Habit design (Hooked)
Variable Reward — Hunt
Unpredictable material rewards
What it is
Rewards of the Hunt are the search for material resources and information. Slot machines, news feeds, infinite scroll, Pinterest, Twitter — anywhere the next item is unknown.
Why it works
Variable rewards rest on findings from Skinner''s operant conditioning work: a behaviour reinforced on an unpredictable schedule is performed more persistently, and extinguishes more slowly, than one reinforced every time. Olds and Milner''s later research clarified the mechanism — dopamine surges in anticipation of a reward rather than on receiving it, so wanting and liking are separate systems and it is wanting that drives repetition. Rewards of the Hunt are Eyal''s category for the pursuit of material resources and information: the next item in a feed, the next result, the next find. The pull comes from the gap between one item and the next, which is why an infinite scroll with unknown contents is compulsive while a paginated catalogue with a visible end is merely useful.
How to apply it in design
Surface variable content (recommendations, drops, hidden finds, randomized features) where the user doesn't know what they'll discover. The hunt is the experience, not the product itself.
Where you'll see it
- A feed whose next item cannot be predicted from the current one
- Drops or restocks where the contents are revealed only at release
- Recommendations that occasionally surface something genuinely unexpected
- Search and discovery surfaces that reward digging with rare finds
- Subscription boxes where the selection, not the delivery, is the variable
Example
A wine subscription's monthly box. The variable is which two bottles arrive — that's the hunt, more compelling than any individual bottle.
Failure mode
Showing all available inventory upfront eliminates the hunt entirely. Predictable, browsable catalogs are utilities, not loops.
When not to use it
Variable rewards are wrong wherever the user has a specific goal. Someone looking for a particular document, product or setting wants deterministic retrieval, and randomness in that path is pure frustration — this is why injecting discovery into search results damages tools people rely on. They are also inappropriate in professional and safety-critical software, where predictability is the feature. And in any product touching money or chance, hunt mechanics move quickly toward gambling psychology and, increasingly, toward gambling regulation.
The ethical line
This mechanism is the engine of both slot machines and loot boxes, and several jurisdictions now regulate the latter as gambling. Two boundaries matter. The first is money: variable rewards purchased with real currency, especially by minors, are the clearest abuse and the one legislators have moved on. The second is exit — a loop the user cannot easily stop is not engagement but compulsion, and the honest version gives natural stopping points, an end of feed, session awareness, and no penalty for leaving. If retention depends on the user being unable to put it down, the design has failed on its own terms.
What to measure
Watch discovery rate — the share of sessions in which the user finds something they were not looking for — rather than raw time in feed. Pair it with a self-reported satisfaction check, because hunt mechanics can raise session length while lowering how good the experience felt.
Frequently asked questions
What is a variable reward of the hunt?
It is Nir Eyal's category for the pursuit of material resources or information where the payoff is unpredictable: scrolling a feed, opening a mystery box, searching for a rare find. The uncertainty about what comes next is what sustains the behaviour, rather than the value of any individual item.
Why is infinite scroll so compulsive?
Because it removes the natural stopping point and keeps the next item unknown. Dopamine responds to anticipation rather than to receipt, so the gap before each new item does the work. A paginated list with a visible end lets the user decide when to stop; an endless one never presents that decision.
Are loot boxes a variable reward of the hunt?
Yes, and they are the clearest example of where the mechanism crosses into gambling. Several jurisdictions now regulate paid loot boxes accordingly, particularly where minors are involved, because a randomised reward bought with real money is structurally a wager.
How do you use variable rewards ethically?
Keep real money out of the randomisation, give the loop genuine stopping points, and never inject unpredictability into paths where the user has a specific goal. If someone is searching for a known item, they need deterministic results, not discovery.
Variable Reward — Hunt in a real redesign
- UX case study: onboarding redesign for a specialty coffee subscription
Replaced five abstract taste options with sliders between ingredients anyone recognises — a choice the body already knows how to make.
Related skills
Source: Eyal — Hooked (Ch. 4)
Practise Variable Reward — Hunt on a real brief
Quest gives you a client brief that targets this skill and 60 minutes to ship an iteration.