2.1. Setting-driven design theory

The theoretical foundation of SDD is rooted in behaviour setting theory (Barker Reference Barker1968; Schoggen Reference Schoggen1989; McGann et al. Reference McGann, Lucas, McHugh and Barrett2024) which is further described and enhanced by a theory of change (ToC) (Aunger & Curtis Reference Aunger and Curtis2016). The ToC, depicted in Figure 2, shows the causal links between interventions, environmental changes and the resulting behaviour. Specifically, an intervention modifies the environment, which then influences how individuals think and feel, leading to a behavioural response. Over time, these behavioural responses contribute to broader outcomes, such as improved health or production efficiency. A behaviour setting constitutes the context in which the environment–brain–behaviour chain is made. This offers a diagrammatic representation of an earlier point that behaviour settings are self-regulating. This self-regulating system consists of several key aspects. The first is the overarching goal of the setting. That is to say that all behaviour works towards some outcome decided at the level of a setting rather than any individual. An understanding of this outcome dictates what is done and what is seen as success within a setting. The next aspect is the creation of a setting with strong synomorphy or “fit” between the environment and the standing pattern of behaviour. In other words, how well the afforded and constrained environment supports the behavioural performance in achieving the setting goal. Indeed, this notion of synomorphy is closely related to the more familiar idea of an “affordance” between objects and people as outlined in several related pieces of work (Gibson Reference Gibson, Shaw and Bransford1977; Lockton et al. Reference Lockton, Harrison and Stanton2010; Lahlou Reference Lahlou2024). The final aspect that governs the self-regulation of a setting is the maintenance of the setting to prevent behavioural drift from the desired synomorphic state of the setting. Maintenance can be proactive or reactive. Proactive maintenance includes the prevention of behavioural drift through behaviours such as modelling and affirming desired behaviour and selective admittance into the setting based on suspected compliance. Reactive behaviours may include anything that corrects setting elements (behavioural or otherwise) to keep the standing pattern of behaviour from drift. Examples of this include normative corrections such as shushing someone to encourage quiet during a lecture and maintenance of the wider supporting environment to maintain synomorphy such as repairing a book in a library to keep it in circulation.

Figure 2. A generic ToC for use with behaviour settings which link interventions to desired outcomes. The behaviour setting comprises the three middle elements of the ToC – environment, brain and behaviour. Adapted from (Aunger & Curtis Reference Aunger and Curtis2016).

The ToC clarifies why certain behaviours (typically the target of behaviour change) occur within a given setting and how interventions can be framed to disrupt existing patterns, leading to lasting change. This framework encourages designers to consider both immediate interventions that address pressing needs and strategies for ensuring long-term sustainability of behaviours. Crucially, successful interventions are not only effective in producing the desired behaviour initially but are also designed to embed and reinforce that behaviour within the setting, fostering a self-sustaining cycle of change over time.

3.1. Objective

The BSC forces teams to be explicit about the objective (which can be different from the motivation they have as a role-player within the setting – see below). The objective is embodied in the setting’s end-state. For instance, a grocery store checkout is a behaviour setting with the objective of having customers pay for purchases. There may also be other objectives such as to provide friendly or efficient service to customers. The objective not only focuses the team in terms of things to include in the setting description but also areas of particular interest that may help or hinder achieving the desired outcome.

3.2. Setting, date and time

In the top middle of the Canvas is a place to note the setting being investigated, the date and the time. This is for practical record keeping, as many different settings may be explored. For instance, this may note a “business meeting” setting within a particular hotel lobby at 2 pm if you are making observations of such settings in various hotel lobbies at different times of the day. The setting can also be any other relevant setting, such as a doctor appointment, factory assembly line, public transport journey or family dinner.

3.3. Stage

This box captures the backdrop in which the setting occurs. Behaviours occur within circumscribed locations; offering a short description of this “stage” is an essential way to provide significant information and set the scene in which behaviour is performed. In a single stage, there could be multiple settings delineated by various temporal, social or physical thresholds. An example might be a swimming pool (stage) that could be the backdrop for several settings, such as lap swimming, pool party or swim meet. Another could be a café area of a building (stage) which, at mealtimes, could support the lunch eating setting but outside of this time could serve as a group study setting or a birthday party setting.

3.4. Agents

All those actors in the setting who have agency are recorded in this box. This includes individuals who contribute to the behaviour (or lack of behaviour). Generally, this could feature the “protagonist” (the person who is doing a target behaviour) but may also include other contributors. In his early work, Barker differentiated levels of involvement and responsibility among inhabitants of a setting from the outermost to innermost including onlooker, audience or invited guest, member or customer, active functionary, joint leaders and finally the single leader (Barker Reference Barker1968). Depending on the setting in question, it can be useful to differentiate people in other ways, including the emotional salience of the experience, familiarity with cultural expectations, expertise or other dimensions. In some cases, you may have other entities that have some autonomy in a setting such as pets or artificial intelligence that exhibit a degree of agency.

3.5. Props

These are objects that are manipulated in the setting during the completion of some aspect of the setting’s behavioural performance. In a physical setting, these are typically products, buttons, packaging, coats, pens, handles, masks, doorknobs or other physical things individuals touch and interact with to achieve a behaviour. In a digital setting, a prop could be an aspect of an interface such as buttons, virtual tools or other interactive elements.

3.6. Infrastructure

The infrastructure includes all physical and digital aspects of the setting that may not be moved, but still contribute to the setting behaviour. Examples of physical infrastructure could be a sink basin for hand washing, a wall where posters are displayed, a kitchen table for mealtime or a whiteboard within a classroom setting. In a digital context, these are similarly supportive elements, such as a database or webpage. The difference between props and infrastructure is whether people actively manipulate it in the setting though such a distinction is sometimes grey, and this does not tend to be an issue practically if documented in one section or the other.

3.7. Roles

Agents (e.g., people), props and infrastructure each have functionally distinct roles that help in the performance of the setting. For instance, in addition to the obvious role a surgeon has within a surgical theatre (to do the surgical procedure), they may also play the role of delegating tasks, explaining the procedure to the patient, as well as more subtle roles such as conveying trust and more. Another example may be with a teacher who, of course, has a role of conveying information but will often also have a role of correcting misbehaviour and generating class participation. The use of functions (Taylor Reference Taylor1969) or “jobs to be done” (Christensen et al. Reference Christensen, Hall, Dillon and Duncan2016) is common in various areas of design, engineering and business and can be analogous to how roles are used here. For props and infrastructure, list the functions that the object has. Refining the title given to a role will help shape how the element is positioned within the broader setting.

3.8. Attributes (and competencies)

People, props and infrastructure all have specific attributes or competencies required to effectively fulfil a role. In other words, attributes are the embodied features and characteristics that help enable a role to be performed. For a surgeon to effectively explain the procedure to the patient (a role of an explainer or comforter), they may need an attribute or competency around good bedside manner to fulfil that role. Even more descriptive would be to say approachable, articulate and knowledgeable. A prop used in a surgical setting such as a scalpel has the role of making incisions. To effectively fulfil that role, the attributes it requires are being sharp, small and fit in a human hand. People have roles which often require competencies rather than attributes. Clarifying the link between roles and the attributes or competencies to fulfil them is an important part of the analysis and subsequent intervention development.

3.9. Motives

The goal or benefit a person hopes to gain by fulfilling a role is their motive. Action in behaviour settings generally utilises short-term rewards to regulate behaviour. Thus, while the section of the canvas related to motives can be used generally, the motives commonly used in BCD are effective here, given their strong theoretical (Aunger & Curtis Reference Aunger and Curtis2013) and empirical underpinning (Aunger, Foster & Curtis Reference Aunger, Foster and Curtis2021; Aunger, Gallyamova & Grigoryev Reference Aunger, Gallyamova and Grigoryev2025).

In addition to identifying which motive(s) best help explain the target behaviour(s) in a given setting, it may be helpful to note if motives are intrinsic (do something for its own sake or personal reward) or extrinsic (doing something for external recognition or reward). Other implementation factors, such as costs (financial, temporal, energetic, etc.), which may be seen as an impedance to acting on a motive, can also be useful to note within this box.

3.10. Norms

Norms are the implicit or informal rules that govern behaviour in a behaviour setting. A simple way to conceptualise this is by considering what is expected in a situation (Deutsch & Gerard Reference Deutsch and Gerard1955). Norms are often thought of in relation to other people (as expectations of someone), but they can also relate to props or infrastructure, e.g., with respect to their use in the expected behaviours. For instance, the normal (and expected) performance of a material, whether an interface feature signals a particular action or expectations around how automatic doors operate. Norms fall into two categories (Cialdini, Reno & Kallgren Reference Cialdini, Reno and Kallgren1990). Injunctive norms are those behaviours seen as approved or disapproved (and hence punishable). Descriptive norms are those behaviours observed to be performed. An example may be seen with public transportation. An injunctive norm may be that people should pay for their fare, but you may see people getting on without paying and thus note a descriptive norm that people travel without paying.

Norms can have strong associations with settings and be strong drivers of behaviour. Indeed, the strength of the norm associated with a setting correlates with adherence to expected normative behaviour. Critically, such behaviour is not driven by mood but expectations of how to act (Aarts & Dijksterhuis Reference Aarts and Dijksterhuis2003). Understanding the manifestation and mechanisms underlying these norms is crucial to drive action at the level of a setting. Particularly useful is to understand how desired normative behaviour is praised and deviations are corrected within a setting – something (Barker Reference Barker1968, p. 169) calls deviation countering mechanisms of behaviour settings which provide the self-regulation of the setting.

3.11. Routine

The routine maps the sequence of behaviours from the initiation to completion of a behaviour sequence in a setting by the agents. It maps how agents, props and infrastructure interact during the enactment of the setting’s objective and shows how each element (agents, props and infrastructure) plays a role over time. Think of it as a script or step-by-step process that guides the actions of everyone involved in the setting rather than a list of isolated actions. By understanding the routine, designers can identify where things might go wrong, where the setting can be improved or how an intervention might be applied. This visualisation focuses on observable behaviours rather than motives, norms, thoughts or feelings. It is similar conceptually to the “user journey” in human-centred design.

4.1. Scoping behaviour setting challenges

The first phase of the SDD process is to scope a challenge whose objective fits within a behaviour setting framework. The most straightforward challenge is one in which an existing setting is to be influenced or disrupted, as would be the case with behaviour change projects. However, SDD’s tools have proven versatile across a range of applications where behaviour may be influenced but the focus is on another challenge, such as introducing novel products, reshaping social norms or refining existing processes (e.g., improving synomorphy within the setting). One common application is a technological intervention in which new technologies augment or automate actions that individuals already perform within a setting. Finally, there may be other relevant challenges, such as creating a new setting altogether, such as when a new policy, technology or other force suggests a fundamentally new way of arranging behaviour and environment to reach an objective. This phase is completed when the team agrees on the challenge, objective and alignment with behaviour settings.

A sample of SDD challenges successfully addressed to date includes the following:

• • Encouraging a patient to take medication using a conversational agent.

• • Improving how people form connections at a networking event.

• • Integrating new sensor and actuator technology to enhance group collaboration in smart offices.

• • Increasing handwashing with soap after defecation.

• • Improving training and performance of personnel in an industrial setting.

• • The limitations and opportunities afforded in extended reality settings.

• • Understanding the features of a car used to support different use cases (i.e., settings).

• • Supporting wider participation among older people in physical activity events to encourage fitness.

• • Normalising desired behaviour such as the prevalence and effectiveness of a designated driver.

There are also many challenges that might sit outside of this frame. For instance, the use of a different theoretical lens through which the challenge is viewed may give precedence to a different framing. The investigation of interactions with a single object across multiple settings (e.g., exploration of phone usage throughout the day) may not always benefit from SDD since there is so much variability in the settings. Finally, there may be an interest in behaviour that exists outside of the standing patterns more generally.

It is common – and expected – for the understanding of the challenge to evolve as the project progresses. The primary goal is to ensure the challenge is behaviour-setting related, providing a clear focus for subsequent phases of the process.

The boundaries of a setting are crucial to defining and understanding its scope. Settings can be delineated by a combination of physical, temporal and social thresholds. A setting is characterised by a “standing pattern of behaviour,” and it is not uncommon to observe multiple settings occurring in sequence, such as shopping in a grocery store followed by checking out at the cashier. Each of these would be considered a distinct setting, even if they are part of a larger behavioural sequence. While scholarly discussions delve deeply into the theoretical and empirical aspects of setting boundaries (Schoggen Reference Schoggen1989; Sasmita & Swallow Reference Sasmita and Swallow2022; Lucas Reference Lucas2024), a pragmatic approach is more appropriate for design work. This involves clearly defining the start and end points of the setting under investigation to ensure a focused and actionable understanding.

A well-framed challenge should explicitly define the overall objective within the setting, as this determines how the setting will be viewed and interpreted throughout the project. For instance, in the example of a grocery store checkout, shifting the objective from efficiency to creating a personal experience fundamentally changes how the setting is analysed and the types of interventions considered. Moreover, a single project may encompass multiple objectives, which should also be noted during this phase to provide clarity and guide decision-making as the work progresses.

4.2. Understanding the behaviour setting

Once the challenge is scoped, the second phase focuses on developing insights through fieldwork. Initial understanding of the behaviour setting often begins with observation, allowing the team to gather foundational information about how the setting functions. To delve deeper into specific elements, contextual interviews with people involved in the setting are typically required. Table 1 provides a summary of appropriate methods for interrogating the various elements of a setting.

Learnings from this phase are captured and summarised in the BSC. Work in this phase is considered complete when the setting is fully understood, and the canvas reflects this understanding. Thus, the canvas would, in this phase, represent the current state or “as is” setting condition and only later be considered with a desired change in mind. While the canvas can document findings from a single setting, it is often beneficial to study multiple comparable settings (e.g., several grocery store checkout settings) to develop an understanding of an archetypal setting. This helps generalise insights and ensures they are applicable across similar contexts. A sufficient analysis is generally understood to be achieved when rich and nuanced information is gathered and additional conceptual insight is no longer emerging from the data collection process. This is consistent with approaches to qualitative thematic analysis that look at adequate coding as determined by the richness of codes as interpreted from the team rather than a strict view of data saturation (Brooks et al. Reference Brooks, McCluskey, Turley and King2015; Braun & Clarke Reference Braun and Clarke2021b, Reference Braun and Clarke2021a) This will mean that the specific detail and way a team engages with data analysis will and likely should vary according to the specific constraints of the brief, the background of the team and the broader context of the project. Indeed, as shown in the diagram, this process will often be iterative as some aspects of a project often cannot be fully understood beforehand and additional understanding is needed to proceed through the remainder of the work. In practice, the BSC then not only acts as a template for coding and organising findings but also provides a structured basis for discussion to confirm that the meaning behind the data is fully understood.

In some cases, this phase may also include additional activities, such as identifying constraints, feasibility requirements or other relevant considerations that influence the design process but that would be documented beyond the canvas.

While there is no fixed procedure for completing the BSC, the following sequence is recommended based on its logical progression and ease of use:

1. 1. Behaviour, Objective, Operational Notes and Stage: Start by filling in the overarching behaviour, the objective of the setting, any operational notes and the general stage in which the behaviour occurs.

2. 2. Routine: Document the routine step by step, capturing the sequence of actions from the initiation to the completion of the behaviour.

3. 3. People, Props and Infrastructure: Identify the agents involved, the props they interact with and the broader infrastructure that supports the setting.

4. 4. Roles and Attributes: For each agent, prop and infrastructure element, outline their roles and the associated attributes or competencies they require.

5. 5. Norms and Motives: Document the norms governing the setting and the motives driving the behaviour of agents.

Following this sequence ensures dependencies are addressed and complexities are unravelled systematically. Only the elements relevant to the setting should be included in the canvas, as additional or tangential details may serve to distract. It is common for this to occur since there is much in a stage that may have little to do with a particular behaviour setting. However, good design work often requires a broader understanding of the challenge, incorporating considerations such as user adoption, institutional constraints or other contextual factors that influence the success or mere possibility of an intervention. While the canvas serves as a central anchor for synthesising and visualising formative research, it is not a substitute for the broader formative research activities needed to comprehensively understand the challenge space.

The BSC serves as a summary of the target setting, highlighting its key elements and their interrelations. While additional notes or descriptions may be required to detail specific elements, the canvas provides a structured, high-level view of the setting. Its primary purpose is to guide the team in understanding why and how the setting functions as it does, enabling informed design decisions. The canvas may also be used for other purposes, such as creating hypothetical scenarios, but its traditional use is to document existing conditions.

In addition to the canvas being completed, the design team should document insights that will guide the next phase of work. We differentiate between two kinds of insights. Descriptive insights give a deeper understanding of why and how behavioural performance occurs in the current situation. Prescriptive insights are information that, when properly contextualised and embodied, can drive the desired behavioural outcomes. Insights should offer a clear understanding of how the setting might be changed to achieve the intended result. The canvas acts as a grounding tool, helping the team evaluate how each element of the setting might be adjusted and what implications these changes may have. Insights should ideally be documented formally to ensure clarity and alignment across the team. This is particularly important when members of the team may change, such as adding additional creative support, moving into the next phase of work, where interventions are developed.

Effective insights share certain qualities:

• • Specificity: Insights should be actionable and directly linked to elements of the canvas. They should reflect the particular fieldwork done.

• • Grounded in Qualitative Data: Insights should be based on fieldwork, including observations and interviews, to ensure they reflect the realities of the setting.

• • Connection to Theory: Insights should draw on the ToC, canvas or other principles of behaviour settings (e.g., deviation countering) to provide a rationale for how changes to the setting will lead to desired outcomes.

Insights may range from observational (e.g., “Gendered roles are critical in how tasks are divided within families”) to theoretical (e.g., “If a core motive is embodied in the intervention, it should drive behaviour change”). A deeper level of insight often combines both, creating a path forward for intervention.

An example of insight: In Tanzania, men are typically viewed as providers of a home, which is seen as a status symbol. However, the toilet is not traditionally regarded as part of this status. By extending the role of “provider” to include the toilet facility, an intervention could leverage the status motive to encourage men to improve residential sanitation facilities (Aunger, Mwambuli & Cardosi Reference Aunger, Mwambuli and Cardosi2023).

A useful structure for documenting insights is to outline the state of the world, the identified issue and a potential way forward. This approach ensures insights are not only descriptive but also prescriptive, providing a clear path to action.

4.3. Intervention development and integration

The third phase in SDD focuses on developing and integrating contextually grounded interventions that address the challenge identified in earlier phases. These interventions should align with the specific elements of the behaviour setting, ensuring they are embedded in and relevant to the context. Success in this phase is achieved when an intervention based on one or more insights from the earlier phases is embodied and effectively integrated into the relevant behaviour setting(s).

Interventions are embodied (i.e., given form) through one or more of the following:

1. 1. Agents: Changes involving agents within the setting, such as training, role creation or behaviour adjustment.

2. 2. Props: Modifications to or the addition of objects or tools that agents interact with during the behaviour, such as redesigned products or interfaces.

3. 3. Infrastructure: Adjustments to the broader physical or digital environment, such as layouts, systems or built spaces.

4. 4. Communications: Messaging, prompts or educational materials that influence behaviour indirectly (e.g., through role modification or increased motivation). Note that communication often exists outside the immediate setting and is thus not an element of the setting but is still a common intervention form that is meant to help drive behaviour by changing how people think within the setting.

Selecting the appropriate form(s) of embodiment is crucial to ensure the intervention is aligned with the behaviour setting and maximises its impact. Ideally, the intervention should be as close to the point of action as possible, directly addressing the behaviour at the moment it occurs.

In many cases, a single intervention may not suffice. Instead, multiple interventions are combined to address different aspects of the behaviour setting, forming an overarching intervention strategy. For instance:

• • Agents: Linking the role of parents to teaching children to wash their hands.

• • Props: Introducing easy-to-use handwashing stations.

• • Infrastructure: Ensuring consistent water supply.

• • Communications: Developing campaigns to motivate participation and normalise desired behaviours.

These interventions, when implemented together, reinforce each other and create a more robust approach to driving sustained behavioural change.

There are at least six ways the setting elements, as outlined on the canvas (i.e., the boxes) and in Table 1, can be modified to develop interventions in the setting:

1. 1. Adding: Introducing new elements to the setting that support the desired outcome.

2. 2. Removing: Eliminating elements to support the desired outcome.

3. 3. Strengthening: Enhancing existing elements to make them more effective at producing the outcome.

4. 4. Weakening: Reducing the influence of elements that promote undesired outcomes.

5. 5. Transforming: Changing the function or role of an element within the setting.

6. 6. Swapping: Replacing one element with another to achieve better alignment with the setting objective.

This list can act as a set of prompts to support ideation, particularly when focusing on changing an existing setting.

The ToC is a valuable tool for guiding this phase, from ideation through prototyping and evaluation. By mapping out the causal links between the intervention, the behaviour setting and the desired outcomes, the ToC helps ensure the intervention is both intentional and effective. Specifically, it plays several critical roles:

1. 1. Strengthening the Causal Links: The ToC ensures that the design is grounded in a logical progression, where each element of the intervention contributes meaningfully to the desired behavioural change. This helps during ideation to identify which elements of the setting need to be adjusted and why.

2. 2. Guiding Ideation: During ideation, the ToC provides a structured framework to generate ideas for interventions that align with the desired outcomes. By focusing on specific points in the causal chain, the team can brainstorm targeted ways to introduce change, ensuring that ideas are contextually relevant and likely to produce the desired effects.

3. 3. Identifying and Mitigating Risks: By anticipating where the intervention might fail (e.g., a node in the ToC does not lead to the next as expected), the ToC helps the team proactively address risks and adjust the intervention to improve its effectiveness. For instance, the change in the environment may produce an expected brain response or feeling but that might lead to a behaviour not expected or desired. Thus, the team can ask how they might mitigate the workaround there.

4. 4. Informing Prototyping: During prototyping, the ToC helps the team identify key elements of the setting to test and validate. It highlights potential weak points or risks in the intervention’s design, such as causal links that might not function as expected, enabling the team to refine the intervention iteratively.

5. 5. Planning Evaluation: The ToC provides a clear framework for identifying what to measure and why during evaluation. By linking metrics to the causal chain, it ensures the evaluation captures both the process and the outcomes of the intervention. For example:

   • ○ Process Metrics: Monitoring whether the intervention is being implemented as designed.

   • ○ Outcome Metrics: Assessing short- and long-term impacts.

6. 6. Executing Evaluation: The ToC helps track progress along the causal chain during implementation. By identifying breakpoints where the intervention may not be achieving the intended effects, the ToC provides actionable insights for improving the design at the point where the issue occurs rather than discarding all work. It also helps evaluate sustainability, ensuring the intervention produces lasting change in the behaviour setting. This latter point is particularly informed by the self-regulating aspect of behaviour settings – what elements help to correct the behaviour when it varies from expectation or ideal.

Explicitly writing out the ToC for one or more interventions at even a conceptual stage is a useful process for facilitating the discussion, refinement and selection of an intervention.

Prototyping is essential for testing and refining interventions. During this phase, the team performs the following:

• • Validates Assumptions: Ensures the intervention aligns with the setting and behaves as intended.

• • Identifies Challenges: Observes how the intervention interacts with the setting and addresses unforeseen barriers.

• • Refines the Design: Iterates on the intervention to improve its effectiveness and feasibility.

Even simple prototyping can greatly reduce costly changes later in the intervention process.

The phase concludes when one or more insights have been successfully embodied and integrated into the behaviour setting through interventions that address the specific challenge. These interventions should be actionable, sustainable and deeply aligned with the setting, ensuring that they influence behaviour effectively and persistently. Important to note with the above are factors beyond the setting, such as institutional constraints or ongoing roles and costs associated with the maintenance of interventions. The team should also consider more fringe interactions that might highlight roles or behaviours that may not be adequately cared for in the original setting. In a grocery store checkout, many of the roles were transferred to the shopper and the technology to support self-checkout. However, employees have continued to be needed for some tasks such as verification of age checks, supporting when technological issues arise or removing anti-theft protection.

5.1. Background to the challenge

Despite its simplicity and proven benefits, handwashing with soap remains a persistent challenge to promote sustainably and at scale. Various public health projects have addressed this issue through education, infrastructure development and supply provision, with mixed results. In this project, SDD was applied to address the unique challenges faced by households in unplanned urban settlements in Tanzania, where there was a limited tradition of handwashing with soap, even after defecation. The objective was to design a behaviour setting-specific intervention that could sustainably increase handwashing with soap in these resource-constrained environments.

5.2. Scoping the challenge

The project began with refining a broad behavioural challenge: increasing post-defecation handwashing with soap among low-income populations in Tanzania earning less than $5 per day. This behavioural challenge was explicitly linked to a specific behaviour setting – the household post-defecation hygiene setting. The scope was further refined to focus on residential use rather than communal settings (e.g., schools or workplaces) and on developing an intervention directly tied to the soap itself rather than communication campaigns or environmental changes as a constraint of the brief due to the interest of the funder and collaborators who primarily have worked on public health campaigns rather than product innovation. Figure 5 provides an example of one of the toilet areas explored in the project.

Figure 5. An example of a toilet setting explored in the Tab Soap case study with internal and external views and a focus on key props (e.g., buckets, soap.).

Initial scoping workshops, involving stakeholders from government, businesses and public health sectors, helped frame the challenge within the behaviour setting context. The workshops identified key barriers, including the lack of soap products clearly positioned for handwashing in low-income households, and established constraints for the project, such as affordability, simplicity and cultural relevance.

By the end of this phase, the challenge was clearly articulated as follows: How can we develop an affordable and contextually appropriate soap product that enables post-defecation handwashing within the household post-defecation hygiene setting? This framing provided the foundation for the subsequent phases of the SDD process.

5.3. Understanding the behaviour setting

The second phase involved fieldwork and analysis to develop a detailed understanding of the behaviour setting and identify actionable insights. The BSC served as the central tool for mapping elements of the setting and structuring the research process. Alongside generating insights, the team also considered constraints (e.g., affordability, production feasibility) and requirements (e.g., cultural appropriateness, ease of use) common to product design and human-centred design work to ensure the intervention would be both practical and impactful.

The research process included the following:

1. 1. Secondary Research: Reviewing existing studies on handwashing behaviour and public health challenges in low-income contexts.

2. 2. Stakeholder Workshop: Engaging government, business and public health stakeholders to consolidate existing knowledge about the behaviour setting. This combined with the secondary research to produce the first draft of the canvas.

3. 3. In-Context Interviews: Conducting interviews with households across several locations in Tanzania to understand soap use, handwashing routines and the social and physical dynamics of the setting. Since observing post-defecation handwashing directly was not possible for ethical and practical reasons, participants demonstrated their routines and described norms, motives and other elements of the behaviour. These activities iteratively developed the canvas further and were refined as data collection continued to understand all parts of the canvas.

Findings from these activities were synthesised into a completed Canvas (see Figure 6) to represent the archetypal setting. The Canvas captured details such as the following:

• • Infrastructure: The roof’s role in protecting soap from theft and weather while defining the physical boundaries of the toilet area.

• • Routine: A clear sequence of actions following defecation, with variations noted across households.

• • Norms and Motives: Cultural perceptions of soap and its uses within the household.

Figure 6. A completed BSC for handwashing with soap in Tanzania. This canvas represents a summary of fieldwork and thus is a pared-back version of the understanding developed. The letter “v” is used to indicate where some variability was seen.

The canvas also critically acted as a boundary object throughout the understanding phase by facilitating directed conversations between the immediate research team and the other relevant stakeholders (i.e., during the stakeholder workshop and later in the ideation part of the next phase of work). Specifically, it helped ensure that the setting was fully understood, grounded conversations about how various setting elements relate and clarified how and what various stakeholders were referring to when communicating about findings as many of these disciplines use the same terms in different ways.

Through this process, several insights were developed, each aligned with different aspects of the Canvas. One particularly critical insight related to a missing role in the props category, multipurpose soap, while commonly available, was not perceived as suitable for handwashing. Participants associated bar soap with tasks like laundry, bathing and dishwashing, while liquid soap was seen as an aspirational product used exclusively for handwashing but priced beyond their reach. This gap highlighted an opportunity to design a soap product specifically for handwashing that would fit within the mental models and price points of low-income households.

This insight, along with others such as a fundamental understanding of human motives that might be embedded in the design, was synthesised into a detailed design brief, providing a clear direction for the intervention phase. The brief emphasised the need for a product that communicates handwashing as its primary function while remaining affordable and easy to use in the household post-defecation hygiene setting, and particularly the stage which is the toilet area itself. This, along with constraints, requirements and other detailed notations from the previous phase, provided a basis for entering the intervention phase.

5.4. Intervention development and integration

The final phase focused on translating insights into actionable, contextually grounded interventions. Using the completed BSC as a guide, the team developed and refined the intervention through a structured process of ideation, prototyping and testing. Ideation relied on a number of insights developed during the previous phase, but also on fundamental truths relating to behaviour settings (see section on Core Tactics for Intervention Development). One intervention that emerged from this process was Tab Soap, a biodegradable swatch of fabric impregnated with soap, designed to promote handwashing in low-income households.

The original concept for the intervention leveraged curiosity as a motive which is understood here as gathering information about whether or not one has won a prize as revealed after the soap is used. This early idea (Figure 7) involved integrating the soap swatch with a reward system: each swatch contained a code that could be revealed when the soap was used, offering users a chance to win a prize, such as a mobile phone. This approach aimed to make handwashing both fun and incentivised. However, while participants expressed interest in the prize-based concept during early interviews and testing, several critical issues emerged:

1. 1. Risk of Misuse: Some participants indicated they might rinse the swatch solely to reveal the code, bypassing its intended use for handwashing.

2. 2. Operational Complexity: Implementing the prize system posed significant logistical challenges, including distributing prizes and managing a redemption process, which would add considerable cost and complexity to the intervention.

3. 3. Compelling Core Features: Feedback showed that participants were highly interested in the product’s core attributes – its portability, single-use design and hygienic appeal – even without the prize mechanism. These factors became stronger motivators than the original curiosity-based approach.

Figure 7. An early sketch of the soap and associated reward system.

Based on these findings, the team decided to focus on Tab Soap as a standalone product without the prize feature. This shift simplified the intervention and allowed resources to be concentrated on refining the core design to maximise its effectiveness and scalability. Figure 8 shows a user providing feedback on the concept.

Figure 8. An image of a user trialling and providing feedback during testing of the Tab Soap concept.

The Tab Soap intervention was developed and refined through a series of iterative steps:

1. 1. Ideation and Refinement:

Initial brainstorming sessions produced a range of concepts, including the prize-based soap swatch idea, as shown in Figures 7, 8 and 9a. These concepts were evaluated against the ToC, with refinements made to align them more closely with behavioural objectives and mitigate identified risks. This was done explicitly by writing out a ToC for each potential intervention along with possible risks and then refining the conceptual idea for the concept. For example, the risk of circumventing the need to wash one’s hands to see the code that would be revealed on the soap led to new considerations on how this might be displayed to force agitation of the swatch for the reveal. This was also trialled through immediate feedback on the concept with potential users of the system to understand what resonated with them and why. Thus, the conceptual design of the concept was anchored in the ToC and how it would be demonstrated in the BSC such that features can better align with this thinking and provide a basis for evaluating prototypes.

1. 2. Prototyping:

Figure 9. Prototypes used in fieldwork when getting feedback on concepts including the implementation of the prize concept (a) and understanding how the soap and casing might be placed in a toilet setting (b).

Prototypes of the soap tabs were created and tested with households over both short (1–2 hours) and longer (one week) durations. Feedback was used to refine the product, emphasising simplicity, affordability and effectiveness while amplifying attributes that resonated with users (e.g., portability, cleanliness). Attention was given to each phase of the ToC outlined at the conceptual stage and iteratively developed throughout testing to understand how it might best fit into the setting, for instance (Figure 9b), and how it might contribute to various motives and how the device might cue handwashing and thus reduce the cognitive need to remember to wash hands. Throughout this period, the BSC was used to ground conversations around how the concept was fitting more broadly. For instance, a critical part of this was the role of a new agent – the soap seller – which included replenishing the soap supply but critically was also a reminder about the desired behaviour.

1. 3. Final Testing:

The final version of Tab Soap was tested in 12 households over two months. During this period, the ToC was iteratively updated to reflect user feedback and ensure alignment with the desired behavioural and health outcomes. Refinements focused on simplifying the product and reinforcing its intended use at the point of action.

The Tab Soap intervention was designed to fit seamlessly within the household toilet setting, addressing barriers to handwashing identified during earlier phases. Its ToC (Figure 10) outlines the links between the intervention’s features and subsequent impact. The intervention itself consists of a new prop in the form of a biodegradable, single-use fabric swatch impregnated with soap, specifically designed to discourage alternative uses (e.g., laundry or dishwashing) while remaining portable and easy to place at the point of use. Within the environment, Tab Soap is positioned for immediate accessibility where handwashing typically occurs. Its novel design acts as a visual reminder and reinforces handwashing with soap at the critical moment. It also motivates use by leveraging novelty and aspirational status and alleviates disgust through its single-use format. The environment also involves the periodic presence of the soap seller who replenishes the soap and acts to remind of and reinforce compliance with the intended behaviour of handwashing. Behaviourally, Tab Soap supports hand washing with a dedicated product that simplifies the action and ensures its intended purpose is clear. The ultimate impact is thought to be lower rates of illness and absenteeism based on the rationale gathered from other research.

Figure 10. The ToC for Tab Soap’s development and evaluation in Tanzania.

The ToC also played a critical role in guiding the design, prototyping and testing of Tab Soap. During prototyping, the ToC highlighted specific areas of focus to ensure alignment with the desired behavioural outcomes. For example, it emphasised the importance of testing the placement of Tab Soap within the environment to verify its accessibility and visibility at the point of action. Additionally, the ToC helped prioritise attributes like portability, single-use design and novelty, which were amplified during refinement to maximise their behavioural impact. Risks identified during the earlier stages, such as potential misuse or low adoption, were evaluated during testing. For instance, the team tested whether the portability of Tab Soap allowed it to be conveniently stored and carried without disrupting its functionality, ensuring that its placement in the household aligned with user needs and behaviours.

By structuring prototyping decisions around the ToC and BSC, the team was able to iteratively refine the intervention, ensuring it effectively addressed barriers to handwashing and integrated seamlessly into the behaviour setting.

5.5. Results and learnings

A small-scale, month-long trial showed strong adoption and consistent use of Tab Soap across all participating households, validating its effectiveness as an intervention in the specific context of the study. Participants highlighted the convenience of the product, its single-use design and its aspirational status as key drivers of adoption. Importantly, the intervention succeeded in addressing the barriers to handwashing identified during earlier phases, with the final design fully aligned with the behaviour setting and user needs.

However, the intervention’s limitations must also be acknowledged. The study was a small-scale pilot conducted with a limited number of households, and while the results indicate potential for broader success, the question of repeat purchase remains unresolved. This reinforces the idea that SDD helps to focus on a specific setting but other factors are at play to make a more holistic offering. Specifically, it is unclear whether households would consistently purchase Tab Soap once external support is removed, particularly in resource-constrained contexts where competing priorities for limited funds may reduce demand over time. This is a critical factor for scaling the intervention sustainably and requires further investigation in larger studies.

Despite these limitations, the study provided valuable insights, including a refined ToC and situated intervention within the context of a BSC that offers a structured understanding of why and how the intervention works. This ToC provides a basis for assessing whether Tab Soap – or similar interventions – might succeed in other contexts. While the unique dynamics of a particular behaviour setting always need to be considered, the ToC offers a flexible framework for evaluating the alignment between the intervention’s mechanisms and the characteristics of new settings. In particular, it helps understand which specific features of the intervention link to behavioural mechanisms. This includes the product itself (e.g., the way the design constrains performance to the desired agitation of soap rather than bathing or alternative uses, the novelty of the interaction leading to status), how it was situated in the context (e.g., the portability allowing placement, the tabs themselves being carried) and wider aspects of the intervention such as the soap seller’s role in reinforcing desired behavioural performance.

In summary, the Tab Soap pilot demonstrated the utility of the SDD approach in designing contextually grounded interventions and highlighted the importance of iterative prototyping and user feedback. The insights gained from this process, particularly the ToC and the canvas, provide a foundation for refining and scaling the intervention while also informing its potential applicability in other settings. This complements a range of other studies looking at design for low-resource settings and contextual consideration more generally that have been referenced in the introduction.

SDD has been applied across a diverse range of contexts with variability in technological focus, team expertise and target output. Depending on the project, the process may focus solely on descriptive analysis or extend through to the development and testing of interventions. In all cases, the BSC serves as a critical tool for summarising findings and guiding decision-making. To support users in understanding the application of the canvas, a number of completed examples are provided in Appendix A. These include settings such as participation in a twitch stream (Figure A2) and engagement in a micro-venue at a conference meant to help people form connections (Figure A3). These additional examples offer practical insights into the flexibility and utility of the BSC in varied design challenges. The appendix also includes a blank canvas (Figure A1) that can be downloaded and used in behavioural design work.