The behaviour change literature has accumulated substantial evidence over the past three decades about what works and what does not. The most consistent finding is that intention-behaviour gap, the distance between knowing what to do and actually doing it, is not bridged by information, motivation, or willpower alone. The interventions with the best evidence work by modifying the environment, reducing friction, creating pre-commitments, leveraging social influence, and building cue-routine-reward structures that eventually become automatic.
Habit Formation Mechanisms
Charles Duhigg popularised the cue-routine-reward loop structure of habits. Research supports this model: behaviours that are consistently performed in the same context, following the same cue, and followed by some form of reward, become automatic more quickly than those performed variably. The practical implication is to identify or create consistent contextual cues for desired health behaviours. Running shoes by the door. A fruit bowl visible on the kitchen counter. A fixed time for meditation on the calendar. The cue does not need to be elaborate; consistency of context is the key variable.
Friction Reduction
Research by Thaler and Sunstein on choice architecture, and by the Behavioural Insights Team on default effects, consistently shows that reducing the friction associated with a desired behaviour is one of the most powerful levers available. Making the healthy option the default, the path of least resistance, removes the need for active choice in each moment. Preparing gym clothes the night before, pre-chopping vegetables for the week, setting a standing appointment for a walk with a friend, and keeping healthy snacks at eye level in the fridge are all friction-reduction interventions with evidence supporting their effectiveness.
Social Commitment Devices
Pre-commitment mechanisms that introduce a social dimension have consistently stronger effects than purely private commitments. Telling someone else what you plan to do, exercising with a partner, joining a class that charges for cancellations, or tracking publicly all increase the cost of not following through beyond the purely internal cost of disappointing yourself. The social dimension works through multiple mechanisms: anticipated accountability activates the threat of social disapproval; shared activity provides co-motivation; and social norms provide guidance for what constitutes adequate effort.
If-Then Planning
Peter Gollwitzer's research on implementation intentions shows that specifying in advance what you will do when a specific situation arises dramatically increases follow-through on intended behaviours. Applied to health: "If it is 6pm on a weekday and I am finishing work, then I will immediately change into exercise clothes before doing anything else" is an implementation intention that performs significantly better than "I intend to exercise more." The if-then structure pre-decides the behaviour, removing the need for in-the-moment decision under conditions of depletion.
Self-Compassion and Relapse Recovery
Research by Kristin Neff and colleagues shows that self-compassion is more effective than self-criticism for recovery from lapses in health behaviour. People who respond to a relapse with self-compassion, acknowledging the difficulty without harsh self-judgement, return to the desired behaviour more quickly and sustainably than those who respond with shame and self-criticism. This is counterintuitive to many people who believe that being hard on themselves is what prevents further lapses. The evidence consistently shows the opposite: harsh self-judgement is associated with higher rates of prolonged relapse, not lower.