Skip to content
  1. Home
  2. Blog
  3. The PBL Fallacy: Why Points, Badges and Leaderboards Don’t Build Habits
Gamification

The PBL Fallacy: Why Points, Badges and Leaderboards Don’t Build Habits

Joris Beerda12 August 2026
Professional woman in a bright glass office looking skeptically at a Login Champion medal, with a gamified learning dashboard on her laptop

A reward system that works gets more expensive every quarter. The first points campaign lifts activity. A few months later the same activity takes double points, then a new badge tier, then a bigger prize at the top of the leaderboard, and none of it can be switched off without the numbers falling. You end up paying more each quarter for behavior that never became a habit.

Simple reward systems fail at long-term habit formation because a habit takes months of repetition, and a fixed reward loses its pull before that. While the reward runs, it also becomes people’s explanation for why they act at all. When it shrinks or stops, the behavior goes with it.

But what about Duolingo? Its streak is about as simple as a reward system gets, and it works. So why does the same streak fail when you copy it onto your product? The answer starts with what Yu-kai Chou named the PBL Fallacy: the belief that points, badges and leaderboards will make a dull activity engaging. They won’t. Once you see why, you can also see when a reward does work.

What a simple reward system is

By a simple reward system I mean any design in which a user takes an action and receives a fixed payout for it, the same way every time. Two different things go by that name, and they fail in different ways.

The first is the paid incentive: cashback, miles, a cash bonus, a prize. It costs the company money and has value outside the product. People do respond to it, and its failure is the one in the first paragraph: the cost keeps rising. The second is PBL, short for points, badges and leaderboards. These cost almost nothing and have value only inside the product. Their failure is quieter. Put them on an activity nobody cares about and users ignore them.

In the Octalysis Framework, both kinds sit on the Left Brain side of the octagon. Points, badges and leaderboards mostly work on Core Drive 2: Development & Accomplishment. Prizes, miles and cashback work on Core Drive 4: Ownership & Possession, and an offer that expires on Sunday works on Core Drive 6: Scarcity & Impatience. All three rely on extrinsic motivation: you act because you want to obtain something. The Right Brain Core Drives on the other side are the intrinsic ones, where the activity is rewarding by itself.

The Octalysis octagon with the Left Brain Core Drives markedAn octagon with the eight Core Drives numbered around it. Core Drives 2, 4 and 6 on the left side are highlighted as the extrinsic drives that simple reward systems use. Core Drives 3, 5 and 7 on the right side are the intrinsic drives.12345678Left BrainRight Brainextrinsicintrinsic
Simple reward systems work on the Left Brain Core Drives: 2 Development & Accomplishment, 4 Ownership & Possession and 6 Scarcity & Impatience. The Right Brain Core Drives are 3 Empowerment of Creativity & Feedback, 5 Social Influence & Relatedness and 7 Unpredictability & Curiosity.

Why do teams start on the left? Because adding a reward to an action is much easier than making the action enjoyable. At the start that’s fine. A reward gives a busy person a justification to try something, and it gives a new user an early win. The trouble begins when, six months later, the reward is still the only thing the experience offers.

The PBL Fallacy: points, badges and leaderboards are the shell of a game

Yu-kai gave the PBL Fallacy its own chapter in Actionable Gamification. Many gamification professionals, he writes, “seem to believe that if you put points on something boring, add some badges, and provide a competitive leaderboard, that once boring product will automatically become exciting.” He calls PBL “the shell of a game experience”: the part you can see when you watch somebody else play.

Ask people who play a lot what makes a game fun and they won’t say the points. In Yu-kai’s words: “They play it because there are elements of strategy and great ways to spend time with friends, or they want to challenge themselves to overcome difficult obstacles.” The points and badges come on top of that. They’re feedback on something that already holds the player.

They’re also everywhere. Points, badges and leaderboards are, as he puts it, “elements that you can also find in boring and unsuccessful games”. Every game on the market has game mechanics, and most of those games are still boring and lose money. So the mechanics can’t be what separates a game people play for years from one they drop after a day.

For a reward system, the consequence is this: a point has no value of its own. It reports progress in something the user cares about. A chess rating means something to you because you care about chess. Put the same number on an activity you don’t care about and it reports nothing.

Take a badge for logging in ten days in a row. What does it tell the person who earned it? That they logged in. They knew that already. Yu-kai is blunt about how this lands: “It is not unusual for users to feel insulted by shallow shell mechanics.”

None of this makes points, badges or leaderboards bad design. What counts, in the book’s phrase, is “how, when, and most importantly, why these game elements appear.” Do you want a leaderboard? Then ask what it ranks, whether the people on it care about that, and what it does to everyone below the top ten. I’ve described the usual outcome in my article on leaderboards that motivate everyone: the top ten love it, everyone else ignores it, and the team concludes that gamification doesn’t work for its audience.

Yu-kai saw that conclusion coming when he wrote the book: “Hey, we tried the points stuff and it didn’t work out. I guess gamification was just a short-term fad.” A company that says this has tested the shell. It hasn’t tested a design built around what motivates its users. Our guide to what gamification is and what it is not explains the difference.

The overjustification effect: what rewards do to interest

PBL on a dull activity gets ignored. Rewarding an activity people already like can do something worse.

In 1973, Lepper, Greene and Nisbett published an experiment with nursery school children who liked to draw. One group was promised a certificate with a gold seal and ribbon for drawing. A second group received the same certificate as a surprise, and a third received nothing. One to two weeks later, during free play, the children who had drawn for the promised certificate spent 8.6% of their time drawing. In the other two groups it was 18.1% and 16.7%.

Psychologists call this the overjustification effect: when you expect a reward for something you already like doing, the reward becomes your explanation for doing it, and your own interest gives way.

One nursery school proves little. In 1999, Deci, Koestner and Ryan pooled 128 experiments on rewards and intrinsic motivation. Tangible rewards that people expected reduced how much they chose to do the task afterwards. That held for rewards given for taking part, for finishing and for performing well. Verbal praise had the opposite effect.

The effect needs interest to undermine. On a dull, routine task there isn’t any, and rewards do no harm to motivation there. I made that case in 2015, in How to use rewards in Gamification without killing motivation, against the writers who treat every extrinsic reward as poison. Is it really that clear cut? I didn’t think so then and I don’t now.

Dull tasks run into a different problem. People adapt to a reward and want more of it for the same activity next time, and the manager who hands out a bonus finds that it has to keep growing. Either way, the reward ends up as the only thing holding the behavior in place. Yu-kai describes what follows once a company stops offering it: “user motivation will often plummet to a level much lower than when the extrinsic motivator was first introduced.”

Why rewards wear off before long-term habit formation

A habit is a behavior that has become automatic in a familiar situation, and it gets there through repetition. Phillippa Lally and her colleagues at University College London followed 96 volunteers who each tried to turn one daily eating, drinking or exercise behavior into a habit, carried out in the same context each day, for example after breakfast. For the 39 people whose data fit the model well, the median time to reach 95% of their peak automaticity was 66 days. The individual estimates ranged from 18 to 254 days.

Those were simple behaviors that people chose for themselves, so 66 days is no rule for a product. What the study gives you is the scale: months of repetition, in a stable situation, with large differences between people.

So couldn’t a reward simply pay for those months? In principle it could. If the payout kept its pull for long enough, and people took the action in the same situation every day, the repetition would do its work. In practice three things get in the way.

The pull doesn’t last. The fortieth payout of 50 points doesn’t feel like the first one. To get the same response you have to raise it, add a multiplier or attach a deadline, which is where the double-points weekend comes from. Navo Orbico, the Procter & Gamble distributor I come back to below, had cash bonuses in place for its sales force, and they had stopped producing extra effort.

The behavior also gets tied to the wrong thing. Lally’s volunteers attached their action to a fixed moment in their own day. A reward campaign attaches it to an offer, and offers change. When the promotion ends, what prompted the action is gone.

And as long as the payout is the reason, the action stays a decision. The user checks, each time, whether it’s still worth it. That’s the nursery school result showing up in a product: the reward has become the explanation. An action that gets weighed every time hasn’t become automatic.

In our work for Pfizer we treated habit formation as three things that have to be there together: a trigger, the ability to act and the motivation to act. You may know them from BJ Fogg’s behavior model. A reward system works on one of the three, and only on the extrinsic half of it.

The reward was supposed to keep people going until the habit formed. A fixed reward wears off first.

One check before you design for a habit at all: does the business need the same action repeated? A hotel booking is a single decision, and our Booking.com case study covers that kind of design. Taking a medication or recording a client visit has to be repeated for months or years.

Why Duolingo’s streak doesn’t transfer

Clients often ask us for “something like Duolingo”. I understand why. Duolingo is very successful, and most people think the streak is the reason.

A streak is extrinsic, Black Hat design: it counts your progress and threatens to take it away. That threat is Core Drive 8: Loss & Avoidance, and it does get people to act today. Duolingo can lean on it this heavily because of what’s underneath. People come to Duolingo because they want to learn a language. The activity is already intrinsically motivating, and the streak keeps them on track with it. Strava’s progress tracking and comparisons work for the same reason: many people already enjoy running or cycling.

Even there the streak has a cost. I kept up a 24-day streak when I wrote my Duolingo review in 2020. At first it felt like an accomplishment. The longer it ran, the more it became something I had invested in and could lose. For most people that daily obligation starts to feel like a burden, and many users who break a long streak quit completely.

Duolingo’s own experiments point in two directions. In a post the company published in 2017, a Streak Wager let learners stake in-game currency on keeping their streak for seven more days, and Day-7 retention rose 14%. A Weekend Amulet let learners take the weekend off and keep their streak. Those learners were 4% more likely to come back a week later and 5% less likely to lose the streak. Raising the stakes worked. So did taking pressure off. When people want to do the activity, you have room to adjust the pressure in either direction.

Neither result says anything about an activity people don’t want to do. Copy the streak onto expense reports, a tax return or a loyalty app that nobody opens for pleasure, and there’s no interest under it. It adds more obligation to an activity that already feels like one.

Duolingo’s streak borrows its motivation from the learner’s wish to speak Spanish. On your product, what would it borrow from?

The same reward can undermine interest or justify it

Go back to the nursery school, because the experiment holds a result that most summaries leave out. The children who got the certificate as a surprise drew for 18.1% of their free time afterwards. That is as much as the children who got nothing (16.7%) and about twice as much as the children who had been promised it.

Same certificate, same gold seal. No damage.

What differed was the job the reward did. Promised up front, it became the reason for drawing. Arriving unannounced, it was a bonus on something the child was doing anyway. So justification works in both directions. In the overjustification effect, the reward takes over as the reason and the interest gives way. In a good design, the reward justifies the time a person spends on something they enjoy. I call that the permission slip. People who consider themselves too busy rarely allow themselves time in an experience just because they like it. Give them miles or points for it and they have a reason they can defend.

This is why points, badges and prizes have a place, and why we use them in our own designs. They do two jobs well. Early on, they create enough interest to pull someone through onboarding. Later, they justify behavior that is already intrinsically motivated. They can’t be the whole engagement strategy.

Our work for LATAM Airlines shows what that looks like in a loyalty campaign. LATAM asked whether a mileage campaign designed around member motivation could outperform its best traditional campaigns. The reward stayed exactly what it was: miles. We changed how members got to them. Standard banners made way for a treasure hunt through the Latin American jungle, with a map to explore and hidden treasures whose rewards varied, so a member couldn’t know what the next find would hold. A member who opens a static offer has all the information at once and only has to weigh it. A member looking at a map with treasures still hidden has an open question, and exploring is how they answer it. The miles made the treasures worth finding. The hunt gave members something to do on the way.

The LATAM Airlines case study reports more than 160,000 unique visits and a 123% increase in mile accumulation compared to historical averages. It was a campaign, so it says nothing about habits. It does show how much the experience around a reward changes the response to the same reward. Most loyalty programs never get that far, because the only way to take part is to spend. I go into that in my articles on loyalty members who rarely buy and on why points programs struggle.

What the reward is paid for matters too. A product that pays for every login, every post or every completed module is paying for labor. Yu-kai’s example in Actionable Gamification is being paid per drawing: you get “the quickest, unrefined drawings possible”. Most gamified communities make the same mistake, as I wrote in The 3 Rules for Community User Engagement Success. They reward liking and commenting, so people like and comment a lot and add little. Reward a contribution when it gets traction with other people, and producers take more care over what they make. A reward tied to results at least tells users how well they did. I hold that as a design position from our own projects. The 1999 meta-analysis doesn’t go that far, since tangible rewards for performing well also reduced interest there. So we pair such rewards with intrinsic design and don’t expect them to replace it.

The design has to change after the first win

What motivates a user on day 2 no longer motivates them on day 200. Most designs ignore that. The veteran gets what the newcomer got, with more of it: more points, more badges, more levels. We design the four Experience Phases separately for this reason.

In Discovery and Onboarding, a linear and predictable design is right, and extrinsic rewards do good work. We want a new user to think: this is easy, I can do this, I came to the right place. It should end in a first Win-State, something they achieved and feel proud of.

Scaffolding is the regular journey of repeated actions that follows, so this is the phase in which an action becomes a habit or gets dropped. If it stays linear, it gets stale and starts to feel like work. Here the design has to bring in the Right Brain Core Drives, and three examples from our own projects show how.

In the Navo Orbico design, each sales rep had a ship to upgrade and chose how to do it. What they developed decided how much cargo the ship could carry, so the choice had consequences they could see and adjust to. That loop of choosing, seeing the result and trying again is Core Drive 3: Empowerment of Creativity & Feedback. If a hundred users all take the exact same actions to reach the Win-State, the design has no meaningful choices in it.

The same reps shared a Port City that suffered when the team missed its targets. A stock check that used to matter only to a manager now mattered to colleagues.

And in the LATAM campaign the treasures varied, so the tenth find was not a copy of the first. When a user can predict every session, there is nothing left to be curious about.

Where it fits, we add a fourth: Epic Meaning & Calling, the sense that the action serves something bigger than the user.

The Endgame is about your veterans, who know every rule and have collected most of what there is to collect. Another prize does little for them. A role does more: mentoring newcomers, setting challenges for others, being the person people follow.

The rewards stay through all of this. Their job changes, from the reason to act to the permission slip. This is the part companies most often miss, or find hard to build.

Two projects where the action had to last

Yu-kai Chou, the creator of the Octalysis Framework, and I founded The Octalysis Group in 2013, and we’ve designed for more than 175 companies since. Two of those projects had exactly the problem this article is about: an action that had to be repeated for months or years, and a standard tool that wasn’t getting it done.

Navo Orbico, a Procter & Gamble distributor: the cash bonus had stopped working

Navo Orbico distributes Procter & Gamble brands such as Ariel and Pampers and operates in 19 countries. Cash bonuses were in place for its field sales force and were no longer producing extra effort. The selling got done. CRM entries, stock checks and training were neglected.

Why? A bonus paid at the end of a period has nothing to say to a rep doing a stock check on an ordinary weekday.

The options were to adjust the incentive or to change what the daily work was like for a rep. We took the second. Each rep became the Captain of a trading ship, in an experience built over the CRM the company already used, so progress came only from real sales work. Hitting KPIs earned health points that protected the Port City the whole team shared. A rep who had a weak period always had a way back: help a colleague with a sale, take a training, or give one. We use the threat of a loss to create urgency, and we give people a straightforward way to recover. A streak that resets to zero offers none.

Our P&G sales gamification case study reports a 60% improvement in KPIs for sales input activities such as up-selling and accurate reporting, which are the daily actions the bonus had failed to reach. Social interaction among the sales staff rose 300%. Revenue increased 28.6% against historical performance, and 99.5% of eligible employees joined a rollout that was not mandatory. It’s the question behind most of our employee engagement work, and behind digital transformation programs too: how do you get people to want to use a system the company has already paid for?

Pfizer: a daily action that has to last for years

Our Pfizer case study reports that 68% of users were still active every day after six months. Six months is close to three times the 66-day median in Lally’s study.

Medication adherence is habit formation at its most demanding: the same action every day, for as long as the treatment lasts. With Pfizer we worked on it for patients with ulcerative colitis and Crohn’s disease. The standard tool is a reminder, and a reminder is only a trigger.

A patient who dismisses the notification did not forget.

Sidekick, the app in the case study, gives patients more to do than confirm a dose. They set their own goals and shape their own health journey. They track sleep, nutrition and mood next to the treatment, so they can see what adherence does for the rest of their life. Mentors and peers mean nobody manages the illness alone. And look at what the rewards are: unexpected health tips and success stories from other patients. Nobody promises them in exchange for a dose. They turn up, like the certificate in the surprise group. Alongside the 68%, the case study reports 30% fewer patient dropouts. This is the approach we take across health and wellbeing projects.

When your rewards cost more every quarter

Back to the company from the first paragraph, the one paying more each quarter for the same activity. What would we look at first?

We’d start with what the reward is paid for. Is it labor, such as a login or a visit, or is it a result? Then the business metric and the Desired Actions behind it. In almost all of our projects, clients find out at this stage that they were designing for the wrong thing. Then the Player Types among the users, because not everybody is motivated in the same way and a single flat payout suits only some of them. And then the question this whole article comes down to: is there anything in the activity for the user once you take the payout away?

Usually the gap is in Scaffolding and the Endgame, where a user with two hundred sessions behind them has nothing to do that a new user doesn’t have. From there we design what is missing: choices with consequences, things to do with other people, surprises, and roles for veterans. The extrinsic design that is already there stays.

Leaving it as it is has a predictable cost. The next campaign will need a bigger payout than the last one, and the users it holds are loyal to whoever pays the most this week. If that describes your program, Contact Us and tell us which action you need people to keep taking, and what you pay for it today.

Frequently Asked Questions

What is the PBL Fallacy?

The PBL Fallacy is the belief that adding points, badges and leaderboards to a boring product will make it engaging. Yu-kai Chou, creator of the Octalysis Framework, named it in chapter 2 of his book Actionable Gamification. The same mistake applies to other game elements: he notes that levels, boss fights and quests fall into the same hole, and that calling a task a “quest” does not make the task fun. Game elements only work when they are chosen for the motivation behind a specific action.

What is the overjustification effect?

The overjustification effect is the drop in interest that can follow when people expect an external reward for something they already enjoyed doing. It takes its name from a 1973 experiment by Lepper, Greene and Nisbett. Two conditions have to be met: the person was interested before the reward, and the reward was expected. In the same experiment a surprise reward did no harm, and in a 1999 meta-analysis of 128 experiments verbal praise increased interest.

Do points, badges and leaderboards ever work?

Yes. They are mechanics, neither good nor bad. Points and levels give users a measure of their progress, and people want to see progress. The LinkedIn profile progress bar is an element from game design that nobody would call unserious, and it works well. In our designs, points and badges do two jobs: they create enough early interest to carry someone through onboarding, and they justify behavior that is already intrinsically motivated. They fail when they are the only thing the experience offers.

Do streaks build long-term habits?

A streak creates urgency through Core Drive 8: Loss & Avoidance. It can support a habit when people already want to do the activity, as with language learning on Duolingo. On its own it builds pressure. A streak that resets to zero treats one missed day as a total loss, while in Phillippa Lally’s 2010 habit study a single missed day made no material difference to how the habit formed. If you use a streak, give people a way back. Duolingo added a Streak Freeze, and in our Procter & Gamble distributor project a rep could recover from a weak period by helping a colleague or taking a training.

Already have a project in mind? Let’s work together!

Our team will listen to your engagement challenges and decide if we are the best match.

Get in touch

Case studies

Measured results from our client work

See all 14 case studies

Hospitality

Hospitality Conversion Gamification

The Octalysis Group created a new industry standard for hospitality and travel booking platforms.

33.4%Increase in gross bookings

Loyalty

Gamified Loyalty Program

Gamified loyalty experience with personalized missions and rewards, leading to record-breaking engagement and conversions.

+153%Increase in credit card acquisitions

Fast-moving Consumer Goods

Sales Force Engagement

Gamified sales platform transforming reps into virtual traders, boosting engagement and performance across 19 countries. Best Gamification Award.

+28.6%Increase in sales revenue