UPDATED September 2021
Audience segmentation is fundamental to social and behavior change interventions. Segmentation serves to align messages, message delivery channels, products, and services with the needs and preferences of an intended audience to maximize program impact. Segmentation divides a population or market into subgroups that have, or are perceived to have, meaningfully similar characteristics, and significant differences from other subgroups. Audiences may be segmented based on demographic, attributional, psychographic, behavioral, or other key variables. Effective segmentation recognizes that the behavior change problem of interest may vary by segment and that different groups will respond differently to social and behavior change (SBC) approaches.
Advanced audience segmentation builds on traditional segmentation approaches by using mixed methods of research to hypothesize and test audience segments. Advanced audience segmentation relies on a phased process of background research, in-depth qualitative research, quantitative research, and advanced statistical analysis to create a representative base of audience segments. It produces a framework to understand how certain beliefs, behaviors, or needs vary across a given population. This rigorous process offers greater nuance and insight into how to best identify and reach priority groups to drive behavior change.
Why use advanced audience segmentation?
Program strategy can be effectively informed by traditional segmentation approaches focusing on demographic, psychographic, or attributional factors. However, SBC interventions can improve their tailoring and targeting of messages, or improve product design, pricing, and positioning by using advanced audience segmentation.
Segmentation can help to:
- Improve understanding of an audience’s experiences, desires, concerns, needs, and behaviors.
- Identify, estimate, and prioritize the potential for behavior change among a specific group .
- Predict the most promising opportunities for resource allocation.
- Tailor services, products, and interactions to specific groups.
- Shape communication efforts to effectively drive awareness, engagement, mobilization, and uptake.
For example, in contexts where the unmet need for family planning is low, advanced audience segmentation can help identify potential new users with the greatest unmet need and the highest willingness to adopt modern methods of contraception.
What are the steps?
Advanced Audience Segmentation Process
- Review your objective
- Synthesize background research
- Conduct qualitative audience research to provide insights into current behavior
- Develop hypothesis segments
- Conduct quantitative audience research to establish segments
- Derive and define audience segments
“Segments are not created but uncovered”
A Breakthrough ACTION How-to-Guide
- Define advanced audience segmentation
- Understand the importance of advanced audience segmentation to SBC
- Understand the advanced audience segmentation process
For comprehensive learning of the Advanced Audience Segmentation Process, with recent case studies (for high as also low data and resource availability), Breakthrough ACTION has created the new Applying Segmentation to SBC in Family Planning course that can be found on the Global Health eLearning Center (GHeL) platform.
Step 1: Review Your Objectives
Before you begin the segmentation process, ensure that your program objectives are clearly defined and you have completed both a situational analysis and audience analysis. This will help you identify which opportunities you intend to address through improved segmentation of the intended audience. The results of your segmentation exercise will vary depending upon your program objective.
Step 2: Synthesize Background Research
The next step is to identify relevant characteristics of the broad target population. Conduct an initial brainstorm based on internal knowledge and a review of existing data to identify characteristics that may be used to segment the intended audience. Gather stakeholders whose work will be most impacted by the results of the segmentation process and engage in discussions on the ideas or assumptions about the broad target population. A synthesis of the knowledge and insights from this brainstorm should be used to identify topics for exploration through primary research.
Then use peer-reviewed and grey literature from public health and other relevant disciplines to identify gaps in knowledge about your target population. Start by investigating existing evidence on knowledge, attitudes, and practices through national or subnational data. If local data is not available or sufficient, review data from contexts with cultural, religious, or other significant similarities. Consider interviewing key informants who can provide added perspective on society, religion, and culture and how those factors might influence the behavior(s) of interest.
Once you have this background information, identify a set of variables that appear most relevant to the target population. It is important to consider the relative strength of each category of variables in predicting the factors that underlie behavior change. For complex or difficult behavior change challenges, it is useful to consider socio-demographic, geographic, and lifestyle variables, but it most helpful to focus on factors that are most likely to drive the intent to change behavior. For example, variables representing the main constructs of the Integrated Behavior Model, are likely better predictors of behavior change than demographic variables. Strongly consider focusing your research plan on investigating the variables in the highlighted boxes below. Next, develop a research plan to explore those variables and to uncover other important characteristics.
Table 1: Types of Variables to Consider
|Age||Urban/Rural/Peri-Urban||Activities||Duration of Behavior||Attitudes|
|Employment||Region||Risk Profile||Frequency of Behavior||Beliefs|
|Status||District||Socially disconnected or disconnected||Habits||Interests|
|Ethnicity||Hamlet||Social Standing||Salience of the Behavior||Intentions|
|Gender||State of ChangeUse of Technology||Opinions|
|Family Size||Use of Technology||Peerceptions of social norms regarding behavior|
|Literacy||Perceptions of self-efficacy in performing behavior|
Step 3: Conduct Qualitative Audience Research to Provide Insights into Current Behavior
Now you have potential variables that may be significant factors in facilitating change among your intended audience and a set of information gaps to investigate. To create a strong segmentation, it is important to understand the range of factors that may influence your audience’s decision-making. To gain a comprehensive perspective on key decision making factors, conduct a qualitative investigation of the needs, attitudes, and behaviors of the population. This helps to identify what the barriers and motivators might be to adopting the desired behavior.
There are a number of qualitative research methodologies that may be appropriate to use. Data may be collected through methods such as focus groups discussions, interviews, consultations, and observations. Participatory methods such as human-centered design workshops and photovoice may also help to uncover differences between subpopulations that will guide your segmentation.
The qualitative audience research should include various populations relevant to the behavior of interest. For example, if your program is trying to increase family planning use among women of reproductive age nationally, your qualitative research should consider sampling a range of women that might include urban and rural women, younger and older women, married and unmarried women, or women at different critical life moments (e.g., recently married, new mothers, etc.). Results of qualitative research should help inform and refine the sampling strategy for the quantitative research.
The results of the qualitative research exercise will provide you with insight into behavioral determinants, including barriers, facilitators, and influences on the intended audience, as well as other socio-cultural factors associated with the desired behavioral outcome.
Step 4: Develop Hypothesis Segments
Now that you’ve conducted your qualitative research and analyzed the data, it is time to hypothesize target segments. Identify the characteristics that make subsets of the audience significantly different from the other subsets. A significant difference is one that requires different messages or approaches to reach and influence the audience.
There are several factors that can be used to segment an audience and each has its advantages and limitations. Again, your approach to segmentation will depend on the program and behavioral objectives identified as a result of your situation analysis, audience analysis, and program analysis. Consider the various types of segmentation to determine which ones would make potentially viable audience segments. At this point, you should hypothesize as many segments as there are meaningful differences in subgroups related to the behavior change objective.
Traits such as age, education level, and income are typically not enough to form the basis of an effective SBC segmentation strategy. Values, tastes, and preferences are more likely to influence how an audience might react to messaging and activities and are, therefore, more useful in identifying segments. The hypothesized segments will help you predict which groups are most likely to adopt the behavior, and the major factors to address in order to enable those behaviors. Before you move into the quantitative research phase, outline what you know about your segments, what you believe to be true about them, and what remains to be uncovered.
Table 2: Segmentation Types
|Segmentation type||Segmentation characteristics||Advantage||Limitation|
|Combination of Needs, Behaviors, and Attitudes||Based on needs, attitudes, and willingness to change behaviorAcknowledges the needs that drive behavioral differences||Helps answer the question, “Which benefits and features matter most to the audience?”Allows you to determine the relevant value proposition||Difficult to identify segmentsMore challenging to target outreach|
|Behavioral||Based on observable behaviorUncovers what people are doing but not why they are doing it||Identifies behaviorUseful for understanding how to market a product or service||Does not explain behavior or predict future behavior|
|Psychographic||Based on the psychological factors of the audienceCan capture some truth about personality traits or values, self-image, and aspirations of the audience||Identifies receptive audiencesUncovers communication preferencesAllows for developing messaging that appeals to emotions||Does not provide specificity to change behaviorThese characteristics can be expected to change with a person’s values and environmentIs weak at predicting what someone is likely to use or adopt with a set of options|
|Attributional||Based on a single attribute such as life stage||Identifies underlying drivers and believesProvides insights for messaging||Ignores attributes that may be greater determinants of behaviorDifficult to identify segments|
|Demographic||Based on factors such as gender, age, income, and geography||Data are often readily availableEasy to understandEasy to target||Assumes common needs or behaviors within or across demographic groupsNot predictive of behavior|
|Adapted from USAID’s Transform/PHARE project|
Step 5: Conduct Quantitative Audience Research to Establish Segments
Once the qualitative research is complete and hypothesis segments are developed, use quantitative research to test the hypotheses. Select a set of variables to test based on factors associated with willingness to practice the behavior of interest. Depending on the results of your earlier evidence-gathering exercise, the factors to test through statistical analysis may include demographic, behavioral, and attitudinal factors associated with willingness to adopt the desired behaviors and/or other key considerations.
Thoughtful survey design and recruitment criteria are integral to producing useful quantitative outputs. The survey instrument should be informed by the results of qualitative research, background research, and insights gathered starting from the initial brainstorm and stakeholder interviews. The screening criteria should be based on the hypothesis segments and the sample should be representative by factors such as geography, education, marital status, and income or socioeconomic group. The instrument should investigate needs, attitudes, behaviors, and other factors thought to influence the behavior(s) of interest.
Following an analysis plan that establishes which questions are to be answered, which sub-groups should be used, and which specific hypotheses are to be tested, the data should be analyzed using the appropriate analysis method. Analysis techniques typically used for segmentation included cluster analysis, latent class analysis, and perceptual mapping.
Two examplesthat used robust statistical segmentation are the National Demand Analyses by Camber Collective in Niger through funding from the William & Flora Hewlett Foundation and in Cote d’Ivoire under the USAID-funded Transform/PHARE project. National demand analysis is a method developed in the private sector to enable companies to understand the needs of their customers and how to best shape the consumer experience to attract and retain them. This quantitative approach to audience analysis was used in Niger and, later, in Cote d’Ivoire to identify the demographic, behavioral, and attitudinal factors associated with willingness to use modern contraceptives; the contraceptive characteristics most appropriate to each segment of women; and the communication channels and dimensions of the provider experience that are most important to each segment of women. Statistical analysis of survey data allowed Camber to estimate the size of each newly identified segment and the potential impact their uptake of family planning could have on the national modern contraceptive prevalence rate.
Step 6: Derive and Define Audience Segments
Reflect on the results of the quantitative analysis and the limitations of the data. Question whether the newly identified segments confirm or dispel your initial assumptions and hypotheses. In some cases, your final segments may look very similar to what you initially brainstormed and, in other cases, quite divergent. Transform/PHARE in Cote d’Ivoireidentified three segments based on life stage as a result of their qualitative investigation: pre-family women; family aspirationals; and family planners. Once the full advanced audience segmentation process was complete, six main subgroups of women were revealed: pre-family women; rural passives; independent matriarchs; struggling aspirationals; family builders; and family limiters. The rigorous segmentation exercise provided Cote d’Ivoirewith a far more nuanced national picture of the type of potential family planning users. This then allowed for more targeted programming and messaging strategies to adequately address the audience’s needs and concerns.
As you refine and finalize your segments, ensure the segments meet the criteria for segmentation. Segmentation is a blend of both art and science, requiring evidence and intuition. However, for the segmentation results to be useful, the segments must, at a minimum, meet the criteria outlined in the table below.
Table 3: Criteria for Segmentation
|Actionable||The program is able to reach the segment with distinctive interventions|
|Homogeneous||Members of the segment are similar in terms of needs, attitudes, and preferences and as well as other significant attributes|
|Heterogeneous||Each segment is relatively unique compared to other identified segments|
|Measurable||Data can indicate the size of the segment|
|Salient||The segment is substantial in either size or potential impact to warrant targeted interventions|
|Reachable||The program intervention can reach the segment|
|Recognizable||Program implementers can recognize the segment|
|Responsive||The segment can be expected to consistently respond better to a tailored approach rather than a generic intervention. If the, the segment can be combined with another.|
|Stable||Stable enough to remain relevant for a reasonable period of time|
|Adapted from Criteria for Market Segmentation|
Finally, validate your segments. Share the results of your segmentation exercise with key stakeholders to assess whether they meet the criteria for segmentation within your context. One way to share the segments is to develop persona profiles or archetypes for each segment. The persona is a descriptive summary of a representative individual within the segment. To develop the persona, consider the needs, motivations, and beliefs that may drive a person within the segment to carry out the desired behavior. The profile should illustrate 3–5 characteristics that differentiate the segment from others. The profiles will help you conceptualize the statistically derived segment as a tangible audience.
Depending on resources, program objectives, and feasibility, you can also consider prioritizing select segments. Segments can be prioritized based on various factors. A segment may be prioritized based simply on its size or ease to access. Another means to prioritize is using statistical analysis to determine which audience has the greatest propensity to change behavior or which will have the greatest potential impact on the outcome of interest. Once again, review the insights you gathered through each step of the process and confirm that targeting the prioritized audiences will help you reach your program’s strategic objectives.
Tips & Recommendations
- To best understand audience behavior, a research approach that combines qualitative and quantitative methods is most effective.
- Observational research allows you to learn about an experience from the perspective of your audience and can be very useful in learning about their decision-making process and preferences.
- In your qualitative study, try to include a cross-section of individuals that will help test initial hypotheses and refine the sampling strategy for the quantitative research.
- The important differences that define your hypothesis segments should be based on their needs, attitudes, and behaviors of the audience segments, not only demographics, geography, and lifestyle characteristics.
- Create as many segments as are needed to reflect meaningful differences across the target population, while keeping in mind segment sizes and the overall complexity of the segmentation. While exceptions may exist, effective and practical segmentations tend to produce between four and seven segments.
- Considering resources and feasibility, your program will likely be most impactful if focused on a few priority segments.
- Using the statistical data to determine which segments have the greatest propensity to change behavior will allow you to prioritize based on, and forecast, potential to the impact program targets.
- Needs, attitudes, and behaviors often change with time so segmentation should be an ongoing process where insights continuously feed into program strategy.
Resources and References
Jvani Cabiness: email@example.com
Jim Malster: firstname.lastname@example.org
Jessica Vandermark: email@example.com