By Jason Hreha

Glossary

The BCT Taxonomy: 93 Behavior Change Techniques, Explained

Reviewed by Jason Hreha

Updated Published 7 min read

What Is the BCT Taxonomy?

The BCT Taxonomy (Behaviour Change Technique Taxonomy version 1, or BCTTv1) is a standardized classification of 93 distinct behavior change techniques organized into 16 groupings. It provides a common language for describing, coding, and comparing the "active ingredients" of behavior change interventions. Developed by a multi-institutional team led by Susan Michie, including Michelle Richardson, Marie Johnston, Charles Abraham, Jill Francis, Wendy Hardeman, Martin Eccles, James Cane, and Caroline Wood, the taxonomy was published in 2013 in Annals of Behavioral Medicine.

The taxonomy answers a practical question: "What exactly did the intervention do?" Before BCTTv1, two researchers describing the "same" intervention might mean completely different things. One study's "goal setting" might be another's "action planning." The taxonomy assigns each technique a precise definition, a numerical label (e.g., 1.1, 2.3), and clear criteria for distinguishing it from similar techniques.

BCTTv1 classifies intervention content. It can be used alongside the Behaviour Change Wheel, but it is not itself a theory of why behavior changes or a ranking of effective treatments.

How the Taxonomy Was Built

The developers used a Delphi process: experts reviewed and refined the definitions over successive rounds.

  1. Step 1: Compilation. The research team compiled candidate techniques from existing published taxonomies, including Abraham and Michie's earlier 26-technique taxonomy (2008, Health Psychology), the CALO-RE taxonomy (Michie et al., 2011), textbooks, and intervention manuals. This produced an initial list of 124 candidate techniques.
  2. Step 2: Expert Delphi exercise. An international panel of behavior change experts participated in iterative rounds of rating and refinement. Experts rated each technique for clarity, distinctiveness, and whether it belonged in the taxonomy.
  3. Step 3: Sorting and clustering. Experts independently sorted the techniques into groups. Hierarchical cluster analysis informed the 16 groupings.
  4. Step 4: Reliability testing. Trained coders independently coded intervention descriptions to test whether they could reliably identify the same techniques. Inter-rater reliability was assessed and reported.

The result was a set of 93 techniques. Each received a standardized name, a numerical label, a precise definition with examples, and clear boundary criteria.

The 16 Groupings

1. Goals and Planning

Includes goal setting for behavior (1.1), problem solving (1.2), goal setting for outcomes (1.3), action planning (1.4), and behavioral contract (1.8). Action planning (1.4) includes implementation intentions but is broader: a plan can specify context, frequency, duration or intensity without using an if-then format.

2. Feedback and Monitoring

Includes self-monitoring of behavior (2.3) and feedback on behavior (2.2). These labels describe observing or recording behavior and providing information about it.

3. Social Support

Three variants: unspecified (3.1), practical (3.2), and emotional (3.3). Advise on, arrange, or provide social support for the target behavior.

4. Shaping Knowledge

Includes instruction on how to perform the behavior (4.1) and information about antecedents (4.2). These address the "I don't know how" barrier.

5. Natural Consequences

Includes information about health consequences (5.1), salience of consequences (5.2), and anticipated regret (5.5). These techniques make the costs and benefits of the behavior psychologically vivid.

6. Comparison of Behavior

Includes demonstration of the behavior (6.1, modeling), social comparison (6.2), and information about others' approval (6.3). Social comparison draws on normative influence. Demonstration draws on Bandura's observational learning.

7. Associations

Includes prompts/cues (7.1), the technique of introducing environmental or social stimuli to trigger the behavior. Also includes cue signaling reward (7.2) and associative learning (7.8).

8. Repetition and Substitution

Includes behavioral practice/rehearsal (8.1), behavior substitution (8.2), habit formation (8.3), and graded tasks (8.7). Habit formation (8.3) specifically prompts rehearsal and repetition in the same context to build automaticity.

9. Comparison of Outcomes

Includes credible source (9.1) and pros and cons (9.2). These help people evaluate reasons for and against changing their behavior.

10. Reward and Threat

The largest grouping with 11 techniques, covering material incentives (10.1), material rewards (10.2), social rewards (10.4), and self-rewards (10.9). The taxonomy carefully distinguishes between incentives (promises of future reward) and rewards (delivered after the behavior).

11. Regulation

Includes pharmacological support (11.1) and conserving mental resources (11.3). These address physiological and cognitive barriers.

12. Antecedents

Includes restructuring the physical environment (12.1) and restructuring the social environment (12.2). These overlap conceptually with choice architecture. Also includes adding objects to the environment (12.5), which could describe adding relevant equipment or supplies, depending on the intervention details.

13. Identity

Includes framing/reframing (13.2) and valued self-identity (13.4). A broad identity-focused approach should not be assigned a code without checking the actual technique definition.

14. Scheduled Consequences

Operant conditioning techniques: behavior cost (14.1), punishment (14.2), remove reward (14.3), and various reinforcement schedules.

15. Self-Belief

Includes verbal persuasion about capability (15.1) and focus on past success (15.3). These address beliefs about capability, including self-efficacy.

16. Covert Learning

Includes imaginary punishment (16.1), imaginary reward (16.2), and vicarious consequences (16.3). This grouping concerns imagined or observed consequences. Mental rehearsal fits here only when it matches one of these definitions.

How BCTs Connect to COM-B and the Behaviour Change Wheel

COM-B helps organize questions about capability, opportunity and motivation. The Behaviour Change Wheel adds intervention functions and policy categories. BCTTv1 can then describe the specific content selected.

Connell and colleagues’ 2019 expert study examined 61 techniques and 26 mechanisms of action. Expert agreement identifies hypotheses to test; it does not prove that a technique operates through a particular mechanism in every intervention. The later Theory and Techniques Tool combines evidence from several studies.

Which BCTs Work Best?

Michie and colleagues’ 2009 meta-regression, which used an earlier 26-technique classification rather than the later BCTTv1, covered 122 evaluations of healthy eating and physical activity interventions. Programs combining self-monitoring with at least one other control-theory technique had larger average effects than other programs (standardized mean differences of 0.42 versus 0.26, with Hedges’ correction for small samples). This comparison was between studies and does not isolate self-monitoring’s causal effect.

There is no universal best technique or combination. Select techniques using evidence relevant to the behavior, population and delivery.

How to Use the BCT Taxonomy

Start with the outcome you want and choose a behavior that can help achieve it. Check whether that behavior suits the people and circumstances involved. A catalogue of techniques cannot make that choice for you.

Use the official definitions to code intervention descriptions. For example, goal setting (behavior) requires evidence that a behavioral goal was set or agreed; a generic health message is not enough.

The original reliability test found adjusted kappa of at least 0.60 for 23 of the 26 techniques that occurred at least five times. This does not establish equally good coding reliability for all 93 techniques. Training, clear descriptions and checks between coders matter.

A BCT label does not fully specify dose, delivery, tailoring, context or fidelity. These should be reported separately. Classification helps comparison; an intervention still needs evaluation.

A worked example: coding a handoff intervention

Suppose a team wants each outgoing worker to leave a useful update for the next shift. The intervention includes agreeing that the worker will write the next action before leaving, deciding where and when to write it, and reviewing obstacles such as missing access to the task system. This is a hypothetical example, not a tested program.

The agreed action can meet 1.1, goal setting (behavior). Specifying the situation in which it will happen can meet 1.4, action planning. Analyzing what prevents the update and selecting ways to address those obstacles can meet 1.2, problem solving. Use the official definitions to check the details. “Improve communication” alone would not establish that any of these techniques occurred.

Now imagine replacing the meeting with an email that says only “Please remember your handoff.” It may serve as a prompt, but it does not document the same goal agreement, planning or obstacle analysis. Calling both interventions “handoff support” would hide a material difference.

Record who delivers the content, how often, and whether workers actually receive it. Then assess whether the next shift can act on the update. The taxonomy can distinguish the intervention's ingredients; the research-methods guide helps choose a comparison that can assess their contribution.

Real-World Applications

The healthy-eating and physical-activity review shows how technique coding can help compare complex programs. In a new project, the same method can make an intervention description clearer: was a goal agreed, did someone record their behavior, and what feedback did they receive?

For smoking cessation or pandemic measures, the labels can describe what a program contains. Its outcomes must still come from evidence about that program. A shared code does not make two interventions equally effective.

Limitations and Criticisms

The taxonomy tells you what an intervention contains. It does not tell you which behavior to choose, which technique will work best, or how well it was delivered. A coach's conversation and a website form can share a code while offering very different experiences. Record those differences instead of treating the code as the whole intervention.

BCT Taxonomy vs. Other Frameworks

  • BCT Taxonomy: What specific techniques does the intervention contain?
  • COM-B: What capability, opportunity, and motivation factors are relevant?
  • Behaviour Change Wheel: Which intervention functions and policy options are being considered?
  • EAST: Could the design be easier, more attractive, more social, or better timed?
  • Fogg Behavior Model: How do motivation, ability, and a prompt come together for this behavior?

These questions can complement each other.

Later Developments

The Behaviour Change Technique Ontology was published in 2023. That paper described 281 techniques within 20 higher-level groups. It extends BCTTv1 through a formal, machine-readable structure. Cite the version used rather than treating the ontology as a synonym for the original 93-technique taxonomy.

Frequently Asked Questions

What is the BCT Taxonomy?

BCTTv1 is a shared classification of 93 behavior change techniques in 16 groupings. Its main purpose is to make intervention content easier to describe and compare.

What are the most effective behavior change techniques?

There is no universal winner. Self-monitoring combined with another control-theory technique was associated with larger effects in the 2009 healthy-eating and physical-activity review. That does not establish the best choice for every project.

How does the BCT Taxonomy relate to COM-B?

COM-B organizes questions about behavior. The taxonomy names specific techniques that a designer might select after investigating those questions.

Is there a BCT Taxonomy version 2?

The 2023 Behaviour Change Technique Ontology is a later development with a larger classification. Use its published name and version rather than treating it as interchangeable with BCTTv1.

Why it matters

Before asking whether an intervention worked, we need to know what it did. The BCT Taxonomy makes that description more precise. It supports better questions; it does not supply the answers about effectiveness.

Sources and further reading