Key takeaways
  • Hooked (Nir Eyal) names a four-step loop: Trigger, Action, Variable Reward, Investment.
  • Teams misapply it by chasing engagement without delivering real value.
  • There's an ethical line Eyal himself raises, and the model doesn't apply everywhere.

Of all the product books on my shelf, the one I lend out the most, and the one I find myself reaching for in meetings six years after first reading it, is Hooked by Nir Eyal. Not because it's the most rigorous or the most academic book in the genre. There are heavier reads with more citations. The reason I keep going back to it is that it puts a single, sharp name on something product people had been intuiting for years without being able to point at: the loop that turns a one-off user into a habit.

The Hook Model is four steps. Run a user through that loop enough times and the product no longer has to fight for attention each time. The trigger has moved from external (your push notification, your ad spend) to internal (boredom, FOMO, status anxiety, the empty pocket reach-for-phone). Once that switch flips, acquisition cost falls off a cliff and retention compounds. That is the entire game for habit-forming consumer product. Eyal didn't invent it, but he named it, and the naming is what makes it useful in a room with twelve stakeholders.

The four-step loop.

Diagram · The Hook Model, the four-step loop named by Nir Eyal
The Hook Model — the four-step loop named by Nir Eyal in Hooked A circular diagram showing the four steps of the Hook Model arranged clockwise around a central badge that reads "Habit forms after enough loops." Step 1 Trigger sits at the top: external cues like push notifications, eventually replaced by internal cues like boredom or anxiety. Step 2 Action sits at the right: the simplest behaviour the user takes in anticipation of reward — a tap, a scroll, a search. Step 3 Variable Reward sits at the bottom and is highlighted in marigold: the unpredictable payoff Eyal classifies as Tribe (social validation), Hunt (information, resources) and Self (mastery, completion). Step 4 Investment sits at the left: the user puts something into the product — data, content, social capital — which improves the next loop and makes leaving costlier. A marigold loop ring connects all four nodes with four arrowheads tangent to the ring indicating clockwise flow. Each successful loop makes the next loop easier and cheaper to trigger. HABIT FORMS after enough loops STEP 01 Trigger external → internal STEP 02 Action simplest behaviour STEP 03 Variable Reward tribe · hunt · self STEP 04 Investment user puts in Each loop makes the next loop easier · the trigger eventually fires from inside the user, not from your push budget.
Eyal's four-step loop, named in Hooked. The variable-reward node is highlighted because it's where most of the model's psychological pull sits, and where most teams mis-apply the model by reading it as gamification.

Trigger. The cue that nudges the user to engage. External at first: your push notification, your email, the ad they saw on Instagram. The model's clever move is that over enough loops the trigger moves inside the user. You reach for your phone because you're bored. You open Twitter because you feel anxious. You check Strava because you went for a run. The cue is now an emotion or context, and the product has won the most valuable real-estate there is: a slot in the user's involuntary nervous system. That's habit.

Action. The simplest possible behaviour the user takes in anticipation of reward. The pull-to-refresh. The tap to like. The swipe right. Each one is the easiest possible payment for the reward the user is about to get. Eyal leans on B.J. Fogg's behaviour model here. Behaviour happens at the intersection of motivation, ability, and a trigger. Reduce friction on ability and you've removed the resistance to action.

Variable Reward. The most powerful step, and the most often misread. The reward the user gets must be unpredictable in some dimension. That's what activates the dopamine-anticipation cycle that keeps them coming back. Eyal classifies variable rewards into three: Tribe (social validation, likes, comments), Hunt (resources, information, money, the slot-machine class), and Self (mastery, completion, the satisfaction of a streak). Most successful consumer products feature at least one. The most powerful feature two.

Investment. The user puts something into the product before the loop closes: data, content, a follower count, configuration, social graph, history. The investment improves the next loop's variable reward (more followers → better feed → bigger reward) and increases the cost of leaving. Investment is the step most teams skip when they read the model. They nail trigger and reward, then forget that the user needs to contribute for the loop to compound.

Trigger and Reward get the user back. Investment is what makes the next loop better than the last. Skip Investment and the loop runs flat.

Where teams get it wrong.

Three common misapplications:

  • Reading Variable Reward as gamification. Badges and points are a shape of variable reward, but they're the weakest one. Real variable reward is structural: the content in the feed, the score on the workout, the unread message. Bolting confetti on a flat product doesn't make it habit-forming, it makes it annoying.
  • Skipping Investment. A product that doesn't ask the user to contribute can drive engagement loops but won't build a moat. Snap had Investment (friend list, streaks). Vine didn't. Snap survived; Vine didn't.
  • Applying the model to the wrong product. The Hook Model is designed for high-frequency consumer products. Apply it to a once-a-year tax-filing tool, a B2B enterprise system used twice a week, or a one-shot booking flow and you get a frustrating product that won't leave the user alone. Frequency of use matters. If a product is naturally low-frequency, forcing a Hook Model loop on it is fighting the user's actual need.

The ethical line — Eyal raises it himself.

The most interesting thing Nir Eyal did after Hooked was write Indistractable, a book about how to resist the kind of products Hooked teaches you to build. That tension is worth sitting with. The Hook Model is, in the wrong hands, a manual for extractive design: apps that earn ad revenue by training users into compulsive behaviour against their own interest. Eyal's response to that critique is what he calls the manipulation matrix: build only products you yourself would use, and that materially improve the user's life. Habit-forming for the user's benefit is fine. Habit-forming for the company's benefit at the user's cost is the line.

It's a soft answer to a hard question, but it's an honest one. If you're applying the Hook Model to a product, the test that matters is: would your user, in their best moment, thank you for the habit you're forming?

When the model applies.

Some products are natural fits and the Hook Model just describes what they already do. Social products (Instagram, TikTok, Reddit). Content discovery (Spotify, YouTube). Status and gamified-progress (Strava, Duolingo, Wordle). Some financial and health behaviour-change tools (Headspace, Calm, Robinhood, though Robinhood is the most contested ethically).

Where it doesn't apply: low-frequency or task-completion tools. Tax-filing software. Conveyancing. Enterprise procurement. Annual insurance renewal. Booking a tradesperson. Forcing a Hook Model loop on a product the user wants to finish and not come back to is the failure mode that causes the worst kind of dark-pattern reputation.

How we use Hooked at Product Pieces.

The Hook Model lives in our diagnostic kit as a question, not a doctrine. When a consumer team brings us a product where engagement isn't landing, we walk the four steps with them. Where's the Trigger? What's the simplest Action? Is the Reward genuinely variable, or is it fixed and known? Most importantly, what's the user's Investment, and does it compound?

Eight times out of ten the gap is at Step 3 or Step 4. Variable Reward is missing or fixed, or Investment is asked too late in the user journey. Once that's named, the fix is usually a single piece of work, not a redesign. That's the value of the model: it puts a name on what was missing.

If you're building a habit-forming product, Hooked is on the short list of books worth re-reading once a year. If you haven't yet — go and read it.

Next issue: N°37. What actually is a product? A thought experiment for anyone who builds them.