Cambridge Analytica has been under a lot of scrutiny over the past 4 months, ever since Channel 4 did an undercover investigation in its ties with Facebook, local governments and other institutions, which CA helped in dubious ways. This article is not about that, but instead about their research and segmentation technique, something they call: pshychographics. They use “data enhancement and audience segmentation techniques” providing “psychographic analysis” for a “deeper knowledge of the target audience” (source: Wikipedia).
Based on the Big 5 personality traits-model, they measure and predict behavior along 5 pillars, which they call the “OCEAN” model:
- Openness – Does he/she enjoy new experiences?
- Conscientiousness – Does he/she prefer plans and order?
- Extraversion – Does he/she like spending time with others?
- Agreeableness – Does he/she put other people’s needs before theirs?
- Neuroticism – Does he/she tend to worry a lot?
Also see this video in which these traits are explained in more detail:
They claim – and I would say rightfully so – that regardless of age, demographics, sex, financial background, etc., people can have the exact same OCEAN typification, as such creating customer segmentation based on psychological needs. This is interesting because for many years, demographics and ‘quantifiable’ information have been the main drivers for creating segmentations and customer profiles (“our customer is a ’40 year old woman with 2 kids, making an average of $25,000 per year”).
See the image above, for example. The other way around is also true: both of these people are roughly the same age, share the same gender, might have a similar salary, but have a completely different OCEAN profile: the left girl is more open en less conscientious, whereas the right girl shows opposite behavior. this, of course, has large implications for your service design.
Anticipating for changing customer behavior
By viewing the world in a different way and typifying people based on the OCEAN model (or any other psychological model, for that matter), we open up the door for creating services and experiences that directly tap into customer needs, not the customer ‘physical’ and measurable attributes.
The beauty of thinking in this way, is that you basically decouple your segmentation results from time and demographics. What I mean with that is: it’s in our nature to change, grow and transform. Transform in our opinion and view on the world, our likings and dislikes, our taste, our needs, etc. One’s financial preferences and needs can drastically change depending on our context and situation. As such, our needs transform or ‘move’ from one point on the segmentation canvas, to another.
Let’s look at a simple example:
At some point in my life, I might be more prone to find security and safety for my finances, whereas in other cases I might want to be more risky because I’m looking to buy a second car, or have a bit more to room to spend and invest, seeking growth in my financial situation (maybe I just got a new, better paying job). Yet, at another point in time, I might be seeking protection of what I have accumulated, for example to facilitate and help my family. These changes in mindset can happen due to a certain life event, which happen once every 5 or 10 years or so – such as getting married or buying a house – but might as well happen overnight (e.g. see the rise of crypto-currency).
Those changes are not bound to passing time, age or sex, but are bound to my psychological state and needs at a given point in time: needs are dynamic, not static. They change when we change, they adapt to our situation and help transform us and become a more complete person. Therefore, it’s dangerous to interlock them with static data as we might consider that as ‘the one truth’, while in reality, there are ‘many truths’ and situations, depending on the context of use.
“Therefore, it’s dangerous to interlock them with static data as we might consider that as ‘the one truth’, while in reality, there are ‘many truths’ and situations, depending on the context of use.”
What can we do with this?
So, what can we do this information? For example, you could follow these steps:
- Do your qualitative research, as always (interviews, focus groups, observations, whatever is most suitable)
- Find patterns in behavior, intrinsic tensions, underlying needs, and other interesting differences in your research results (opposing behavior often points in a direction that you’re onto something: find out what’s behind it!).
- Use the OCEAN model as a base for ‘search domains’ for those behaviors: find differences and juxtapositions in openness, agreeableness and neuroticism (for example)
- Summarize and plot these findings in a way your prefer (e.g. in a 2-axes matrix as depicted above)
- Find 4-8 psychological distinct segments that are truly different (in needs, behavior, goals, etc.)
- Start your design project with these psychological segments in mind, without making personas as you’re used to!
(By making traditional personas, you run the risk of characterizing and visualizing the target group and falling back into ‘old patterns’)
Good luck! This approach has increased my focus on the problem and design challenge, anticipating changing customer behavior and the ‘fluidity’ of our everyday lives, frankly.