncertain. Personas are only used to create content, but what the customer sees is based on how he/she responds, and that continually changes. 

5.  How effective is PersuasionAPI? What typical results do you see with it? 

Persuasion principles generally have a positive effect. Improvement will depend on what you're already doing now and the frequency of customer visits to the site. 

6.  How long does it take to zero in on what makes a person tick? 

After 25-30 interactions, we start to get fairly large groups of people whose estimates are not changing a lot anymore. So we're kind certain that we're addressing them in the right way. 

However, we don't need 30 interactions before we make a decision. The second interaction will be slightly better than the first one because the estimates will have changed based on a click or a purchase. At a certain point, you get a certain profile, but that's not the aim of what we're doing. The aim is not to perfect the profile. The aim is, given the history, make things more effective. 

7. Are the principles different for B2B and B2C? Are there differences in terms of industry or specific application?  Would it be more effective in certain circumstances than others? 

We don't have very good cases to compare B2B to B2C; but in whichever case, it's still humans making the decisions , so there should not be too big a difference . 

In terms of how this relates to certain products or categories, for some clients certain strategies work really well. On average, one strategy is by far more effective. This is because the group of people a company attracts is dependent on the company's business model (e.g. companies, like Groupon, who offer last-minute deals will attract people who respond to scarcity). 

8. This is a learning algorithm so you can't really do this with one-time visitors; the more people navigate on the site or interact with the CTA, the better, right?

With one time visitors, you just have to do a fancy A/B test. 

Watch the full webinar to listen to Maurits answer more questions about PersuasionAPI.