Temporal Self-Regulation Theory: S-Tier Behavioral Designer’s Guide

Every behavioral designer I know has watched the same disaster movie a hundred times. A user signs up brimming with intent. They will exercise. They will learn the language. They will quit smoking, save more, study harder, ship the side project. Two weeks later the app is uninstalled, the streak is dead, and the dashboard reads zero. The intention was real. The behavior never followed.

Most designers respond by stacking more motivation on top — bigger rewards, louder badges, redder progress bars. Almost none of it works, because intention is not the bottleneck. The bottleneck is the architecture between intention and behavior, and Hall and Fong’s Temporal Self-Regulation Theory is the most precise map of that architecture ever published.

This pillar walks you through TST the way I teach it to advanced Octalysis students: not as a health-psychology curiosity, but as a load-bearing model for anyone designing a product that asks humans to do something today for a payoff that arrives tomorrow.

Speed Run Notes

• The intention-behavior gap is not a motivation failure. It is a structural mismatch between an intention’s slow future payoff and the immediate, automated cues fighting it.
• Hall & Fong’s TST formula: Behavior = (Intention × Self-Regulatory Capacity) − Behavioral Prepotency. Three knobs, not one. Most products tune the wrong one.
• Self-Regulatory Capacity (SRC) fatigues under load. Stack three willpower checkpoints in a row and a healthy intention will lose to a weak habit by lunchtime.
• Behavioral Prepotency is the dark matter of UX. Every existing habit, default route, and salient cue counts as competing prepotency. You cannot out-motivate it — you must redesign around it.
• Temporal valuation is trainable. Octalysis CD1 (Epic Meaning) and CD4 (Ownership) compress the felt distance to a future outcome, raising its weight inside the intention.
• Design the contest, not just the motive. The Octalysis Apply section maps each TST term onto a specific Core Drive, with one mechanic to raise it and one to lower it.

Author Credibility: Yu-kai Chou

Yu-kai Chou created the Octalysis Framework after studying gamification since 2003 — years before the term entered mainstream vocabulary. As a Human-Systems Architect & Behavioral Designer, his framework has been applied by LEGO, Microsoft, Porsche, Coca-Cola, Salesforce, and MrBeast, impacting over 1.5 Billion Users.

Chou has taught the Octalysis methodology at Harvard, Stanford, Yale, Tesla, Google, BCG, and IDEO.

His work has been cited by Harvard, Stanford, MIT, Forbes, Wall Street Journal, Wired, US Department of Energy, NIST, NSF, NCBI, US Department of Education, ClinicalTrials.gov, and Google Scholar — with 3,700+ more academic publications. Explore his books here.

TEDxLausanne · Gamification to Improve Our World17 min

What Is Temporal Self-Regulation Theory?

Temporal Self-Regulation Theory was introduced by Peter Hall and Geoffrey Fong in a 2007 Health Psychology Review paper that has quietly become one of the most-cited integrative models in behavior-change research. They were trying to solve a specific embarrassment in the field: the Theory of Planned Behavior could predict roughly a third of the variance in intention, but the link from intention to actual behavior kept evaporating. Intent-only models worked beautifully on paper and badly in clinics.

Hall and Fong’s argument is structural. Future-relevant behaviors — eating less sugar today to avoid diabetes in 2046, drafting a chapter today to ship a book next year — happen inside a temporal mismatch. The cost of the behavior is now. The reward is later. Meanwhile, every competing behavior with an immediate payoff is sitting right there, fully practiced, low-friction, neurologically pre-baked.

So they wrote down the contest explicitly: Behavior at time t is a function of three forces: the intention’s strength, the person’s self-regulatory capacity at that moment, and the prepotency of the competing automatic response. The intention does not win because it’s right. It wins when the SRC (Self-Regulatory Capacity) multiplier is big enough to overcome the BP (Behavioral Prepotency) drag.

That single move — turning the intention-behavior gap into an equation with three handles instead of one, is why TST has aged better than the dozens of pure intention models that came before it. It is not a motivation theory. It is a contest model. And the contest is winnable by design.

The Three Components Hall & Fong Actually Specified

This is where casual readings of TST go wrong. People remember “temporal” and assume the theory is about delay discounting. It isn’t. The temporal piece is one input out of three, and the moderators. SRC and BP: are where most of the predictive power lives.

Component 1: Intention (with Temporal Valuation Baked In)

An intention in TST is not just “I want to exercise.” It is a weighted vector of perceived future consequences. Two people can have the identical conscious intention to save for retirement, but one of them weights the 2056 outcome at near-present strength, and the other weights it at roughly zero. Same intent string, completely different behavioral force.

This is where Hall & Fong absorb a chunk of Bickel and Marsch’s delay-discounting literature without saying so loudly. The intention’s magnitude is the product of conscious goal endorsement and the subjective steepness of one’s temporal discount curve. Steepen the curve (make tomorrow feel like next year) and the intention gets crushed before the behavior even contests for it. Flatten the curve (make 2046 feel close, vivid, ownable) and the intention shows up to the fight already heavier.

I labor this point because most behavioral designers treat intention as an input to optimize on the surface — better copy, sharper goal-setting flow, clearer benefit statements. Those help. But they help by raising the conscious goal-endorsement number while leaving the temporal-valuation number untouched. The conscious half is the small half.

Component 2: Self-Regulatory Capacity

SRC is the broad executive-function bundle: working memory, inhibitory control, attentional regulation, the prefrontal capacity to override a prepotent response in favor of a goal-relevant one. In TST, SRC is the multiplier that turns intention from a stored value into actual behavior.

The crucial property of SRC is that it varies. It varies by trait (some people start every day with more), by state (a bad night’s sleep, a tense meeting, an alcoholic drink, hypoglycemia, all drag it down), and by recent load (every prior act of inhibition this hour pulled from the same pool). This is the cleanest place TST connects to the broader self-control literature, including the now-troubled ego-depletion findings, we will get to those when we walk through where the theory holds and where it gets fragile.

The design implication is brutal: an intention that wins at 9 a.m. on a rested Monday can lose at 9 p.m. on a stressed Thursday with no change in the intention itself. If your product asks for behavior during low-SRC windows, it had better not require fresh inhibitory effort to comply.

One subtlety worth flagging early: SRC is not synonymous with willpower in the folk sense. The folk notion is moralized. Some people have it, others lack it, the difference is character. The TST construct is structural: capacity at this moment is what is available to this intention after every prior demand on it has been paid. The same person, identical character, will have radically different SRC across an ordinary day. A behavioral designer who treats SRC as character will keep punishing the user for losing contests the architecture set up to be unwinnable.

Component 3: Behavioral Prepotency

Behavioral Prepotency is everything trying to win the moment without SRC’s help. It is the strength of the competing automatic response, with two sub-components Hall & Fong make explicit: habit strength (how often the alternative has been executed in this context) and cue salience (how loudly the environment is currently shouting for the alternative).

Designers consistently underestimate BP. The vending machine on the office hallway is not a passive object. It is a fully-trained, perfectly-placed prepotent response to mid-afternoon glucose dip, accumulated over hundreds of trials. The healthy-snack drawer two floors away is a fresh intention with no prepotency at all. The contest is not even close, and no amount of intention-side intervention will balance it as long as the architecture sits where it sits.

This is the single most-skipped insight in behavior-change design. We act as if a stronger goal beats a weaker one. TST’s claim — supported by twenty years of follow-up evidence, is that intention loses to prepotency unless either SRC is large enough to override it or the prepotency itself has been engineered down.

One way I like to test whether a designer has absorbed this: ask them what would happen if the user’s intention got 50 percent stronger overnight. Most answers describe a 50 percent better behavioral outcome. The TST answer is more honest. If the user’s SRC is exhausted at the compliance moment, a 50 percent stronger intention buys little. If the competing prepotency is overwhelming, the same is true. Intention strength scales linearly only in environments where the other two terms are not bottlenecks. And those environments are rare in real life.

What Hall & Fong Got Right

Nineteen years after publication, four parts of TST have aged exceptionally well, and any designer using it should understand exactly which load they carry.

1. The Intention-Behavior Gap Is a Multiplicative Problem

Before TST, the dominant framing was additive: more intention plus more skill plus more environmental support equals more behavior. Hall & Fong wrote it multiplicatively, which means a near-zero on any one term collapses the whole product. A perfect intention with no SRC produces no behavior. Great SRC with weak intention does nothing either. Both have to be present at once, and a prepotent habit can swamp both. This explains why interventions that improve only one term so often fail outside the lab.

2. Self-Regulation Is a Capacity, Not a Trait or a Mood

Hall & Fong treated SRC as a fluctuating resource with both stable and dynamic components. That move: pre-empting the entire ego-depletion debate — means TST gracefully absorbs both findings: yes, baseline executive function predicts long-run behavior, AND no, you cannot count on full SRC every moment. Designers who internalize this stop building products that demand willpower at the worst hour of the day.

3. Habit Strength Belongs Inside the Theory, Not as an Afterthought

Earlier intention-behavior models treated habit as noise to be controlled for. TST puts habit strength into the BP term as a load-bearing variable. This is empirically vindicated: meta-analyses of behavior-change interventions consistently find that habit measures account for substantial variance even after intention is controlled. TST predicts this. TPB does not.

4. The Temporal Frame Is Adjustable

The temporal valuation inside the intention term is not a fixed personality parameter. It moves with framing, with mental contrasting, with vivid future visualization, with identity work. This is what makes TST a designable theory rather than a deterministic one. If users discount the future too steeply, you can show them the future. We will return to exactly how in the Octalysis section.

Where TST Falls Apart

No theory survives twenty years of replication unscathed, and TST has three honest weaknesses every practitioner should hold in mind. I list them not to dismiss the model, I will still recommend it as the best contest framework in the literature. But because design built on a partially-broken theory is worse than design built on a theory you understand the limits of.

1. SRC Is a Construct Sandwich

The Self-Regulatory Capacity term aggregates working memory, inhibitory control, attentional regulation, set shifting, and emotion regulation into a single variable. Cognitive-neuroscience research since the 2010s has shown these are dissociable and only weakly correlated. A person with strong inhibitory control can have terrible emotion regulation. A high-working-memory adult can fail a simple delay-of-gratification task on a tired afternoon. Treating SRC as one number is a useful simplification, but it predicts less than the four-or-five components measured separately.

The practical fallout is that “build self-regulation” interventions get mixed results because we are not consistent about which sub-component we are actually training. Mindfulness raises attentional regulation but barely touches inhibitory control. Implementation intentions strengthen prospective memory but do nothing for emotion regulation. If a product’s use case demands inhibitory control specifically, training mindfulness will look like it should work and won’t.

2. Behavioral Prepotency Is Measured Backwards

The standard way to estimate BP in TST studies is to ask people how often they have performed the competing behavior, and how automatic it feels. This is reasonable in survey research but quietly circular when used to predict the same behavior class. The harder problem: objective measurement of cue strength in the user’s actual environment — remains unsolved. Until that’s fixed, BP is partially confounded with whatever the dependent variable already is.

For designers, the takeaway is humility about your prepotency estimates. The environment matters more than the survey says, and the competing prepotency is usually stronger than you think.

3. The Ego-Depletion Anchor Has Cracked

TST originally drew support from Baumeister’s ego-depletion work, which suggested SRC behaves like a fatigable resource at minute-to-hour timescales. The replication crisis has reduced confidence in the strongest depletion effects, especially at short timescales. SRC still varies, sleep, glucose, stress, and time-of-day effects are well-replicated. But the “every act of self-control immediately pulls from a finite pool that drains your next decision” story is much weaker than it looked in 2007.

This does not invalidate TST. SRC variation is real; the question is what drives it. Practical design under post-2020 evidence: treat SRC variation as multi-causal (sleep, mood, context, glucose, prior cognitive load) rather than as a single muscle that gets tired with each rep.

What’s Really Happening Inside the Brain

TST has a beautiful neurobiological correlate, even though Hall & Fong’s 2007 paper was deliberately psychological. The contest the theory describes maps onto a contest the brain runs constantly between two systems, and understanding that mapping makes the design implications click.

The intention side of the equation lives largely in the prefrontal cortex, particularly the dorsolateral and ventromedial regions. These areas encode goal representations, simulate future states, and exert top-down control over downstream action selection. When a future-relevant goal “feels real,” these regions are active and projecting onto motor and reward circuits. When the goal is abstract or distant, the same regions encode it weakly and project less.

The prepotency side runs through the basal ganglia and the dorsal striatum, the brain’s habit machinery. Repeated behavior in a stable context gets compiled into this system as a stimulus-response routine that the prefrontal cortex no longer needs to compute. That is wonderful for tying your shoes; it is terrible for eating one cookie instead of seven if you have eaten seven from this bag every previous night.

The SRC multiplier corresponds to the prefrontal cortex’s real-time ability to override the basal-ganglia output before it becomes a motor action. When SRC is high: rested, fed, calm, low cognitive load — the prefrontal veto wins more often. When SRC is low — tired, hungry, stressed, multitasking — the basal-ganglia routine wins by default.

This is why behavior-change strategies that “make the right thing the path of least resistance” work better than strategies that “make the right thing more appealing.” The latter raises the intention term. The former lowers the BP term directly, and you no longer need a high SRC to win. The brain is not being inspired into compliance; the contest has been rigged in advance.

The dopamine system deserves one paragraph of its own here. Schultz’s reward-prediction-error work showed that dopamine fires not on reward itself but on the gap between predicted and actual reward, and the system learns to fire earlier and earlier as a cue becomes reliable. This is exactly the substrate that compiles a basal-ganglia routine into prepotency. Every time the user reliably gets the cookie when the 3 p.m. slump hits, the dopamine response moves earlier in the chain, the cue grows more salient, and the prepotency tightens. Behavioral designers can use the same mechanism in the desired direction: reliable, well-timed reward on compliance compiles the new behavior into prepotency over weeks, and the SRC tax drops correspondingly. A streak that survives ten days is partially self-sustaining for that reason.

TST vs Other Theories of the Intention-Behavior Gap

TST sits in a crowded neighborhood. Five neighbors are worth comparing it to, because each handles one piece of the same problem differently and most behavioral designers I meet are operating with a confused mash-up of them.

TST vs Theory of Planned Behavior

Ajzen’s Theory of Planned Behavior predicts intention beautifully but is famously weak at predicting behavior from intention. TST is best understood as a TPB-completion theory: keep TPB’s intention-formation machinery (attitudes, subjective norms, perceived control), then add SRC and BP as the moderators that explain why intention so often fails to translate. If you use TPB without TST, you have a complete model of why people want to act and an empty model of why they actually do.

TST vs Implementation Intentions (Gollwitzer)

Gollwitzer’s if-then plans are tactical: they shift prospective memory and pre-link cues to responses, raising the probability that a particular intention gets executed in the moment. TST is the larger architecture inside which implementation intentions are one of the strongest known SRC-supplements. Translated: implementation intentions are a brilliant SRC-augmentation move, but they are still subject to BP if the competing prepotent response is stronger than the if-then link.

TST vs Dual-Process Models (Kahneman, Strack & Deutsch)

Dual-process accounts split cognition into System 1 (fast, automatic) and System 2 (slow, deliberate). TST is what you get when you treat that distinction seriously and ask: under what conditions does System 2 actually override System 1 in this moment? The intention term is a System 2 quantity. BP is a System 1 quantity. SRC is the bandwidth of the System 2 channel. TST and dual-process theory are talking about the same machine from different angles.

TST vs Habit Loop / Cue-Routine-Reward (Duhigg / Wood)

The habit literature, particularly Wendy Wood’s recent synthesis, focuses on how to build prepotent responses for the behavior you want. TST is the contest model that explains why your existing prepotencies keep beating your new intention even after months of effort: and why habit work is so high-leverage. Habit design lowers BP for the desired behavior and raises BP for the undesired one. Done correctly, the contest tilts even when SRC is low.

TST vs Control Theory (Carver & Scheier)

Control Theory frames self-regulation as a feedback loop: reference value, current state, comparator, output. TST is, in a sense, a forward-time loaded version of Control Theory in which the reference value sits far in the future and the comparator has to do extra work to stay calibrated. If you have read the Control Theory pillar and want the temporal-mismatch corner of the same problem, TST is the right next stop.

TST vs Health Action Process Approach (HAPA)

Schwarzer’s HAPA splits behavior change into a Motivational phase and a Volitional phase, separated by action planning and coping planning. TST and HAPA agree about the architecture of the gap but emphasize different mechanisms inside it. HAPA stresses planning (the if-then content) and self-efficacy (the perceived capability). TST stresses the contest itself (intention vs prepotency, moderated by capacity). Used together, HAPA tells you what to build inside the user’s head; TST tells you what to build outside it. Most TST-informed designs are quietly using HAPA-style planning to raise the SRC term, then using environmental redesign to lower the BP term.

TST vs Temporal Motivation Theory (Steel & König)

Temporal Motivation Theory is procrastination’s preferred mathematical model: motivation = (expectancy × value) / (impulsiveness × delay). TMT is brilliant on the discount-curve mechanics but treats the user as a single decision-maker against time. TST adds the prepotency contest TMT abstracts away. Together they form a complete picture: TMT tells you how steep the discount curve is, TST tells you who is winning the moment-by-moment contest under that curve. If you have read the Temporal Motivation Theory pillar, TST is the structural companion you want.

TST in the Real World

TST is academic in tone but the predictions are stark on the ground. Four domains show what it actually buys you when you use it as a design lens rather than a research label.

Fitness & Apple Watch Rings

The standout TST-shaped product is the Apple Watch Activity rings. Apple did not cite Hall & Fong, but the system reads like a clinic in TST application. The Move ring compresses the temporal frame of cardiovascular health from decades to a single day, which is a direct upward push on the intention’s temporal valuation. The wrist-tap and ring graphic are designed-in prepotencies for movement: a low-cost, high-salience cue that competes with the prepotencies trying to keep you on the couch. The seven-day streak adds Core Drive 4 (Ownership) and Core Drive 8 (Loss) layers that protect the behavior on low-SRC days when the intention alone would lose. Strava’s segments work similarly with CD5 (Social Influence) added as a third SRC-substitute. Old Foursquare badges were doing the same thing with primitive tools in 2010.

Language Learning & Duolingo

Duolingo’s genius is that almost nothing it does requires high SRC in the moment of compliance. The lesson is bite-sized. The streak repair token is one tap. The owl notification is a designed prepotency strong enough to compete with whatever else was prepotent at 7:55 p.m. The hearts system is a Core Drive 8 (Loss) move that protects intention on low-SRC days; the leagues are a CD5 (Social Influence) supplement. The intention itself — “learn Spanish”, is a textbook long-horizon goal, and Duolingo’s daily-streak frame is what compresses its temporal valuation enough to compete with TikTok.

Personal Finance & Acorns / Wealthsimple

Round-up savings apps like Acorns are the cleanest behavioral-prepotency design move in consumer finance. They do not raise your intention to save. They route every existing spending prepotency through an automated savings transfer, so the contest never reaches your SRC at all. The intention is satisfied as a side-effect of behavior you were going to perform anyway. Wealthsimple’s daily rewards layer adds CD7 (Unpredictability) to convert what would have been a pure System 2 task into a partial System 1 reward loop.

Education & Khan Academy / Anki

The education stack is a TST playground because the payoff horizon is brutal: maybe in twenty years a student’s career benefits from learning calculus today. Khan Academy’s mastery-points system collapses that horizon into per-skill micro-accomplishments (CD2-mediated SRC supplement), while Brilliant’s daily problem and Anki’s due-card count are designed prepotencies competing for the same opening-the-phone-after-dinner moment that Instagram and TikTok already own. The systems that work in education are almost always the ones that match BP against BP rather than asking SRC to keep beating Instagram in a fair fight. Notice that you cannot out-motivate Instagram on cognitive-benefit grounds, because the temporal valuation gap is too large. You can, however, install a prepotency for “open Anki when you sit down on the couch” that wins half the contests on its own.

Smoking Cessation & Behavioral Pharmacology

Hall & Fong are tobacco-control researchers, and the original TST paper is shaped by smoking-cessation data. The empirical pattern is unmistakable: intention to quit predicts ten-year quit rates poorly. Intention combined with high SRC predicts them well. Intention combined with low BP. Smokers who have moved cities, ended a long relationship, or otherwise disrupted their cue environment: predicts them better still. The strongest predictor in the modern literature is the combination: intent, SRC support (nicotine replacement reduces the SRC-tax of resisting cravings), and BP collapse (changing the home, work, and social cue environment).

The Elephant in the Room

Most articles about TST stop here, around the application examples. I want to spend two paragraphs on what gets conspicuously avoided in the academic literature, because it is the conversation behavioral designers actually need to have.

TST predicts something uncomfortable: an enormous fraction of people will fail at their stated long-term intentions, and not because they are weak. They will fail because their SRC is structurally lower than the cohort whose advice they are following, or because their BP environment is structurally heavier, or both. Poverty raises cognitive load and reduces SRC bandwidth in ways measured cleanly in the scarcity-cognition literature (Mullainathan and Shafir). Shift work disrupts sleep-modulated SRC. High-cue food environments (vending machines, advertising, free office snacks) inflate BP for behaviors that healthcare campaigns are trying to suppress. None of these are personal failings. They are TST inputs.

This is also the elephant for designers. If you ship a product that quietly assumes the user has the SRC of a rested, well-fed PhD student in a stable cue environment, you have built something that works for the demographic least in need of help. The honest version of TST-informed design is: lower the SRC tax wherever possible, engineer the prepotency contest in the user’s favor in advance, and only ask intention to do the easy half of the work. Anything else is shifting blame to people whose contest you set up to lose.

There is a deeper ethical move available here, which I’ve been making in client work for years. The same TST architecture that lets us help people quit smoking, save money, and exercise more is the architecture every adversarial product is using against the user. Slot apps, infinite-scroll feeds, and dark-pattern checkout flows are all running TST against the user: they amplify a low-value intention through high-CD7 unpredictability, drain the user’s SRC with attentional fragmentation, and engineer prepotency in their own favor through cue saturation. If you understand TST and ship White-Hat-grade design, you are not merely helping users do what they say they want. You are competing against an army of products designed to exploit the exact gap you are trying to close. The bar is not “is my product nice.” The bar is “does my product give my user a real chance against the prepotencies the rest of their environment is building.”

How to Apply Temporal Self-Regulation Theory with the Octalysis Framework

This is the section every designer reading TST wants. The three TST terms map onto the eight Core Drives of the Octalysis Framework in a way that gives you a complete design checklist for any intention-behavior gap problem.

Raising the Intention Term (Compress the Temporal Frame)

The intention term in TST is weighted by temporal valuation. Two Core Drives do most of the work here.

Core Drive 1 (CD1): Epic Meaning & Calling raises the intention’s value by connecting the present-day behavior to a larger story whose payoff feels immediate because it is identity-anchored rather than time-anchored. A user who sees their daily five-minute Duolingo session as “being someone who keeps faith with a small daily promise” experiences the reward in the act, not at the imagined fluent-Spanish horizon ten years away. CD1 is the cleanest way to flatten the temporal discount curve without lying about future timelines.

Core Drive 4 (CD4): Ownership & Possession raises temporal valuation through endowment. A streak you have built, a graph you have grown, a tree you have watered for forty days — these feel present-tense valuable in a way an abstract future does not. Forest, the focus app, is a near-perfect CD4 instantiation: every minute of focus grows a tree you now own, so the cost of distraction is the immediate loss of an owned thing rather than the discounted loss of a productivity outcome.

Raising the Self-Regulatory Capacity Term (Supplement Executive Function)

SRC supplements are mechanics that reduce the inhibitory load of compliance, so a smaller native SRC is enough to win the moment. Three Core Drives handle the bulk of this.

Core Drive 2 (CD2): Development & Accomplishment supplements SRC by inserting interim rewards that close the temporal gap. The user does not need executive function to hold a distant goal in mind; they get a progress bar tick, a level-up, a visible accomplishment that lets the basal-ganglia reward circuit fire while the prefrontal cortex is busy with something else.

Core Drive 5 (CD5): Social Influence & Relatedness substitutes social-immediate consequences for delayed individual ones. Strava’s segment leaderboards are not motivating Olympians; they are converting a delayed cardiovascular outcome (matters in 20 years) into an immediate social outcome (matters when you upload the ride tonight). The intention does not need to win against couch-prepotency on the strength of cardiac-aging worry, it only needs to win against couch-prepotency on the strength of social standing this evening.

Core Drive 7 (CD7): Unpredictability & Curiosity supplements SRC by routing a chunk of the compliance reward through a System 1 dopaminergic channel. Variable-reward mechanics in moderation lower the SRC tax of compliance because they introduce a present-tense pull on the desired side of the contest. Used heavily this becomes manipulative; used lightly it converts willpower work into curiosity work, and curiosity is much cheaper.

Lowering the Behavioral Prepotency Term (Engineer the Contest)

The single most-skipped move in behavior-change product design is direct BP engineering. Three Core Drives are particularly powerful here.

Core Drive 3 (CD3): Empowerment of Creativity & Feedback lowers BP for the desired behavior by giving users active routes through their own environment that increase the prepotency of the new path. A meal-planning app that helps you redesign your weekly kitchen so the healthy option is the path of least resistance is doing CD3-mediated BP engineering.

Core Drive 6 (CD6): Scarcity & Impatience raises BP for the desired behavior selectively, by making the window for compliance feel narrow enough to override the competing prepotency. Daily-reset streaks, limited-time events, and “today only” framings all use CD6 to make the desired behavior the prepotent one for a short window. CD6 is a heat-of-the-moment lever. Use it where you actually want sub-day urgency, not as a constant background pressure.

Core Drive 8 (CD8): Loss & Avoidance raises BP for the desired behavior by making the cost of skipping it concrete and immediate: a lost streak, a broken commitment, a public failure to a small accountability group. CD8 deserves a warning label: it works powerfully in the short run but consumes SRC over weeks, because every CD8 mechanic is asking the user to overcome a loss frame, and loss-frame overrides are SRC-expensive. Use CD8 in punctuated doses, not as a constant pressure.

The White-Hat / Black-Hat Reading of TST

Pull the camera back and the Octalysis quadrants give a clean prescription. The top four Core Drives (CD1, CD2, CD3, CD4) are White Hat — they raise the intention term and supplement SRC without bleeding capacity. The bottom four (CD5 partial, CD6, CD7 with caveats, CD8) are Black Hat, they win moments by introducing urgency, scarcity, social pressure, or loss. Black Hat works fast and is SRC-expensive over time. White Hat works slowly and is SRC-replenishing over time.

The TST-informed design heuristic: lead with White Hat to keep SRC high across the user’s lifetime with the product. Use Black Hat in punctuated doses for the specific high-BP windows where SRC is structurally low. A product that runs Black Hat constantly. Constant streak fear, constant scarcity timers: will look great for six weeks of cohort data and then collapse, because it has been silently mining the SRC that made the early compliance possible.

Practical Steps for Designers

If you take one workflow away from this article, take this one. The next time you sit down to design a behavior-change product, run this checklist before you touch the screen.

1. Name the long-horizon intention out loud. Not the metric you optimize. The thing the user wants in three years that this product will help with. Write it on the wall. If you cannot name it, you do not have a product yet.
2. Estimate the temporal valuation gap. How far away does the payoff feel to your target user on a bad day? Hours, days, years? The bigger the felt distance, the more CD1 and CD4 work you need to compress it.
3. Audit Behavioral Prepotency, ruthlessly. List every competing behavior in the cue environment. Vending machines, social apps, default route home, friends’ habits. These are your real competition, not other apps in your category.
4. Estimate the user’s SRC at the compliance moment. Time of day, energy level, context. If your compliance moment is 9:30 p.m. after work and parenting, you are operating in a low-SRC window. Design accordingly.
5. Choose two CD-mechanics per term. Two for intention amplification (typically CD1 + CD4). Two for SRC supplementation (typically CD2 + CD5 or CD2 + CD7). Two for BP engineering (typically CD3 + CD8 in punctuated doses).
6. Ship White Hat as the constant, Black Hat as the pulse. Streak fear and scarcity work in small doses and corrode in large ones. Build the daily product on CD1-CD4, reserve CD6 and CD8 for specific high-stakes moments.
7. Re-measure prepotency after eight weeks. The desired behavior should be building its own prepotency by now. If it is not, your SRC supplements are too high a tax for the BP environment you are operating in. Lower the SRC ask, not the intention message.

Temporal Self-Regulation Theory Was the Beginning, Not the End

Hall and Fong wrote TST inside a health-psychology frame, and most subsequent research has stayed inside that frame. The deepest opportunity for behavioral designers is to recognize that the theory is bigger than the literature has noticed. Anywhere a user is asked to do something today for a payoff tomorrow — education, finance, climate behavior, creative work, civic participation, the same three-term contest runs, and the same design moves apply.

The piece I have been arguing with Hall & Fong about for years is also the piece they got right: that the contest is winnable. Not by inspiring people harder. By rigging the contest in advance. Compressing the temporal valuation through CD1 and CD4, supplementing SRC through CD2, CD5, and CD7, and engineering BP through CD3, CD6, and CD8 in disciplined doses. Build that, and the intention-behavior gap stops being the universal failure mode it has been for half a century of behavior-change products.

If you build one, tell me. I want to see the next Apple-Watch-rings-grade application of TST. The literature is full. The product shelf is mostly empty.

Frequently Asked Questions about Temporal Self-Regulation Theory

What is Temporal Self-Regulation Theory in one sentence?

Temporal Self-Regulation Theory says behavior happens when a person’s intention, multiplied by their available self-regulatory capacity, exceeds the strength of the competing prepotent habit: so closing the intention-behavior gap is a contest you can engineer, not a motivation problem you can shout at.

Who created Temporal Self-Regulation Theory?

Peter Hall and Geoffrey Fong, both tobacco-control and health-behavior researchers, introduced TST in a 2007 Health Psychology Review paper. Hall later extended the model with a neurobiological version in 2015 that mapped the three TST terms onto specific brain systems.

How is TST different from the Theory of Planned Behavior?

The Theory of Planned Behavior is excellent at predicting intention but weak at predicting behavior. TST keeps TPB’s intention-formation machinery and adds two moderators — Self-Regulatory Capacity and Behavioral Prepotency, that explain why intention so often fails to translate. The cleanest framing is that TST is the missing back half of TPB.

What is Behavioral Prepotency, exactly?

Behavioral Prepotency is the strength of the competing automatic response in the moment of decision. It has two components Hall & Fong specify explicitly: habit strength (how often the alternative has been performed in this context) and cue salience (how loudly the environment is currently calling for the alternative). The vending machine in the hallway has prepotency; the salad you packed at home does not.

Does the ego-depletion critique kill TST?

No, but it changes how you read the SRC term. Hall & Fong wrote SRC as a fluctuating capacity, which is consistent with current evidence: SRC really does vary with sleep, glucose, stress, mood, and recent cognitive load. The stronger ego-depletion claim. That every act of self-control immediately drains a finite pool: is much weaker than it looked in 2007, and TST does not require it.

How does TST apply to product design specifically?

Every product asking users to do something today for a future payoff faces a three-term contest. Most products spend all their design budget on the intention term (better copy, sharper benefit framing, louder motivation). TST predicts this is the smallest lever. The bigger payoff is in supplementing the user’s SRC at the compliance moment and engineering the prepotency contest in advance. The Octalysis Apply section above gives the specific Core Drive mapping.

Is TST evidence-based?

Yes. The 2007 paper is supported by twenty years of follow-up research, including health-behavior meta-analyses where TST’s SRC and BP terms add predictive variance over and above intention alone. The strongest empirical support is in physical activity, dietary behavior, alcohol use, and smoking cessation. Hall & Fong’s 2010 Health Psychology Review follow-up reviews the evidence base in detail.

What is the TST equation?

The informal version: Behavior = (Intention × Self-Regulatory Capacity) − Behavioral Prepotency. The formal version inside Hall & Fong’s paper is a moderated-moderation model in which SRC and BP jointly moderate the intention-behavior link rather than acting as additive predictors. The product form is the right one to design from.

Can TST be combined with Implementation Intentions?

Yes — cleanly. Gollwitzer’s if-then plans are one of the strongest SRC-supplement techniques in the literature. Inside TST, an implementation intention raises the probability that the desired behavior wins a specific cue-driven moment by pre-linking cue to response. They are tactical tools that work because TST’s architecture allows them to work.

Where should I read next if I want the design-side of TST?

Read Yu-kai’s Octalysis Framework pillar for the eight-Core-Drive model that the Apply section above uses, then the Control Theory pillar for the feedback-loop sibling of TST, and finally the Action Identification Theory pillar for the abstraction lever that sits inside the intention term.

References

1. Hall, P. A., & Fong, G. T. (2007). Temporal self-regulation theory: A model for individual health behavior. Health Psychology Review, 1(1), 6–52.
2. Hall, P. A., & Fong, G. T. (2010). Temporal self-regulation theory: Looking forward. Health Psychology Review, 4(2), 83–92.
3. Hall, P. A., & Fong, G. T. (2015). Temporal Self-Regulation Theory: A neurobiologically informed model for physical activity. In Conner, M. & Norman, P. (Eds.), Predicting and Changing Health Behaviour. Open University Press.
4. Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50(2), 179–211.
5. Gollwitzer, P. M. (1999). Implementation intentions: Strong effects of simple plans. American Psychologist, 54(7), 493–503.
6. Gollwitzer, P. M., & Sheeran, P. (2006). Implementation intentions and goal achievement: A meta-analysis of effects and processes. Advances in Experimental Social Psychology, 38, 69–119.
7. Heatherton, T. F., & Wagner, D. D. (2011). Cognitive neuroscience of self-regulation failure. Trends in Cognitive Sciences, 15(3), 132–139.
8. Hofmann, W., Friese, M., & Wiers, R. W. (2008). Impulsive versus reflective influences on health behavior: A theoretical framework and empirical review. Health Psychology Review, 2(2), 111–137.
9. Allom, V., & Mullan, B. (2014). Self-regulation versus habit: The dual role of self-regulation in healthy eating. British Journal of Health Psychology, 19(2), 264–282.
10. Booker, L., & Mullan, B. (2013). Using the temporal self-regulation theory to examine the influence of environmental cues on maintaining a healthy lifestyle. British Journal of Health Psychology, 18(4), 745–762.
11. Brown, D. J., Hagger, M. S., & Hamilton, K. (2020). The mediating role of constructs representing reasoned-action and automatic processes on the past behavior-future behavior relationship. Social Science & Medicine, 258, 113085.
12. Bickel, W. K., Koffarnus, M. N., Moody, L., & Wilson, A. G. (2014). The behavioral- and neuro-economic process of temporal discounting: A candidate behavioral marker of addiction. Neuropharmacology, 76, 518–527.
13. Wood, W., & Neal, D. T. (2007). A new look at habits and the habit-goal interface. Psychological Review, 114(4), 843–863.
14. Carver, C. S., & Scheier, M. F. (1981). Attention and Self-Regulation: A Control-Theory Approach to Human Behavior. Springer-Verlag.
15. Locke, E. A., & Latham, G. P. (1990). A Theory of Goal Setting and Task Performance. Prentice Hall.
16. Chou, Y. K. (2015). Actionable Gamification: Beyond Points, Badges, and Leaderboards. Octalysis Media.

Related Reading

• The Complete Octalysis Framework, the eight-Core-Drive system used in the Apply section above.
• Control Theory (Carver & Scheier). The feedback-loop sibling of TST.
• Action Identification Theory (Vallacher & Wegner): the abstraction lever that lives inside the TST intention term.
• Goal Systems Theory (Kruglanski) — the means-ends architecture behind multifinal goal-pursuit.
• The Behavioral Framework Library, the full index of pillars on yukaichou.com.

WOULD YOU LIKE YU-KAI CHOU TO WORK WITH YOUR ORGANIZATION?