I can predict the shape of your retention curve before I see it. If your product grew on points, streaks, badges, or a daily reward of any kind, the curve looks like this: a hopeful spike, a brief plateau, then a slow bleed that no bonus calendar or double XP weekend can reverse. I have watched this pattern repeat across industries for more than a decade, from airlines to education apps to fitness platforms. The mechanics differ. The failure does not.
The painful part is that the reward system worked. That is exactly why it failed.
Let me explain, because this is the sentence most product teams never hear: a simple reward does not build a habit. It rents behavior. And the rent goes up every quarter.
The failure pattern: the reward becomes the reason
Yu-kai Chou calls the naive version of this the PBL Fallacy: the belief that if you put points on something boring, add some badges, and attach a leaderboard, the boring thing automatically becomes exciting. Entire platforms exist to sell you this at scale, which is why so many outsiders conclude that gamification equals points, badges, and leaderboards. They are not wrong that PBLs motivate behavior. They do. Briefly.
Here is what actually happens inside a user’s head when you attach a reward to an action.
Week one, the reward is a pleasant surprise. The user does the behavior, gets the thing, feels a small hit of progress. By week four, the reward is no longer a bonus. It is the price of the behavior. The user’s brain has quietly renegotiated the deal: I do this, you pay me. Psychologists documented this decades ago as the overjustification effect: offer an external reward for something a person might have done willingly, and intrinsic interest drains out of the activity. Yu-kai describes the cliff that follows in Actionable Gamification: once companies stop offering the extrinsic motivator, user behavior drops to a level even lower than before the reward was first introduced.
Read that again. Lower than before you started. The engagement you thought you bought was rented all along, and the rent bill keeps arriving.
The diagnosis: Left Brain rewards, Right Brain habits
The Octalysis Framework sorts the eight Core Drives of motivation into two families. Left Brain drives are extrinsic: you act because of a goal, a reward, a consequence. Development & Accomplishment (CD2), Ownership & Possession (CD4), and Scarcity & Impatience (CD6) live here. Right Brain drives are intrinsic: the activity itself is the point, no reward required. Empowerment of Creativity & Feedback (CD3), Social Influence & Relatedness (CD5), and Unpredictability & Curiosity (CD7) live there. (Epic Meaning & Calling and Loss & Avoidance sit at the top and bottom of the octagon, outside the left/right split.)
Now look at what a simple reward system actually activates. Points, badges, and leaderboards all pull Core Drive 2: development and accomplishment, but of the score, not the skill. An expiring bonus is CD6 urgency. A streak you can lose is CD8 anxiety dressed up as engagement. Every one of these is extrinsic. Every one points at the prize, not the practice.
Durable habits, the kind that survive a bad month, a vacation, a competing app, are held together by Right Brain drives. The designer who opens the tool because making things in it is enjoyable (CD3). The runner who runs because her running group is waiting for her (CD5). The explorer who returns because every session hides something new (CD7). Nobody has to pay these people. The behavior pays them.
This is the core failure of simple reward systems, and it is structural rather than cosmetic: they activate extrinsic Left Brain motivation and never engineer the transition to intrinsic Right Brain motivation. Worse, the reward actively blocks the transition. The user never builds an identity around the behavior, because the identity that forms is “person who collects the reward.” The moment the reward stops feeling scarce or valuable, there is nothing underneath.
A real product under the lens: Duolingo’s streak
Take Duolingo, the most successful habit product in consumer education. Ask a product manager why users return daily and you will hear one word: the streak.
Now analyze the streak honestly. A streak is CD8 Loss & Avoidance with a CD4 counter attached. Break it and users feel genuine pain, which is exactly why Duolingo sells streak freezes and repairs. That is not a habit mechanism. That is a hostage mechanism, and Duolingo is smart enough to know it. Watch what the product builds around the streak: visible skill progression and fluency gains (CD2), leagues and friend quests (CD5), surprise chests and varied lesson formats (CD7 and CD3). The streak gets users to the door. Everything else works to convert the visit into something they would miss for its own sake.
The lesson is that a streak is scaffolding. Duolingo succeeds to the degree that, somewhere in the first hundred days, the user stops opening the app to protect a number and starts opening it because learning Spanish feels like becoming someone. When a simple reward system is the entire product, there is no somewhere to arrive at.
What we saw at LATAM Airlines
I watched this exact failure play out at scale when we worked with LATAM Airlines, the largest airline group in Latin America. Their frequent flyer program was a classic earn-and-burn machine: accumulate miles, redeem miles, go quiet until the next booking. Transactional behavior, no lasting loyalty. Millions of members, and the program had to keep paying them to care.
The motivational architecture told the story. Expiring miles and annual status resets hammered CD6 and CD8: urgency through anxiety, relief when you kept your status, resentment when you lost it. The miles themselves were CD4 gone wrong, because a collection that can be taken away is a loan, and a loan can never feel like possession. And CD1, Epic Meaning & Calling, was absent, which is remarkable when you consider that travel is one of the few categories where meaning comes almost for free. People define themselves by where they have been. The program tracked spending, not identity.
The redesign added what the reward never could: an identity layer that turned travel history into a personal narrative, a rebalanced ownership layer with less anxiety, and a discovery layer built on CD7 surprise. You can see more of this kind of work on our case studies page. The question we asked during that engagement is the one I now hand to every client: when a member looks at their profile, do they see an identity they are proud of, or a balance they are afraid to lose? The answer tells you whether your program builds loyalty or rents it.
My hard threshold: the 30-day reward holiday
Here is a falsifiable claim, and I invite you to test it against your own data. If you pause every extrinsic reward in your product for 30 days and daily active usage drops by more than 50 percent, your system has produced zero durable habits. Not few. Zero. What you were measuring was never engagement with your product. It was engagement with your payout, and the day a competitor pays slightly more or slightly faster, your “loyal” users will leave without a flicker of guilt.
Most teams refuse to run this test. Deep down, they know what it would show.
The four-question reward audit
You do not need the full framework to start fixing this. You need honesty and four questions. Sit down with your product and answer them in writing:
1. If the reward disappeared tomorrow, what would users still do? Whatever survives is your real product. Everything else is a subsidy.
2. Does the reward point at the behavior or at the prize? If users talk about the points, the streak, the discount, the reward owns the relationship. If they talk about the skill, the progress, the identity, you own it.
3. Can a user describe who they are becoming, or only what they are earning? Habit lives in identity. “I am becoming a runner” survives a broken streak. “I am earning a voucher” does not survive a better voucher.
4. Is the experience more interesting on day 100 than on day 10? Right Brain motivation compounds: richer creativity, stronger social ties, deeper curiosity. Extrinsic rewards only inflate. If your day-100 experience needs a bigger prize than day 10, you are on the treadmill.
Score poorly on two or more of these, and your reward system is a recurring marketing expense wearing a gamification costume. The costume is the only gamification in it.
Build the transition, not the reward
The fix is not to delete rewards. PBLs have their place; they are excellent at triggering first actions and bridging cold starts. The fix is to treat every reward as scaffolding with a demolition date. Let CD4 and CD6 get the behavior started, then deliberately hand motivation over to CD3 creativity, CD5 belonging, and CD7 curiosity. The reward should become less necessary with every repetition, and your design roadmap should state exactly how.
If you ask gamers what makes a game fun, none of them say “the points.” They talk about strategy, challenge, friends, the world. Games without a single extrinsic reward, including text-based adventures with no graphics at all, have held players for decades. The fun was never in the payout. Your habit will never live there either.
If your retention curve shows that familiar slow bleed and you want a second pair of eyes on the motivational architecture underneath it, Contact us and we will examine it together.
