(Behavioural Science) #19 Temptation Bundling

Principle · Habit formation category

Temptation bundling

The practice of pairing a behavior you want to do (an immediately rewarding "want") with a behavior you should do but tend to avoid (a future-beneficial "should"). By making the tempting activity contingent on the virtuous one, the immediate reward of the want is harnessed to motivate the should — converting present bias from an obstacle into a fuel source.

51%

more gym visits when audiobooks restricted to gym use only (Milkman et al.)

29%

increase in goal completion when temptation bundling vs. willpower alone

2014

Milkman, Minson & Volpp — the coining study

High

replication rate — robust across fitness, studying, saving, and healthcare domains

1. What it is and the science behind it

Most behavior change approaches try to increase willpower, strengthen intentions, or remove friction from the desired behavior. Temptation bundling takes a different route entirely: it stops fighting present bias and starts using it. The core insight is that the same psychological force that causes people to skip the gym in favor of watching television — the overweighting of immediate pleasure relative to future benefit — can be redirected if the immediate pleasure is made contingent on the future-beneficial behavior.

Katherine Milkman, who coined the term in her 2014 paper, frames it as a solution to the "want/should" conflict. Wants are activities with immediate rewards and delayed or no costs — watching a gripping TV show, eating indulgent food, scrolling social media. Shoulds are activities with delayed rewards and immediate costs — exercising, studying, completing administrative tasks. Present bias systematically favors wants over shoulds. Temptation bundling restructures the choice so that the want is only available during the should.

The core structure

The bundling equation

Should (low appeal, high future value)

+

Want (high appeal, immediate reward)

→

Bundle (want unlocked only during should)

Going to the gym

+

Listening to a gripping audiobook

→

Audiobook only plays at the gym

Doing tax returns

+

Watching a favourite TV show

→

TV only allowed during admin tasks

Why it works — four mechanisms

Present bias redirectionRather than fighting the pull toward immediate reward, temptation bundling attaches that immediate reward to the virtuous behavior. The same present-biased brain that would skip the gym now has an immediate reason to go — the audiobook it wants right now is only available there.

Opportunity cost reframingSkipping the should no longer means "I avoid the unpleasant task." It now also means "I lose access to the want." The cost of avoidance becomes more tangible and immediate, raising the psychological barrier to skipping.

Positive association buildingRepeated pairing of the should with a genuinely enjoyable want builds a conditioned positive association with the previously aversive behavior. Over time, the gym begins to feel associated with pleasure rather than effort — reducing the long-term reliance on the bundle.

Intrinsic motivation preservationUnlike external reward systems (gold stars, cash incentives), temptation bundling uses rewards the person already wants for their own reasons. It does not crowd out intrinsic motivation or create dependence on external validation — the want is self-chosen and self-sustaining.

Key studies

Audiobooks and gym attendance — the coining study

Milkman, Minson & Volpp, 2014

University staff were given iPods preloaded with gripping audiobooks (page-turners chosen by participants) and randomly assigned to one of three conditions: keep the iPod at home (control), keep it at the gym but use it freely (weak bundle), or only access the audiobook at the gym (strong bundle). The strong bundle condition — where the want was strictly contingent on the should — produced a 51% increase in gym visits compared to the control group over nine weeks. The weak bundle, where access was not restricted, showed a much smaller effect, confirming that the contingency structure — not merely the presence of the enjoyable activity — is the active ingredient.

+51% gym visits with strict contingency vs. control

Temptation bundling and flu vaccination

Milkman et al., 2021 — NBER Working Paper

A large-scale field experiment with over 8,000 employees offered flu vaccination bundles — pairing the vaccine appointment with a small indulgence (a preferred coffee drink, a gift card to a preferred retailer). Uptake was significantly higher in bundled conditions than in standard reminder conditions. The effect was strongest for employees who had previously reported vaccination as unpleasant but intended to get vaccinated — exactly the want/should conflict that temptation bundling is designed to resolve. The bundle provided the final motivational push at the moment of scheduling. Bundling increased vaccination uptake significantly vs. standard reminder

Self-control and wanting vs. liking in habit formation

Woolley & Fishbach, 2016 — Journal of Personality and Social Psychology

Participants who reported enjoying their exercise routine — regardless of whether the enjoyment was intrinsic or from a paired reward — showed significantly higher long-term adherence than those who exercised purely out of obligation. Critically, immediate rewards (enjoyable music, social company, interesting podcasts) were stronger predictors of continued exercise than delayed rewards (weight loss, health outcomes). This confirms the mechanism: immediate enjoyment during the behavior, not anticipated future benefit, drives habit persistence. Temptation bundling works by engineering the immediate enjoyment condition artificially.

Immediate enjoyment during exercise predicted adherence better than delayed health outcomes

Retirement savings and "want/should" conflicts

Milkman, Rogers & Bazerman, 2008

An earlier framing study showed that people reliably choose healthier foods and more virtuous options for the future than for the present — ordering salad for next week's lunch but pizza for today. This "want/should" asymmetry is the foundational evidence that present bias creates a stable, predictable gap between intended and enacted behavior. Temptation bundling works precisely in this gap: it does not require the person to resolve the conflict — it changes the structure of the choice so the conflict is bypassed entirely.

Future-self choices reliably more virtuous than present-self choices — the gap temptation bundling targets

2. Real application examples

Business

Fitness and wellness apps

Apps like Peloton and running apps pair the workout (should) with curated playlists, podcast integrations, or social leaderboards (want) that are only accessible during active sessions. The entertainment is contingent on the workout — a product-level implementation of temptation bundling that drives both retention and session completion rates.

Business

Workplace learning platforms

E-learning platforms that integrate high-quality short-form video content — genuinely interesting and entertaining — alongside required compliance or skills training apply a bundling logic. Learners who choose their own "reward" content (a fascinating talk, a skills masterclass) in exchange for completing required modules show higher completion rates than those given purely obligatory content.

Business

Loyalty programs and desirable behavior

Retailers and banks bundle access to desirable exclusive experiences (early sale access, exclusive events, premium customer service) with behaviors they want to encourage (regular purchases, higher account balances, product diversification). The want — status, access, exclusivity — is contingent on the should — continued engagement and product usage.

Public policy

Vaccination bundling

Following Milkman's research, some healthcare systems and employers have experimented with pairing vaccination appointments with small desirable treats — preferred beverages, retail vouchers, or entry into prize draws. The bundle is not a bribe (the want is incidental, not the primary motivation) — it provides an immediate positive at the point of scheduling for a behavior whose benefits are distant and abstract.

Public policy

Healthy eating in institutional settings

School and hospital cafeteria designs that pair healthy food choices with access to preferred seating, social environments, or entertainment screens create environmental temptation bundles. Choosing the salad gets you the nice window seat; choosing the burger gets you the fluorescent-lit back row. Subtle, but measurably effective in cafeteria trials.

Public policy

Active travel and commuting

Cities and employers that provide secure, convenient, enjoyable cycling infrastructure — good showers, bike cafés, scenic routes, social cycling groups — are bundling the should (active commuting) with wants (pleasant experience, social connection, preferred facilities). The bundle is built into the infrastructure rather than added as a reward on top.

Personal habit

Exercise and entertainment

The classic personal application: only allowing yourself to watch a favourite TV series, listen to a specific podcast, or read a page-turner audiobook while exercising. The want becomes strictly contingent on the should. Many people report this as their most effective personal behavior change tool — partly because it is self-designed and therefore perfectly matched to their own wants.

Personal habit

Unpleasant admin tasks

Pairing tax returns, email triage, expense filing, or other aversive administrative tasks with a genuinely enjoyed indulgence — a specific coffee, a favourite music playlist, a pleasant location — converts the should into a bundle. The indulgence is reserved exclusively for the admin session, maintaining the contingency that is the active ingredient.

Personal habit

Learning and skill development

Pairing language learning (should) with content in that language that is genuinely interesting — watching beloved films with subtitles, reading about a passion topic, following social media in the target language — converts obligatory study into desired consumption. The should becomes the vehicle for accessing the want, rather than a sacrifice made in place of it.

3. Design guidance — when and how to use it

The central design insight

The active ingredient in temptation bundling is the contingency — the want must only be available during the should. Allowing the want independently (listening to the audiobook on the commute as well as the gym) dramatically reduces the effect. The bundle must be strict enough that the want creates a genuine pull toward the should, not just a pleasant accompaniment to it.

Four rules for effective bundle design

Rule 1

The want must be genuinely compelling

A mild preference is not enough. The want needs to be something the person actively craves — a page-turner, a favourite show, a deeply enjoyed indulgence. The motivational pull must be strong enough to compete with the aversion toward the should.

Rule 2

The contingency must be strict

The want is unavailable outside the should. Half-measures — "I try to save the audiobook for the gym" — produce much weaker effects than genuine restriction. Milkman's weak bundle condition confirmed this: contingency, not proximity, is the mechanism.

Rule 3

The want must be compatible with the should

The want should not impair the should. A gripping audiobook works at the gym; it does not work during a task requiring deep cognitive focus. Match the type of want to the attentional demands of the should — sensory or narrative wants work for physical tasks; social or ambient wants work for administrative tasks.

Rule 4

Let people choose their own wants

Self-selected wants produce stronger and more durable bundles than assigned ones. Milkman gave participants a menu of audiobooks rather than prescribing one. Personal choice increases the pull of the want and maintains intrinsic motivation — the reward stays genuinely desirable over time.

When this principle works well

Use when

There is a clear want/should conflict — the person intends to do the should but consistently delays or avoids it in favor of more immediately rewarding alternatives.

Use when

The should involves repeated behavior over time — exercise, studying, healthy eating, professional development. The bundle builds habit and positive association cumulatively.

Use when

A compelling, self-relevant want can be identified — one the person actually craves, not one they think they should enjoy. The want must be genuinely tempting to that specific person.

Use when

The want and should are physically or temporally compatible — they can occur simultaneously, or the want can bookend the should as an immediate reward at completion.

Avoid when

The should requires deep focused attention and the want is distracting. Pairing complex analytical work with entertaining media impairs performance on the should.

Avoid when

The goal is to build intrinsic love of the should itself. For activities where the long-term goal is autonomous enjoyment — learning a musical instrument, developing a reading habit — bundling can delay intrinsic motivation development.

Step-by-step design process

1. Identify the should precisely — the specific behavior that has a persistent intention-action gap. "Exercise more" is too vague; "go to the gym on Tuesday and Thursday mornings" is the actionable should.
2. Inventory genuine wants for that person — not what they think they should enjoy, but what they actually crave and tend to consume in moments of low resistance. Favourite podcasts, specific TV shows, particular foods, social activities, games. The more personally resonant, the stronger the pull.
3. Match want type to should type — physical/repetitive shoulds can accommodate narrative or entertainment wants (audiobooks, TV, podcasts). Administrative shoulds work with ambient or sensory wants (preferred music, favourite café environment, indulgent food). Cognitive shoulds require lower-distraction wants (pleasant environment, preferred beverage).
4. Enforce the contingency — create a structural mechanism that makes the want unavailable outside the should. This might be a physical rule (only listen to this podcast at the gym), a digital mechanism (app access unlocked by a health tracker), or a social commitment (only watch the series with a friend who also has a should to complete).
5. Refresh the want periodically — wants habituate. A series that is gripping in episode 1 is mandatory by episode 8. Plan for want rotation: new audiobooks, new series, new indulgences that keep the pull strong over months, not just weeks.
6. Plan the exit strategy — the goal is eventually to internalize the should as a positive habit, at which point the bundle can be relaxed. Track whether positive association with the should is building independently of the want: "do I look forward to the gym even on days I'm not listening to the audiobook?" That is the signal the habit has been established.

Before and after — design framing

Exercise habit — personal design

Weak (willpower approach)

"I need to be more disciplined. I'll force myself to go to the gym three times a week even though I don't want to."

Strong (bundle approach)

"I'm only allowed to listen to the next episode of [favourite podcast] while I'm at the gym. I download it Sunday night so it's waiting for me Monday morning."

Workplace learning — product design

Weak (obligation framing)

"Complete your mandatory compliance training by end of quarter. Modules are available on the learning portal."

Strong (bundle design)

"Complete your compliance module to unlock this month's masterclass: [topic the employee chose at onboarding]. Takes 20 minutes — the masterclass is 45 and worth it."

Healthy eating — cafeteria design

Weak (information only)

"Calorie information is displayed next to all menu items. Choose wisely."

Strong (environmental bundle)The salad bar is positioned next to the window seats with the best view and the coffee station. The burger station is at the back near the kitchen. Choice architecture creates a soft bundle between the healthy option and the desirable environment.

Two failure modes to design against

The first failure mode is a weak contingency — allowing the want outside the should "just this once," which rapidly collapses the bundle. The second is choosing the wrong want: one that is mildly pleasant but not genuinely tempting, or one that the person feels they should enjoy rather than one they actually crave. Both produce bundles that feel good in design but generate no meaningful pull at the moment of decision. The test of a well-designed bundle is simple: does the person feel a genuine pull toward the should because of the want? If the honest answer is no, the want needs to change. This is why self-selection of the want is so important — no designer knows another person's genuine cravings as well as the person themselves.