Nudge Wiki

Concept

Variable Reward Schedule

Also known as: Intermittent Reinforcement

operant-conditioning reward addiction

A reward delivered unpredictably — sometimes present, sometimes not — produces more persistent, harder-to-extinguish repeated behavior than a reward delivered every single time. First demonstrated by B.F. Skinner with pigeons trained to peck a disc for food (see Skinner Pigeon Variable Reward Study); the same mechanic drives engagement with gambling, social media, news, and games in humans.

Nir Eyal categorizes the variable rewards used by habit-forming products into three types: rewards of the tribe (social validation from other people — e.g. upvotes on Stack Overflow), rewards of the hunt (the search for information or material resources — e.g. endless scrolling for news), and rewards of the self (intrinsic feelings of competence, autonomy, and mastery — e.g. checking off a to-do list). A fridge you've stocked yourself is more compelling to open repeatedly than an empty one precisely because its contents are variable.

Nathalie Nahai adds the neuroscience: dopamine, discovered in the 1950s by researchers Carlson and Hillop, drives the wanting/seeking behavior that keeps someone checking a feed, most strongly via D2 receptors in the brain's mesolimbic pathway — while a separate opioid system governs the liking/satisfaction once a reward actually lands. Because online platforms deliver rewards unpredictably, the seeking loop can run even when the outcome itself isn't satisfying (see Habit Loop (Cue-Routine-Reward)). A practical marketing application: adding a spinning prize wheel with unpredictable rewards to marketing emails raised conversion from about 0.5% to about 3% (see Neil Patel Spinning Wheel Email Variable Reward).

Louis Grenier cites TikTok's recommendation algorithm as a modern application: rather than showing visually or emotionally similar videos on every scroll, it deliberately mixes good, bad, funny, and interesting content in an unpredictable "Goldilocks" blend, keeping users scrolling for the next satisfying hit the same way Skinner's pigeons kept pecking.

Applied to loyalty rewards (Episode 172): Pret a Manger gives staff a weekly quota of free items to hand out entirely at random, rather than a predictable punch-card loyalty scheme, making recipients disproportionately happy for the same average giveaway cost (see Pret a Manger Random-Reward Loyalty Case).

A controlled demonstration and a public-policy application (Episode 179): a 2014 study found 70% of participants completed a task for an unpredictable $1-or-$2 reward, versus only 43% for a certain $2 reward — a higher completion rate for the lower-expected-value option (see Luxi Shen Variable-Reward Task-Completion Study). Rory Sutherland cites this as an argument for paying welfare benefits as variable lump sums rather than small, fixed, predictable weekly payments.

A precise neuroscience demonstration and a personal application (Episode 240). Bas Wouters cites a 2015 study monitoring a monkey's dopamine response: dopamine spiked on seeing an image signaling an upcoming banana-juice reward, and fell once the juice was actually delivered ("anticipatory enthusiasm"), then doubled again once the reward was made uncertain (delivered only 50% of the time) rather than guaranteed (see Anticipatory-Dopamine Monkey Study). Phill Agnew found the same pattern in his own behavior: a cafe table sign offering an undisclosed "surprise" in exchange for a QR-code review was enough to make him leave a review he otherwise wouldn't have, precisely because the reward's uncertainty made it more compelling than a known, named reward would have been (see Nudge Podcast Surprise-Reward Cafe Review Case).

A brand stunt built on the same principle (Episode 239). Chris Baker cites Tony's Chocolonely's advent-calendar stunt — removing the chocolate from day 8 to illustrate supply-chain inequality, then including two chocolates on day 9 — as a deliberately variable reward pattern, attributing the mechanism to B.F. Skinner and a 1970s follow-up by Mazur (see Tony's Chocolonely Advent-Calendar Variable-Reward Case).

Consumers actively prefer surprise over certainty, under specific conditions (Episode 279). Mehdi Bouhassan cites Eva Buechel and Ryan Li's 2022 eight-study paper finding people choose a random/mystery option over their own stated favorite 69-89% of the time across ice cream flavors, stress balls, hotels, and songs — but only when the options are "horizontally" different (a matter of taste, no real quality gap); when options are "vertically" different (one is objectively better), certainty is strongly preferred instead (see Buechel Li Mysterious-Consumption Studies). HelloFresh's early meal-kit growth leaned on exactly this horizontal-surprise appetite, then had to abandon it once retention required the vertical certainty subscribers wanted long-term (see HelloFresh Effort-and-Surprise Growth Case).

Octopus Energy's meter-reading wheel of fortune (Episode 296). Pete Miller cites B.F. Skinner's founding 1938 rat-lever research directly as design inspiration, and a 2015 University of Chicago study (Lucy Shen, Ayelet Fishbach, Christopher Hsee) found a 50/50 chance of $1-or-$2 drove 70% task completion versus 43% for a guaranteed $2 — an objectively worse expected payout still outperforming certainty (see Octopus Energy Wheel-of-Fortune Meter-Reading Case, Shen Fishbach Hsee 2015 Uncertain-Reward Study).

Discussed in

  • Episode 13 — 13- What makes products like Instagram so addictive
  • Episode 17 — 17- How marketers alter our behavior
  • Episode 18 — 18- 8 nudges you need in 2020
  • Episode 60 — 60- How not to suck at customer interviews (TikTok's algorithm)
  • Episode 172 — 172-why-these-irrational-prices-actually-work (Pret a Manger random rewards)
  • Episode 179 — 179-it-s-how-mcdonald-s-make-men-binge-rory-sutherland (Luxi Shen study; welfare-payment application)
  • Episode 239 — 239-six-eco-brands-using-psychology-to-sell (Tony's Chocolonely advent-calendar stunt)
  • Episode 240 — 240-learn-7-scientifically-backed-marketing-tips-in-27-minutes (2015 monkey dopamine study; Nudge's own cafe-review case)
  • Episode 279 — 279-the-psych-trick-behind-one-of-the-decade-s-fastest-growing-orgs (Buechel Li mysterious-consumption studies; HelloFresh)
  • Episode 296 — 296-how-we-used-nudges-reach-12-billion-in-sales-octopus-energy-s-pete-miller (Octopus Energy wheel of fortune; Shen Fishbach Hsee study)

Related