How to embed behavioural insights into design teams — Part 1: from problems to insights

Sharing my personal reflections from multidisciplinary design projects in public health. This post continues from How and when to embed behavioural insights into design teams on digital projects.

If a behavioural scientist joins one of our project teams in PHE Digital in an advisory capacity, they’ll typically meet a couple of designers (service, interaction, content etc), a user researcher, probably a product manager, perhaps a delivery manager and, if they’re lucky, a dedicated domain expert (e.g. a cardiovascular disease prevention policy specialist, or an expert in childhood weight management). Here’s what I’ve learned from our collaborations so far:

1. Defining the problem

Consult a behavioural scientist early on if they could contribute existing knowledge of the topic area or if you’re considering a literature review.

→ Sharing relevant past work allows teams to build on things that have worked in the past, or avoid things that have failed. Formats like expert interviews or show-and-tells could work here, upfront or when you commence design research.

→ Literature reviews can help inform your design research by systematically exploring existing evidence about people’s behaviours and what’s been proven to work in the past. They can take time, so consider them early on and how and when you’ll blend in what you learn from them.

Defining the problem starts with the team defining the ‘goal’ of using a service, from a user’s point of view. Through a behavioural insights lens, the ‘goal’ is framed in terms of behaviour, i.e. the outcome an ‘intervention’ is aiming for. Get aligned at the start to inform your design research or later idea-generation. Map their existing experience and evidence of a topic area onto your initial impression of what’s going on, so that you can ask more focused questions about how design could elicit certain behaviours.

A behavioural scientist could review past work and evidence to about:

• User actions: The steps a user needs to take (in Behavioural Insights terminology, the ‘Target behaviours’ that are most conducive to outcome, referencing for example the N.I.C.E. recommendations for a particular health domain)
• Pain points: What hinders them reaching their goal (‘Barriers’)
• High points: What helps them reach their goal (‘Facilitators’)
• Decision-making: Which choices do people make that are habitual (‘Automatic’)? Which are considered and deliberate (‘Reflective’)?
• Contexts: How do people’s environments or social norms influence what they do?

Tip: If at the outset there is not a clear reason to accompany your design research with a literature review, or it isn’t practical to integrate the outcome of it into a project plan, then think twice! Delay it until you have research goals that cannot be addressed solely through design research/user research, or consider if it’s really additive to your project phase. Similarly, if existing knowledge or evidence is functionality or feature-level but your research goals are taking a wider lens, you might be wasting your time involving behavioural scientists at this stage and might get more through later design crits.

2. Planning research

Consult a behavioural scientist if you need to understand the less-conscious reasons that people do the things that they do, or if you need particularly robust recruitment profiles.

Once your research plan is underway, discussing it with a behavioural scientist can help you frame research goals about the less conscious reasons that people do the things that they do, or focus your methods-mix to explore gaps where existing evidence is light.

→ Collaborating on recruitment helps gauge to what extent your plans satisfy the robustness required for the phase of work you are in. Behavioural scientists can help ensure you recruit a representative range participants considering socio-economic, behavioural and cognitive characteristics. This might include recommending standardised scales to include in screeners or in a pre-interview survey (for example, to measure how much knowledge people have about health), which also allows you to gather people’s perceptions to compare what they think and what they say.

→ Sharing your discussion guides means they could contribute questions focused on behavioural patterns of interest to dig into specific aspects of participants’ nuanced habits, beliefs, emotions and how different modes of service delivery fit their lives, or prompt people to talk about less-conscious factors.

The same is true for Expert Interview guides — contributing questions that help identify needs of the target population or knowledge of what works.

Existing behavioural frameworks could inform the whole range of research activities. For example, a behavioural scientist might refer to practical evidence like:

• the COM-B model, to consider whether people have the three necessary components for behaviour change — Capability, Opportunity and Motivation
• the Theoretical Domains Framework, to explore determinants of behaviour — the barriers and need for a digital intervention

We’ve noticed that collaborating with behavioural scientists during research planning helps stakeholders relate to the recruitment profiles and trust the process because the tools are more familiar and have often been used across other types of study.

Tip: In my opinion, the Design Researcher or User Researcher is the ultimate architect of the research plan for a project. Behavioural scientists can offer a wealth of knowledge, so be selective about what you need and what you don’t. If you’re on a tight project timeline, you might choose to skip this stage and involve them during or after synthesis or when characterising users, to use their frameworks as a reflective tool or to raise unanswered questions for a further phase or review of evidence. As in the previous problem-definition stage, if behavioural scientists’ contributions are feature-focused and you’re searching for wider opportunities, their time might be better applied later to critique and refine ideas.

3. Sense-making

Consult a behavioural scientist if their existing knowledge or evidence tells you which users need to do what differently, when, where or how.

If this is the case, your goal is to weave these findings to strengthen your insights, your map of the current experience for users (if you create one), or focus your design opportunities.

User Researchers drive the distillation of learning at this stage to ensure designers are inspired and get what they need from research. In my experience, that’s a minimum of:

1. User journey — a map of people’s current experience, the steps they take between realising their goal and achieving it, how they are interacting and feeling at each step.
2. Insights — statements about people’s challenges, behaviours and motivations that are actionable and drive design.
3. User needs — to describe how to enable users to do the things they need to do, when they need to do them. Perhaps mapped onto #1 as ‘Jobs to be done’. Translated into…
4. Opportunities — concise statements or questions that designers can respond to with ideas.

→ If you included pre-interview questions or want to position what you learned about people in a broad context, involving a behavioural scientist during share-backs could offer a helpful lens on responses to questions asked before or during user interviews. This might also help guard against biases in how you interpreted what you heard.

→ If you want to make your user journey map more comprehensive, involve a behavioural scientist to map evidenced findings alongside your primary research (for example, ‘target behaviours’ against user actions, ‘facilitators’ against highs and ‘barriers’ against lows).

→ If you want to sharpen your design opportunities, you could involve a behavioural scientist to suggest more specific verbs in your opportunity statements. They might refer to behavioural frameworks like the Behaviour Change Wheel (which builds on COM-B mentioned earlier) to recommend the broad types of intervention that are most likely to be successful. These ‘Intervention Functions’ (for example, education, persuasion, incentivisation, training, and enablement, and others) might offer structured starting points for design if inserted into an opportunity statement (i.e. ‘How might we educate/persuade/incentivise… and so on). Else, each opportunity statement could be accompanied by a list of relevant Intervention Functions to act as design cues or ways to categorise ideas (see step 5 on generating ideas)

Tip: Analysing research for design demands very different outputs compared to research for other purposes. In my opinion, behavioural scientists won’t play a key role in interpreting research for design but can be valuable thought partners from the perspective of rigour and comprehensiveness. Behavioural tools like the ‘intervention functions’ are really only helpful when combined with user insight to form design starting points.

4. Characterising users

Consult a behavioural scientist if the most helpful way to group your users relates to their beliefs, attitudes or motivations, and if you’re looking for possible axes or language to describe these factors.

When we use User Type and Persona frameworks to define and understand different groups of users, we aren’t so interested in statistical or demographic generalisations, but in the nuanced differences in people’s beliefs, attitudes, and motivations, their needs, skills and contexts. Since psychology and behavioural science is particularly good at describing these things, collaborating at this point could offer a fast-forward.

→ If you want to experiment with a range of segmentation axes, involve a behavioural scientist to suggest various behavioural frameworks as ways to group people, and work with the team to try applying them (for example, scales to measure capability, or describe types of motivation).

→ If you want a hand describing groups you’ve already found, you could ask a behavioural scientist how they see each group’s characteristics. Then, with your team, check these are actionable (for example by looking for emotive or practical reasons that users have different experiences or needs). The language they use might also help you find suitably memorable names for the groups.

→ If you want to generate ideas specific one group or another, involve a behavioural scientist to suggest a specific type of intervention might be relevant (for example, people with high health knowledge might find positive health choices easy and logical, but could need a surprising amount of help to start with a manageable step over a wildly ambitious goal). Mapping out the differences in how each group of users experiences the service today can help here.

Tip: In my opinion, the template for a typology or persona is determined by a design researcher. Personas for behavioural insights are typically quite different from those for design. Theirs are intended for researchers to inform further research. Ours are often lighter and are made for designers to prompt, provoke and remind them of differing needs.

Continue reading about embedding behavioural insights when you’re generating, prototyping, testing and evaluating ideas in Part 2: from insights to action