Psychometric personality-based factors in (OCEAN) targeting
What is targeting by psychometric factors?
Data plays a significant role in the targeting of individuals and segments online with promotional messaging. The advertising industry has seen a huge growth in data usage over the last decade. I have been vocal about the issue that data has now become an over-traded commodity, thus making it hard for marketers to make informed decisions or indeed, decisions that lead to incremental value.
Marketers face a sometimes impossible task of breaking down the data barrier to target customers.
A marketeer will start with a customer hypothesis built from data on its known customers or perceived customers.
Age | Gender | Location | Affinities - Interests.
The above is a great start but because data has become such an immense marketplace, effectively targeting based on the attributes of above the marketeer is left with a campaign that no longer delivers significant ROI, value because the net is cast so wide and offers little in the way of distinctive differentiation.
Behaviour, motivations and personality.
Now let's look at motivational and behavioural segmentation, introducing one of the big 5 personality matrix assessments, OCEAN or psychometric indicators. OCEAN happens to be just one of the known behaviour classifications, possibly one of the most well known what does OCEAN stand for O (Openness) C (Conscientiousness) E (Extraversion) A (Agreeableness) N (Neuroticism).
Openness to experience (inventive/curious vs. consistent/cautious)
Conscientiousness (efficient/organised vs. extravagant/careless)
Extraversion (outgoing/energetic vs. solitary/reserved)
Agreeableness (friendly/compassionate vs. challenging/callous)
Neuroticism (sensitive/nervous vs. resilient/confident)
More interestingly if we look at the IBM Watson wheel of personality profiling, the image above we is a sunburst of my personality snapshot from IBM in relation to my Twitter profile https://personality-insights-demo.ng.bluemix.net/sunburst. I am not a huge twitter user but some of this is amazingly accurate, as with always some of it you wish wasn't but when you dig deep you know it is. You can also see that they have swapped out Neuroticism for Emotional Range, possibly as emotional range sounds better in contrast.
Creating the nudge experience.
So why is personality, nudge or even psychometric targeting so exciting, it is because it adds the dimension that should in theory combine creative and data together. I have worked on multiple projects across retail, travel and finance and can confirm that when you combine and layer data correctly, use of 1st Party CRM data, combined with the addition of consensual opt-in collected personality metrics, engagement and performance see a direct impact, over seasonal campaigns, this can equate to 150% YoY growth and product launch luxury retail brand 350% increase in engagement.
So how does this work, what does this look like?
The beauty with behavioural science and data-driven creative lies in the process of getting to know your customers, the motivations, the way they want to be communicated with, the style, the imagery, the tone of voice.
Once you have understood what your ideal customer looks like what the hypothesis is, broad demographics and affinities you can look to build out a customer survey alongside a wider survey.
That survey research tool enables you to start to really identify the nuances and build a more in-depth picture of both existing customers but also potential new segments of customers currently not catered for within the proposition.
Conducting research like above is a little more in-depth than a simple online survey containing 5-10 questions, we are looking at discerning more information to enrich the customer experience, for reference is a little more in-depth than below.
For example, part of the survey includes questions like:
Q. What does luxury “insert product” mean to you? Rank these images from 1/top, sums it up the best, to 9/bottom, sums it up the worst. Just go with your gut instinct.
Images presented are depicting - Sensual, Glamour, Status, Love, Indulgence, Fresh, Fashion, Luxury, Wealth.
So why is the above important? This is where the data and behaviour science team starts to infer insight into the imagery that will best engage the audience, more importantly, segment imagery and creative against different audience sets, leaning towards the nudge approach for different segments.
We can look at proposition, brand, mission statement and tone to see if it resonates with survey sample we can look at propensity to buy metrics for retail eg.
Q. Apart from the “product”, what would encourage you to buy a “generic product”? Pick your top two reasons.
Options could include: Low price, High price, Familiar brand, Recommended by someone I know, Good reviews, Appealing packaging, Appealing bottle, Appealing advertising/branding, Handcrafted/Artisanal, Environmentally-friendly, Unique or different, Locally-sourced, Easy to find/buy
Notice above we are also starting to discern perceptions and motivational and psychological reasons to buy.
We can also identify the audiences that will re-engage, purchase more frequently with the LTV (Life Time Value) and can focus creative and targeting in new acquisition strategies.
There are obviously a raft more questions including disqualifies and flow routes but ultimately the goal here is to really get to know your consumer and remove the bias of large customer segmentation that derails value and engagement.
So let’s look at an example of how this may play out.
We have a luxury brand specialising in the accessories sector, their product is expensive, their product has a distinct style, following and heritage - sales are good but have plateaued over the last 24 months, the brief is to identify growth opportunities, new segments, new customers and test narrative hypotheses.
Process:
Design survey methodology, questions and qualify existing customer hypothesis.
Survey UK Quantitative sample 5,000 completes (you may have to reach out 10,000 but only get 5,000 completes, but bear in mind the data of the other respondents is not useless it actually can provide great insight into who is not in your customer sphere.
3. Once the survey is completed we look to run data/behavioural analysis through K-Means or indeed Gaussian Mixture models to cluster the segments into workable groups.
big-data - Data Science | Behavioural Science
4. Ascertain the creative differentiation and huddle data science, creative and planning together to create the narrative, tone, design and plan activation across relevant channels like Social Media, PPC, Content Marketing etc..
5. Produce data-driven creative inline with clustered segmentation and review your 4-5 clusters of audience. I have always been a fan of naming conversion like:
Behavioural Segmentation
Stylish Susan | Perfect Peter | Leisure Louise | Negative Nick
The above gives you an instant impression of who this audience is but the exciting part is yet to come when you drill into the detail further, let's take a look at Stylish Susan.
Stylish Susan
Represents 31% of of the segment
54% More likely to buy multiple times a year as a gift
29% More likely to buy multiple times a year for themselves
Engages rich vibrant tones and abstract imagery
Love, Sex and Status drive this consumer
2nd Wealthiest segment, 71% Female, 29% Male age range 25-54
47% of this segment is Married,
Highly Extroverted, Open and Contentious.
Social Media Usage
Facebook 84.3%
Youtube 77.7%
Instagram 47.5%
Twitter 46.9%
Online Behaviour
Social Media, streaming videos, reading news, shopping, & research
Adverts Clicked
Entertainment, news, technology, fashion, food
Messaging - Tone
Clear, clean and concise, luxury, vibrant, aspirational - slightly out of reach but attainable. Platforms of focus programmatic display, Facebook and youtube.
The above is a snapshot, and the analysis gained from these types of exercises can deliver amazing insights that can truly move marketing campaigns to the next level, building cohort audiences from opt-in data, custom audiences across multiple platforms, advertiser|publisher data matching (merging to 1st party data sets)
Sound interesting drop me a line to find out more, Oliver
