Gamification is not a game

Gamification is not a game because it usually imports the parts of games that are easiest to see—points, badges, leaderboards, streaks—and leaves behind the parts that make a game work: choices that matter, outcomes that are genuinely uncertain, and the fact that nobody has to play. What remains is a scoring system attached to an activity that was already there, often an obligation. That can be useful, but it is a different thing, and calling it a game hides the difference.

What the word was meant to mean

The most careful definition of gamification comes from a 2011 paper by Sebastian Deterding, Dan Dixon, Rilla Khaled and Lennart Nacke, written as the term was spreading through the technology industry. They defined it as “the use of game design elements in non-game contexts.” The definition is deliberately narrow and neutral. It says nothing about which elements, and nothing about whether the result is good.

What makes the paper useful is the map it draws around the term. Deterding and his co-authors place gamification on two axes. One runs from whole to parts: is the thing a complete game, or does it merely use elements of games? The other runs from gaming to playing, a distinction borrowed from Roger Caillois: rule-bound, goal-directed play on one side, free and open-ended play on the other. That produces four neighbors. A complete game built for a purpose beyond entertainment is a serious game. A complete thing for free play is a toy. Elements of free play added to a non-game produce playful design. And elements of rule-bound games added to a non-game—that is gamification. The paper’s title pairs the term with gamefulness, the experience those elements are supposed to produce, and that pairing is worth remembering.

The distinction between a serious game and gamification is the one that matters most here, and it is older than the word. Clark Abt’s Serious Games argued in 1970 that complete games could teach and train. Gamification proposes something else: that you can keep the activity as it is and add the game on top.

The visible parts

In practice, the elements that get added are a short and familiar list. Kevin Werbach and Dan Hunter, in For the Win, named the most common combination “PBL”: points, badges and leaderboards. To those, consumer software has added the streak.

The pattern is easy to recognize across very different products. Nike+ (2006) put a sensor in a running shoe, sent the data to a website, and turned runs into logged distances, goals and challenges against friends; later Nike hardware converted activity into a branded points currency. Stack Overflow (2008) awarded reputation points for upvoted answers and badges for particular behaviors. Duolingo (2012) gave language lessons experience points, weekly leagues with leaderboards, and above all a streak: a count of consecutive days on which the learner completed a lesson, which can be protected, repaired or lost.

These elements were chosen for understandable reasons. They are the most visible layer of a game, the part that shows up in a screenshot. They are cheap to build, because they are mostly counters. And they are measurable, which matters to organizations that need to show a number went up. But they share a property that the enthusiasm around them tended to skip. Points, badges and leaderboards are feedback. They are a game’s report on what happened during play. They are not the play.

Points are what a game says about play. Gamification often keeps the report and drops the play it was reporting on.

Margaret Robertson made this point in 2010, before the term had settled, proposing that much of what was being called gamification should be called “pointsification.” Her argument was that points and badges are among the least essential parts of a game, and that a system built mainly from them had borrowed the scoreboard and left the sport behind.

What gets left behind

If the scoreboard is the visible layer, what is underneath? Three structures do most of the work, and gamified systems most often lack them.

Meaningful choice

In a game, the player makes decisions, and those decisions change what happens. The points that follow are interesting because they reflect choices: a high score in a puzzle game records a series of good decisions under pressure. A Duolingo streak records one binary choice repeated daily—did you open the app and finish a lesson? The streak is indifferent to how you learned, what you struggled with, or whether you chose an easier lesson to keep the count alive. It measures attendance.

Uncertainty

A game’s outcome is not known in advance; Caillois counted uncertainty among the defining qualities of play. Most gamified systems are designed to be certain. Do the thing, get the points. Complete the profile, unlock the badge. A progress bar that always fills at the same rate is a receipt, not a challenge.

Voluntary play

Bernard Suits defined playing a game as “the voluntary attempt to overcome unnecessary obstacles,” and both halves of that phrase matter. Gamification is typically applied to activities whose obstacles are entirely necessary—learning a language, getting fit, doing a job—and often to activities that people are required or pressured to do. The attempt is not voluntary and the obstacles are not optional. The scoring layer cannot supply the attitude Suits described; at most it can reward people for persisting.

That is where the critics landed. Ian Bogost’s 2011 essay “Gamification Is Bullshit,” published on his website and republished by The Atlantic, used Harry Frankfurt’s philosophical sense of bullshit—speech indifferent to whether it is true—to describe gamification as a marketing term that claimed the authority of games while using their least important parts. Earlier the same year he had proposed a blunter name, “exploitationware,” for systems that use the appearance of games to extract labor or attention. Bogost’s own work on persuasive games makes clear that he was not objecting to games being used for purposes; he was objecting to the claim that points constitute a game.

There is a psychological version of the same worry. A 1999 meta-analysis by Edward Deci, Richard Koestner and Richard Ryan found that expected, tangible rewards tend to undermine intrinsic motivation for tasks people already find interesting, while positive feedback did not have that effect. Gamification sits awkwardly across that line: it presents itself as feedback, but it is often built and experienced as a reward scheme.

Three systems, and why they work differently

None of this means every system with points is a failure. It means the points are rarely the reason a system succeeds.

Stack Overflow is the clearest case. Its reputation score works, where it works, because it is not pretending to be a game. Reputation comes from peer votes, so it measures something the community actually values, and it unlocks real privileges in the site’s moderation system. It is a trust and permission system with a scoreboard on top, and the choices it reflects—which questions to answer and how well—are real ones.

Foldit, released in 2008 by researchers at the University of Washington, is the opposite move. It is not gamified protein science; it is a game made out of protein science. Players manipulate the three-dimensional structure of a protein, and the score reflects how plausible the structure is. Every structure is a genuine choice, nobody knows the best answer in advance, and nobody is required to play. In 2010 its developers reported in Nature that players could match or outperform the lab’s automated methods on some problems, and in 2011 Foldit players produced a model that helped researchers solve the structure of a retroviral protease that had resisted them for years. That is a serious game in the full sense: all of a game, put to work.

Running offers a smaller contrast. Nike+ logged runs and set goals. Zombies, Run! (Six to Start, 2012) did something structurally different: it turned a run into an audio story in which the runner collects supplies and must speed up when pursued. It added unnecessary obstacles, uncertainty and a fiction to an activity, rather than only counting it. That is closer to what Deterding and others mean by gameful design: designing for the experience of playing, not for the appearance of a score.

The empirical record supports this caution. A widely cited 2014 literature review by Juho Hamari, Jonna Koivisto and Harri Sarsa found that most studies reported some positive effects from gamification, but that the effects depended heavily on context and on who the users were, and that many of the studies were methodologically limited.

The honest label

The early enthusiasm drew partly on a larger argument. Jane McGonigal’s Reality Is Broken (2011) made the case that games are unusually good at producing engagement, cooperation and a sense of meaning, and that the world could learn from them. That argument is defensible. But the lesson it points to lives in the structure of games, not in their scoreboards. Deterding himself has argued that what designers should take from games is the way they structure motivating experiences—goals worth pursuing, challenges calibrated to skill, freedom to choose—rather than their reward systems.

So the useful question for any gamified system is not whether it is a game. It almost never is. The useful question is which game structures it actually contains. A streak counter has goals and feedback. It rarely has choice, uncertainty or play. A system that admits this can be designed honestly, as a habit tracker or a reward scheme or a reputation ledger, each of which is a perfectly good thing to be.

The problem with gamification is not that it uses parts of games. It is that it treats the parts that are easiest to count as the parts that matter.

The companion essay Where the game lives turns this into a general lens: six structures, and a way of asking which ones any game-like thing has and which it lacks. The same export of visible parts happened inside games too, when progression systems spread from role-playing games into nearly every genre; Why “RPG” stopped meaning anything traces what that did to a genre label. In both cases, the part that was easiest to copy turned out to be the part that explained the least.