What Is the AACTT Framework?

AACTT is a framework for specifying behavior that Justin Presseau and colleagues introduced in Implementation Science in 2019. The five letters stand for Action, Actor, Context, Target, and Time. The framework describes an observable action, the person or group that performs it, and the conditions in which it takes place. It also identifies the people the action is directed toward, carried out with, or done for.

The authors build on Fishbein and Ajzen’s (2010) TACT framework, which grew out of social psychology research on predicting behavior. One of its principles is to measure intentions and actions at the same level of specificity. Presseau and colleagues make the actor explicit for situations in which one person acts with or for someone else. That distinction matters at work, where the people preparing a decision, making it, and benefiting from it often have different roles.

How Do You Define a Target Behavior with AACTT?

Back to billing. The team picks one action to study on projects where the client must approve the month’s billable work. For each eligible project, the project manager sends the client the monthly billing summary through the portal for approval by the first business day after the month-end close. The wording specifies what the project manager is expected to do and when the team has an opportunity to observe it.

Element Question to Ask Billing Example
Action · Observable action What exactly should we be able to see someone do? Send the monthly billing summary for approval.
Actor · Person or group Who takes the action? The project manager responsible for the account.
Context · Setting and circumstances In what setting and under what conditions does it happen? In the portal, for projects where the client must approve the month’s billable work.
Target · Recipient or other party Who is the action directed toward, carried out with, or done for? The client who receives the summary for approval.
Time · Timing When should the action happen? By the first business day after the month-end close.

In that sentence, the Target is the client who receives the summary. The target behavior is the project manager sending it under the conditions specified. The business outcome is fewer late invoices. Whether earlier delivery helps achieve that outcome is a hypothesis. To track the action, the team divides the number of summaries sent on time by the total number due for the period. If summaries go out on time but approvals still come in late, the next step is to look at what happens after clients receive them.

The right level of detail depends on the decision at hand. The exact time of day adds little if the portal accepts submissions all day and the deadline is set by date. It matters if sending after a cutoff pushes processing to the next day. Keep the detail that changes how the team recognizes the action or makes sense of its absence.

When Should You Specify More Than One Behavior?

Suppose a hiring manager sends their decision to a recruiter after a final interview. The recruiter then gives the candidate the news. The recruiter is the recipient of the first action and the actor in the second. A single measure of how long candidates wait shows the total delay. Timestamps for the two handoffs show where it occurs.

Once the team tracks the actions separately, it may find that the recruiter responds quickly as soon as the decision arrives. It could also find that the recruiter has the decision and is waiting for approval to extend an offer. Where the delay occurs determines whose work the team needs to look at next.

How Do You Go from Specifying Behavior to Diagnosing It with COM-B?

A behavior specified with AACTT gives the team a clear question to investigate. Someone may take the expected action when things are running smoothly and put it off under pressure. The COM-B model, introduced by Michie, van Stralen, and West (2011), helps the team look at what changes between those situations. Can the person still access the information they need? Does another task take priority? Comparing instances in which people acted with those in which they put off the action helps narrow down the possible barriers.

The diagnosis guides the choice of intervention. If people struggle with harder cases, the team can have them practice those cases and give them feedback on their decisions. If most people know what to do and are waiting for another department to grant access, the process owner has a decision to make about permissions. The Behaviour Change Wheel (in French) helps connect the barriers the team finds to relevant intervention functions.

How Do You Measure the Target Behavior?

To build a behavioral indicator (in French) for frequency, first define the occasions when people are expected to act. Count the occasions when they meet the criteria. The total number of actions can go up simply because the volume of work has increased. A rate lets the team compare periods with different workloads, provided it applies the same counting rules.

A timestamp tells you when a record was updated. Before using it as evidence of a check or a handoff, find out what actually triggers the log entry. If the system records those events the same way, the team can review a sample of cases and compare the logs with what people did. Presseau and colleagues (2019) point to this trade-off between measuring the behavior precisely and using data the organization already collects.

Where Does AACTT Fit in the MEET© Protocol?

In MEET©, teams use AACTT starting in Model. They identify an initial behavior based on the business outcome and see whether it holds up when they observe the work. If the person they had in mind lacks the authority to act, or if a different action turns out to matter more, they revise the specification.

In Engineer, the specification gives the intervention a clear behavior to address. In Experiment, it gives the team a consistent way to compare what people do before and during the pilot. In Transfer, the same criteria help the team check whether the practice continues in day-to-day work as pilot support tapers off.

Frequently Asked Questions

Can AACTT be used to describe group behavior?

Yes, when the group takes an action that can be observed at the group level. Two people who jointly approve a decision can be the actor. If one person prepares a recommendation and the other gives final approval, two specifications make each person’s contribution visible.

What if the system leaves no record of the action?

Decide before the pilot which situations to observe and over what period. An observer might track a practice in a sample of meetings, using the same criteria before and during the intervention. The findings describe what happened in the meetings observed. Choose the meetings so those observations can be compared meaningfully.

Does “Time” have to specify a frequency?

A frequency works when the action follows a regular schedule. A trigger or deadline is more useful when the opportunity may arise at any time. “Within two business days of receiving the complete application” establishes when the clock starts and which applications count.

References

Fishbein, M., and Ajzen, I. (2010). Predicting and changing behavior: The reasoned action approach. Psychology Press. 10.4324/9780203838020

Michie, S., van Stralen, M. M., and West, R. (2011). The behaviour change wheel: A new method for characterising and designing behaviour change interventions. Implementation Science, 6, 42. 10.1186/1748-5908-6-42

Presseau, J., McCleary, N., Lorencatto, F., Patey, A. M., Grimshaw, J. M., and Francis, J. J. (2019). Action, actor, context, target, time (AACTT): A framework for specifying behaviour. Implementation Science, 14, 102. 10.1186/s13012-019-0951-x

About the Author

Maxime Rabéchault is the founder and president of UNREST, a change enablement consultancy. He applies behavioral science to organizational transformation and leadership development.