Social psychologist Wendy Wood published Good Habits, Bad Habits: The Science of Making Positive Changes That Stick with Farrar, Straus and Giroux in 2019. The book draws on her habit research to challenge a familiar story: people do not carry out every repeated behavior by consciously deciding again. In stable contexts, practiced responses can be cued with little deliberation.
This does not make intention irrelevant. Intentions help select a direction, arrange a context, and begin repetition. They are often weak controllers in the exact moment when fatigue, distraction, or an established cue favors the old response. Wood's practical shift is from repeatedly demanding motivation to designing conditions in which the desired response becomes easier to repeat.
The book is closer to an evidence synthesis than many habit titles, but it remains a trade book, much of the research comes from Wood and colleagues, and laboratory concepts do not map perfectly onto every life. Use its mechanisms as testable models rather than neural certainties.
Habit is not the same as frequency
A behavior can occur often because it is scheduled, socially required, or consciously valued. Researchers typically use “habit” more narrowly: a learned cue-response relation in which encountering a context increases the impulse to perform a response automatically.
That distinction matters. Reading every evening may be frequent but still require a deliberate choice. Unlocking a phone upon feeling a vibration may be strongly cued. The goal is not merely to count repetitions; it is to ask what starts the behavior and how much deliberation it requires.
Choose one target and observe it for seven days. Record:
- the location, time, preceding action, people present, and emotional state;
- the first visible behavior in the sequence;
- the immediate consequence;
- exceptions, especially when the usual context changed.
Patterns reveal potential cues. Do not assume that a polished “cue-routine-reward” story is correct after one observation. The site's guide to what the habit loop simplifies explains why the popular three-part loop is a useful prompt rather than a complete theory.
Build stable context before demanding automaticity
For a desired habit, specify a repeatable context and a small response: “After placing my breakfast plate in the sink, I will take the medication prescribed for that time,” or “When I sit at the desk after lunch, I will open the language exercise and complete one item.”
The cue should be observable and likely to occur. “When motivated” is not stable. The response should be small enough to repeat under ordinary conditions but still represent the intended behavior. Preparing exercise clothes can support exercise; it is not exercise itself.
Reduce friction by placing needed materials where the cue occurs, pre-deciding the first step, and removing unnecessary choices. Increase friction for an unwanted response by adding distance, delay, or a deliberate checkpoint. The site's friction-design guide offers examples without pretending that inconvenience alone resolves dependence or compulsion.
Repeat under a similar context and record completion. If the setting changes constantly, use a portable cue such as a preceding routine rather than a clock time tied to one place.
Rewards teach; they do not need to be prizes
Wood emphasizes that rewarded responses are more likely to be repeated and encoded in habit learning. “Reward” here can be immediate satisfaction, relief, social connection, sensory pleasure, or successful completion. A delayed abstract benefit may be too remote to shape the moment.
Do not infer that every behavior needs a purchased treat. Pairing action with an enjoyable element, making progress visible, or choosing a response that feels intrinsically workable may be enough. External rewards can also displace the purpose or become the only reason to continue.
For an unwanted habit, identify the immediate payoff before prescribing a substitute. Doomscrolling may offer relief from an ambiguous task; snacking may interrupt boredom; overchecking may briefly reduce uncertainty. A replacement must address some of that function or the old response remains attractive.
There is no universal formation deadline
Popular challenges promise a habit in 21, 30, or 66 days. The evidence does not support one deadline for every behavior and person. A recent systematic review of health-related habit formation found wide variation across behaviors and individuals, with limited and heterogeneous studies. Context stability, repetition, enjoyment, planning, and the behavior itself all mattered.
The right question is not “Which day will this become automatic?” It is “Is the cue-response connection strengthening enough to reduce effort and survive normal variation?” The site's page on the 21-day habit myth is a useful guard against false precision.
Measure three things weekly: how often the cue occurred, how often the response followed, and how mentally effortful starting felt. Self-report is imperfect, but the pattern is more informative than a ceremonial streak. Missing one opportunity is not a moral failure. Resume at the next valid cue and investigate repeated misses.
Bad habits are not bad people
The same learning process can support beneficial or harmful behavior. Moralizing a cue-response pattern adds shame without changing its conditions. Yet environmental design is not unlimited. Addiction, eating disorders, obsessive-compulsive symptoms, trauma responses, and some self-harming behaviors cannot be responsibly reduced to moving an object or adding friction.
For medication, nutrition, sleep, substance use, or mental-health targets, behavior design should respect professional guidance and physical safety. Sudden changes can carry risk. If a pattern is dangerous, compulsive, or causing substantial impairment, seek qualified help rather than interpreting difficulty as inadequate habit technique.
Resources also shape context. Advice to redesign a kitchen, commute, schedule, or neighborhood may assume money and control that many people do not have. Look for the smallest available lever while acknowledging the constraint. Shared environments require consent; one person's friction can become another person's barrier.
A four-week context experiment
In week one, observe without changing the system. In week two, choose one cue and reduce one source of friction. In week three, protect the immediate payoff or make progress visible. In week four, introduce one planned disruption—a different location or busy day—and see whether the response survives or needs a portable cue.
Review:
- Did behavior follow the intended cue more often?
- Did starting require less conscious negotiation?
- Did the design create costs for anyone else?
- Is the target still valuable?
- Which feature should remain after the experiment?
Good Habits, Bad Habits is strongest when it removes moral drama from repetition. It does not say that people are robots or that environments determine everything. It shows why conscious intention needs material support. Choose a cue, arrange the first step, repeat in a stable context, and let observed behavior—not a motivational slogan—tell you whether the design works.
Sources
Authorship, subtitle, Farrar, Straus and Giroux imprint, copyright, and 2019 publication were checked on Macmillan's official book page. The definitions and practical limits were bounded with a systematic review of habit concepts in health behavior and the 2024 systematic review and meta-analysis of habit-formation time. Those reviews support context-dependent repetition and wide variability; they do not make every trade-book example a tested intervention.