Why Good Intentions Need a Trigger

Implementation intentions—usually called if–then plans—are the subject of this series. The question is not whether a sentence beginning with “if” can control a life, but whether linking a recognizable situation to an actionable response helps a person carry a goal into behavior. By the end of this series, readers should be able to tell a genuine cue-linked plan from a vague intention, judge what the evidence actually shows, and decide when the strategy is worth trying without treating it as magic.

The gap is at the moment of action

Most behavior-change efforts begin with a reasonable statement of intent: eat more vegetables, drink less, attend lectures, exercise regularly, or remember to make an appointment. Such a statement matters. It expresses motivation and gives a person a direction. But it is made at one time, often when the person is calm and thinking deliberately, while the behavior must be carried out later, amid hunger, fatigue, deadlines, social pressure, habitual routines, and competing options.

Consider someone who decides to eat more vegetables during a nutrition conversation. The decision may be sincere. Several hours later, that person is hungry at work, the nearest option is familiar and convenient, and the earlier motivation is no longer especially vivid. The old choice does not necessarily win because the person stopped caring about health. It may win because the situation has activated a well-rehearsed routine before the goal has been retrieved in a useful form.

This is the intention–behavior gap: the imperfect relationship between what people plan or want to do and what they subsequently do. The gap is not evidence that intentions are useless. It identifies a missing part of the problem. A goal tells someone what outcome matters; it does not always specify how to recognize the decisive moment or what response to make when that moment arrives.

The same pattern appears in less obviously health-related settings. A student may intend to attend every lecture, then face a deadline, bad weather, travel time, or the temptation to stay home. Someone trying to reduce drinking may want to decline another drink, but the relevant moment may involve an offer from friends, a round already being ordered, or a familiar setting in which drinking is expected. The obstacle is not always a lack of information. Often it is the transition from a general commitment to a concrete response under pressure.

That distinction changes what an intervention should try to do. A motivational message may strengthen the goal at the time it is delivered. An implementation intention is designed for the later point of contact with a cue. It asks, in effect: when this recognizable situation occurs, what will I do next?

flowchart LR A["Goal intention: I want to change"] --> B["Everyday critical situation: a cue appears"] B --> C{"Is the cue noticed?"} C -->|Yes| D["Then: actionable, goal-directed response"] D --> E["Enacted behavior"] C -->|No| F["Old habit or missed opportunity"] B --> G["External constraint or competing social cue"] G --> F P["If–then plan: proposed bridge, not a guarantee"] -. "links" .-> B P -. "links" .-> D

The diagram’s fork is important. An if–then plan is proposed as a bridge between the goal and the critical situation; it does not remove the competing habit, create a resource that is absent, or guarantee that a response will be noticed and carried out. Its more limited ambition is also its analytical advantage: it targets a particular moment rather than promising a total change in character.

An if–then plan adds a condition

An implementation intention has a different structure from a goal intention. “I intend to eat more fruit and vegetables” names a desired direction. “If I feel hungry in the office, then I will go to the grocery store and buy a salad” links a situation to a response. The first sentence says what the person wants. The second specifies when the plan should become relevant and what action should follow.

Three features make that link meaningful. First, the if-part must describe a perceivable cue or critical situation. Hunger in the office, being offered an alcoholic drink, or feeling tempted to miss a lecture can in principle be recognized when it happens. A cue that is too abstract—“when I need to make better choices”—offers little help at the moment of decision. The plan needs an event, feeling, setting, or opportunity that can function as an anchor.

Second, the then-part must be a directly actionable, goal-directed response. “I will live more healthily” is another goal, not a next action. “I will ask for a non-alcoholic drink” or “I will remind myself why attending this lecture matters and leave for campus” is closer to a response, although its usefulness still depends on whether the person can actually do it in that situation. A response can be precise without being feasible; specificity alone is not a substitute for time, money, transport, social support, or control.

Third, the two parts must be explicitly linked. The point is not simply to write a list of situations and a separate list of desirable actions. The person forms a conditional relation: if this situation, then that response. The strict evidence reviewed in the dietary meta-analysis treats the link as essential because the proposed mechanism depends on a cue being connected to a response. If one component is missing, the intervention may still provide information or encouragement, but it is testing a different form of planning.

This is why a calendar reminder is not automatically an if–then plan. “Go for a run on Thursday at 6 p.m.” can be useful, but it is a time-based instruction. It may not identify what the person should do when the relevant moment is disrupted, nor does it necessarily connect a perceivable cue to an immediate response. A broad “when, where, and how” plan can overlap with implementation intentions, but the categories should not be collapsed merely because both involve planning.

The distinction is more than terminology. It tells us what to look for in evidence. If an intervention labeled “if–then planning” contains only general advice, a scheduled activity, or an unlinked collection of options, a positive outcome cannot be cleanly attributed to the cue–response mechanism. Conversely, a plan can meet the formal definition and still fail because its cue is rarely encountered, its response is impractical, or the behavior is controlled by forces outside the person’s reach.

The theory behind the strategy proposes that a formed link can make the critical cue easier to detect and the associated response easier to initiate. That is a claim about how behavior may be supported, not a direct observation that every plan becomes an automatic habit. The evidence in this series will therefore keep two questions separate: did the behavior change, and was the proposed process itself demonstrated?

“Works” depends on what is being measured

The available studies make the subject useful precisely because they do not produce one undifferentiated verdict. They examine different behaviors, populations, comparison groups, delivery formats, and measures. Each changes the meaning of success.

The most tightly focused evidence comes from a 2026 systematic review and meta-analysis of strict if–then planning for adult fruit and vegetable consumption. The authors screened studies not merely because they used the label “implementation intentions,” but because their instructions contained a perceivable situation, a directly actionable response, and an explicit link. They included ten articles containing twelve comparisons and 2,399 participants, with follow-ups ranging from one week to 24 months. Compared with active controls, the planning groups reported consuming about 0.29 more portions of fruit and vegetables per day. The standardized effect was small, at 0.20.

That result is neither trivial nor transformative. It suggests that a low-resource planning component can shift a reported dietary behavior by roughly a quarter of a portion per day on average. It does not show that people reliably reach dietary recommendations, that every person benefits, or that the change is large enough to produce a clinically meaningful reduction in disease risk. All of the included dietary measures were retrospective self-reports, and the review notes that the participants were predominantly Western European, female, Caucasian, and young-to-middle-aged adults. Even the prediction interval allowed a future study to find a null effect.

The alcohol evidence makes the outcome problem sharper. A separate meta-analysis of sixteen studies found a small reduction in weekly alcohol consumption when people formed implementation intentions: the average standardized effect was −0.14. But the effect on heavy episodic drinking was essentially null, at −0.01. Reducing a total number of drinks across a week is not the same behavioral target as preventing a concentrated episode of heavy drinking. Social settings, cultural expectations, and the person’s control over the situation may be especially important in the latter case.

The alcohol review also found signals that effects were larger in community samples than in university samples, with paper delivery than online delivery, and with volitional help sheets than traditional implementation-intention formats. Those findings are clues, not a clean ranking. Delivery mode was entangled with sample type, and the format comparison included far fewer help-sheet samples. Longer follow-up was associated with smaller effects. A plan that helps in the first few weeks may need renewed attention, adaptation, or a different intervention if the relevant context changes.

Finally, a randomized trial provides a different kind of evidence. At one Scottish university, 252 psychology students were assigned to a volitional help sheet, the same help sheet plus an encoding-facilitation task, or an active control. Attendance was recorded over an eleven-week semester rather than recalled later. The control group attended about 64% of lectures, compared with about 76% for the help-sheet group and 74% for the group that also rehearsed and restated the plans. The help sheet improved attendance in this setting, but extra rehearsal did not add a detectable benefit, and memory ability did not change the size of the help-sheet effect.

These findings can coexist. They are not competing answers to one simple question. They show why “does it work?” must be completed by several others: for which behavior, under which conditions, against what comparison, measured how, and for how long? An if–then plan is best understood as a targeted way to improve the odds of a goal-directed response at a recognizable moment. The next question is what, exactly, is supposed to happen between forming that conditional link and encountering the cue later.

References