Guide
Behavioral Design Process
What is a behavioral design process?
A behavioral design process organizes the investigation, development, and testing of changes intended to influence behavior. The term describes a family of practices rather than one universally agreed sequence.
For my own approach, see the SPARK Method. It starts with the desired outcome, develops several possible behaviors, and chooses among them by examining how well each fits the person or group. That selection comes before efforts to simplify or reward the behavior.
Steps in a behavioral design process
Select the behavior
Begin with the outcome and consider which actions could achieve it. Select a behavior that fits the people and circumstances before designing ways to encourage it.
Investigate the setting
A team needs to identify whose behavior matters, what action is being considered, when and where it occurs, and why the existing situation might impede it. Research should examine the environment and available resources as well as individual beliefs or preferences.
Form a testable proposal
An explicit proposal then connects a design change with a predicted result. For example, clearer document requirements might reduce interrupted applications. That prediction remains to be tested.
Testing, iteration, and evaluation
The MRC–NIHR guidance on complex interventions offers a relevant framework for development, feasibility, evaluation, and implementation, with iteration and attention to context. It is not a standardized definition of every commercial behavioral-design process.
A worked example: completing an application
Imagine a service whose applicants repeatedly leave before submitting the required documents. The goal is a correct, completed application, not simply a longer visit to the website. This is a hypothetical example.
First compare possible actions. Asking people to find every document before opening the form might suit applicants who already know what they need. Saving an unfinished application and adding documents later might suit people who must request records from someone else. Neither choice is automatically best. Watching applicants work through the task can reveal which constraint matters.
Suppose that research identifies uncertainty about acceptable documents. A testable proposal is to place a short list of accepted documents beside the upload control. The prediction is that fewer applicants will stop to ask for help or submit an unsuitable file. If records are expensive to obtain, clearer wording alone will not remove that expense.
Try the prototype with applicants, then choose an evaluation design suited to the decision. A randomized comparison can test whether the changed instructions improve completed, valid applications relative to the existing instructions. Record rejected uploads, support requests and completion time as well. A higher submission rate accompanied by more rejected applications would not solve the original problem.
The research-methods guide explains how to choose between interviews, observation and experiments. They answer different questions within the same process.
Where is behavioral design used?
The same process can frame different practical questions:
- Health: what makes a planned activity or care task feasible?
- Environmental sustainability: what do people encounter when recycling or choosing transport?
- Financial decisions: how do the service, information, and available options affect saving or debt-management actions?
- Education: what helps a learner begin and complete a particular learning task?
Common mistakes
Naming a cognitive bias or selecting a familiar technique does not demonstrate that it explains a local problem. The process should make assumptions visible and compare them with evidence. A change that increases clicks may not improve the user’s eventual outcome, so the chosen measure must match the intended purpose.