A psycho-behavioral segmentation approach aims to provide a more nuanced view of COVID-19 vaccine hesitancy in order to create segment targeted interventions for COVID-19 vaccine uptake.
As there are large variations and complexity in individual tendencies and environments, barriers to COVID-19 vaccination are diverse as well - these can include low risk perception, procrastination, hot state heistency and vaccine skepticism or distrust. Given this diversity, the current one-size fits all approach to vaccination is not enough.
Rather than “generating demand”, the psycho-behavioral segmentation approach aims to activate the “latent demand” which exists in the form of needs, preferences and tendencies. Latent demands can be converted into actual demand with the right context and cues to drive uptake.
A segment targeted approach helps build an understanding of the context, core barriers and drivers for each of the population relevant segments. This understanding can be used to design a diverse range of solutions such as targeted or differentiated communication design and service delivery design to drive vaccine uptake.
Need for Segmentation
Norm in Private Sector
Segmentation involves clustering individuals by shared characteristics.
First developed in private sector for three key purposes:
niche, ease of engagement
higher engagement, better subjective experiences and market expansion
efficiency in distribution and supply
In contrast to the traditional demand generation approach which is heavy touch, a behavioral psychology-driven approach posits that demand isn’t ‘generated’, but exists in the form of needs, preferences and tendencies of individuals, which can be converted into actual demand, given the right context and cues. We call this latent demand. Latent demand is often non-conscious and inexplicit, therefore it has to be inferred and decoded.
When the design of products, services and communications is aligned to the behavioral drivers and latent demand of the target population, they are intrinsically driven to engage with the product/service, without any external pushes or influences. This outcome is known as ‘self-selected attention’. Diversity and variability of latent demand necessitates multiple, differentiated solutions for self-selected attention.
Unique Benefits in Development Sector
When used in activating latent demand for global health and development programs, many of the benefits of this approach in private sector, like higher engagement and efficiency, translate well to the sector. Others, like prioritization, do not align with the needs of the sector. Beyond its purpose in private sector, segmentation provides some unique benefits in the development sector.
More inclusive and equitable outcomes due to customization for diverse needs and preferences
Reduced externalities and collateral impact of strategies that may work for some but not for others
Reduced resource and time requirements due to a light touch approach
Managing complexity and scale of behavior change by identifying smaller, meaningful clusters
Universalization is only possible with a differentiated approach that serves even the hardest to reach
Stability, Scalability and Predictive Value
Typical Knowledge-Attitude-Preference (KAP) Surveys
Based on self-reports of individuals’ attitudes, beliefs, preferences and intentions
Self-reports don’t capture non-conscious tendencies, therefore insufficient for latent demand
Attitudes and preferences are unstable and context dependent, therefore not predictive of real-world behavior
Gives us the current preferences but not the strategies to change them
Deconstruct decision-making, intent formation and preference construction
Capture the components and processes
Build a psycho-behavioral model to not only understand current preferences, but also predict preferences in other contexts
Response to programmatic interventions
Quantitative Survey Design Overview
The journey framework and the enabler-barrier maps, which were the output of the qualitative research, formed the basis of the design of the quant research instrument. The survey questions were designed to quantify and measure these journey dynamics and enablers/barriers, thereby capturing the psycho-behavioral drivers of decision-making, attitude formation and preference construction. The survey content was uniform across geographies, and these were conducted in local languages by trained moderators using a custom-designed tablet interface.
Semi-Supervised AI/ML-Driven Iterative Exploratory Approach
An iterative exploratory approach using semi-supervised Artificial Intelligence/Machine Learning tools was used to build predictive segmentation models . The factors underlying the model were not predetermined. The iterative analytic process and competing models identified them as existing in the data to use in a final model.