Jason Duke, Founder, Kronaxis
Tag: Insights
Demographics explain roughly 10 to 15 percent of the variance in consumer purchase behaviour. This is not a controversial claim. It is the consistent finding across decades of marketing science research, from Ehrenberg's repeat purchase analysis through to the IPA's effectiveness database. Age, gender, income, and location tell you something about what people buy. They tell you almost nothing about why, and even less about when.
The industry knows this. Every brand manager has encountered the limits of demographic segmentation. Yet most research budgets are still structured around demographic panels, demographic quotas, and demographic reporting. The tools enforce the model, and the model stopped being adequate somewhere around 2005.
The Prediction Hierarchy
There are three layers of consumer behaviour prediction, each more powerful than the last.
Demographics capture the outer shell: who the person is on paper. Two 35 year old men in Manchester with £50,000 household income. Same segment in every demographic model ever built. Same targeting criteria in every media plan.
But one of them is high Impulsivity, low Discipline. He sees a new subscription product, reads the headline, and clicks "subscribe" before reaching the feature list. The other is low Impulsivity, high Discipline. He opens three tabs, reads the comparison page, checks Trustpilot, calculates the annual cost versus the monthly cost, and decides to wait until next month. Same demographics. Opposite behaviour. Demographics cannot distinguish them because demographics do not measure the psychological dimensions that drive the decision.
Psychographics (lifestyle segments, attitudes, values) improve prediction to roughly 30 to 40 percent of variance. They capture motivation and context better than demographics alone. But most psychographic models are categorical: you are either a "conscious consumer" or you are not. The real world does not work in categories. People sit on continua, and where they sit on one dimension interacts with where they sit on every other.
Personality conditioned models with economic and emotional context push prediction to 50 to 60 percent of variance. Continuous personality dimensions, combined with the persona's current financial state and emotional context, predict specific behaviours: time to purchase, comparison shopping depth, sensitivity to social proof, engagement with technical detail, likelihood of impulse upgrade, response to scarcity framing.
What Personality Adds
DYNAMICS-8 measures eight continuous dimensions, each scored from 0 to 1. The dimensions were not selected because they are theoretically interesting. They were selected because each one predicts observable consumer behaviour.
Impulsivity predicts time to decision. High Impulsivity personas convert faster on limited time offers, respond more strongly to scarcity framing, and are more likely to upgrade at checkout. Low Impulsivity personas ignore urgency entirely and may actually convert less when pressured.
Discipline predicts comparison shopping depth. High Discipline personas research alternatives methodically, calculate cost per unit or cost per use, and respond to clear value quantification. Low Discipline personas make gut decisions and rarely check competitor pricing.
Mercuriality predicts emotional response to price changes, brand messaging, and service failures. High Mercuriality personas experience genuine anxiety at price increases even when they intend to stay. They are disproportionately likely to leave negative reviews and share complaints socially. Low Mercuriality personas treat price as information and process service failures without emotional escalation.
Yielding predicts response to social proof. High Yielding personas are heavily influenced by "most popular" badges, review counts, and endorsements. Low Yielding personas distrust consensus signals and may actively avoid the popular option.
Acuity predicts engagement with technical detail. High Acuity personas read specification pages, API documentation, and comparison matrices. Low Acuity personas skim headlines and respond to visual hierarchy. The same product page converts differently depending on which dimension is dominant in your audience.
Novelty, Candour, and Sociability each predict their own behavioural clusters. Together, the eight dimensions create a fingerprint that determines not just what a persona chooses, but how they arrive at that choice.
The Economic Context Layer
Personality alone is not enough. A high Impulsivity persona with £200 in the bank behaves differently from one with £20,000. The same psychological drive (respond quickly to reward cues) expresses differently under financial constraint (suppress the impulse because the consequence is real) versus financial comfort (act on the impulse because the downside is trivial).
DYNAMICS-8 treats economic state as a first class input. Income, savings buffer, financial anxiety, debt load, and risk tolerance all modulate how personality expresses in purchase decisions. A high Discipline persona with high financial anxiety becomes hyper rational about every purchase. The same Discipline score with low financial anxiety produces a relaxed, methodical buyer who compares but does not agonise.
This interaction between personality and economics is where most psychographic models break down. They assign a label ("price sensitive") without distinguishing whether that sensitivity comes from psychological disposition or economic circumstance. The two require completely different commercial responses. A dispositionally frugal customer needs value framing. An economically constrained customer needs payment flexibility.
The Emotional Context Layer
A persona who just received good news responds differently from one who is anxious about their job. Emotional state modulates how personality expresses. The same personality under different emotional conditions produces different behaviour.
High Mercuriality personas amplify emotional context. Good news makes them more impulsive than their baseline. Bad news makes them more risk averse than their Discipline score alone would predict. Low Mercuriality personas show minimal emotional modulation: their behaviour is remarkably stable across emotional states.
This matters commercially because your customers are not in a laboratory when they encounter your product. They are in a mood. That mood interacts with their personality and their financial state to produce the decision you observe. Modelling behaviour without emotional context is modelling a version of your customer that does not exist in the real world.
The Practical Difference
Consider subscription churn prediction. A demographic model says "users in the 18 to 24 bracket churn at 8% monthly." That is descriptive. You cannot act on it because the only intervention it suggests is "acquire fewer 18 to 24 year olds," which is rarely a viable strategy.
A personality conditioned model says "high Novelty, low Discipline personas churn when the product stops feeling new, regardless of age. The median time to churn is 4.2 months. High Discipline, low Novelty personas churn only when a competitor offers measurably better value, and the median time to churn is 14 months."
One insight describes a population statistic. The other tells you exactly what to build. For the first group: introduce feature releases, visual refreshes, and engagement surprises on a 90 day cycle. For the second group: publish transparent pricing comparisons and quantified value metrics. Same product. Different retention strategy. Different personality driver.
Where Demographics Still Matter
Demographics are not useless. They constrain the possibility space. A 22 year old student and a 55 year old surgeon have different economic realities regardless of personality. Regional differences in cost of living, cultural norms, and product availability are demographic facts that personality models should incorporate, not ignore.
The argument is not "discard demographics." The argument is "demographics are the floor, not the ceiling." Start with demographic targeting to define the addressable market. Then layer personality dimensions to understand the decision mechanics within that market. The combination predicts more than either alone.