Last week, we presented a handy reference chart of our Gamer Motivation Model based on our data from over 140,000 gamers. Here it is again for ease of reference.
Want to know how you compare with other gamers on these motivations? Take a 5-minute survey and get your Gamer Motivation Profile.
As we mentioned, motivations in the same column tend to be correlated, while motivations in different columns tend to be less correlated. But that doesn’t mean that each column is completely unrelated to each other. And indeed, some columns are more related than others. So wouldn’t it be nice to get a visual map of these motivations that could capture these relationships better? Luckily for us, there’s a statistical method for that.
Wouldn’t it be nice to get a visual map of these motivations?
Multidimensional Scaling (MDS) is a technique that compresses the distances between a set of variables into a 2D map while preserving the original distances as much as possible. Variables that are more correlated are put closer together, while variables that are less correlated are put further apart. When we applied this technique to the motivation data, we found a higher-level structure to the motivations.
At a high-level, 3 clusters of motivations emerged.
- In the bottom-right orange cluster, there’s an Action-Social cluster that combines the interest in fast-paced gameplay with player interaction.
- In the left yellow cluster, there’s an Immersion-Creativity cluster that combines the interest in narrative, expression, and world exploration.
- In the top blue clusters, there’s a Mastery-Achievement cluster that combines the appeal of strategic gameplay, taking on challenges, and becoming powerful.
We found motivations that act as bridges between these major clusters.
- Discovery is a bridge between the Immersion-Creativity cluster and the Mastery-Achievement cluster.
- Power is a bridge between Action-Social and Mastery-Achievement.
We didn’t find a bridge between Immersion-Creativity and Action-Social. This map might be hinting to us that there should be something here. On the other hand, there’s no reason why motivations need to fall neatly into balanced models where everything is bridged.
Think of it as a proximity map.
What the map is showing is the relatedness between all these motivations. So if someone scores high on a particular motivation, they are more likely to score high on the nearby motivations.
The opposite isn’t true though. Motivations that are farther apart are independent of each other; they don’t suppress each other. For example, gamers can score high on Action-Social and high on Mastery-Achievement.
The axes point to 2 primary dimensions on which motivations vary.
The axes produced by Multidimensional Scaling are not always easy to interpret. In this case though, we think they point to two interesting dimensions.
- Acting on the World vs. Acting on Other Players: We think the horizontal axis reflects a distinction that Richard Bartle made in his Player Types model. The left side of the map emphasizes action on elements of the world and its narrative, and the right side emphasizes interaction and action on other players.
- Cerebral vs. Kinetic: In terms of the vertical axis, we think the top part of the map emphasizes careful, long-term oriented gameplay while the bottom of the map emphasizes dynamic, fast-paced gameplay.
This motivation structure is consistent across all the regions we have data for.
We found this exact same structure across our data from North America (N = 81k), South America (N = 7.6k), Western Europe (N = 13.8k), Eastern Europe (N = 6k), Southeast Asia (N = 14.3k), and Australia + New Zealand (N = 6.3k).
We’ve been hinting at this structure in our web app and reference chart.
This higher-level structure is the rationale behind our layout of the radar graph on the Gamer Motivation Profile, and the ordering of the motivations on the reference chart. In both, the motivations that are more related are placed closer together.
Stay Tuned
In our next blog post, we’ll describe how these 3 high-level clusters intersect with personality psychology and explain why the unintuitive Action-Social cluster actually makes a lot of sense.
[…] Whatever feature you have come up with, they will have to make it happen down to the smallest detail, and in doing so they will be carrying the design mantle that you hand off once you have outlined the feature. Be precise when discussing implementation with programmers. Gaming Motivations Group Into 3 High-Level Clusters – Quantic Foundry. […]
Perhaps between Immersion-Creativity and Action-Social is Self-Expression—the freedom to create and the desire to show to others. Think of the mod community—Garry’s Mod, Elder Scrolls, Halo Forge, and The Sims. In these games, players design their own worlds and fantasies and then ask other players to discover them. In doing so, they act both on the world and on other players’ expectations.
And there are both elements of cerebral play and kinetic play—creating is cerebral, but you never know how other players will react to it. This interplay/kinesis blurs the lines between creators, co-creators, and players. Sometimes, these games/mods spontaneously develop their own challenges, strategies, and even traditions. Garry’s Mod, the popular “world-building” mod of Half-Life 2, has enabled the emulation of nearly every game genre.
Self-Expression would not be an obvious bridge between Immersion-Creativity and Action-Social for a few reasons. For one, it may seem insignificant because few games cater to this audience. However, the few games that do have very devoted members. After all, they must go through the trouble of modifying other games first.
Furthermore, measuring self-expression as a motivation may be difficult because self-expression is an primary characteristic of the game medium. Unlike other media like books or film, games require people to actively manipulate the world and other players—and that performance is a form of self-expression. Because self-expression is an inherent quality of games, it would therefore seem invisible.
Hello Nick
Of course principal components analysis and its variations fundamentally depend on the sample. Bad, biased or incomplete sampling mostly leads to artefacts.
How did you get the 140k players’ data ?
Do they all play the same game (f.ex Wow) or the same kind of games (MMO) ?
Is some kind of games excluded from the sample (f.ex CS or WoT) ?
There must be Something wrong with what is called Community and its proximity to notions like Destruction and high “kinetics”.
Generally people who like chatting, helping and social interaction (aka role playing) are not destructive or achievement oriented.
On the other hand people that destruct and play high kinetics for excitement (aka PK) are not chatting or helping and their social interaction uses only one style – destruction and griefing.
I do not find in this model these 2 typical kinds of primarily socially oriented players that everybody knows on an MMO.
On one hand the helping, guiding and chatting guy.
On the other hand the KSing,, PKing and cheating guy.
Even if both share a high interest with player interaction (e.g they are “right side” of the diagramm) they are very far apart in what motivates them to do what they do.
I couldn’t believe seriously a model that would tell me that helpers and PKs are the same thing and if it happened I would be sure that there is a missing dimension which would separate them if it was taken in account.
The sampling frame was people who play video games and wasn’t targeted at gamers of specific titles or genres. You can see the detailed sampling notes here: https://quanticfoundry.com/the-v2-sample/. Also see this post on the favorite games by motivation to get a sense of the diversity of games listed by respondents as their favorites: https://quanticfoundry.com/2015/08/11/most-popular-games-by-gaming-motivations/.
Your notes on RP vs PK are highly MMO-specific. It’s interesting that you express concern about genre bias in your first paragraph, and then apply MMO stereotypes to all gamers in your second paragraph. You’re also conflating a lot of different motivations in your stereotypes of RP and PK even within MMOs. Many MMO gamers socialize without ever role-playing (an activity that is now so niche in MMORPGS that only a small handful of WoW servers are designated as RP). And there are plenty of MMO gamers in raiding guilds for whom socializing and achievement are equally important. Finally, excitement-seeking is not the same as griefing/PK-ing. For example, spawn camping or stalking can be slow, strategic, and precise rather than kinetic and chaotic.
The motivation map shows how motivation factors are often correlated. It shows the relationships of motivations rather than showing buckets you can put players in. Motivation factors are configural. And correlations are just that: correlations. The map isn’t deterministic at the individual level. So someone can score high on Community and high on Fantasy, but low on Competition (in an MMO, this might be an RPer). Or someone can score high on Competition, Destruction, and Power, but low in Community (in an MMO, this might be a PKer).
The 12 motivation factors are facets of each gamer. And a gamer can score high or low on any of those factors. What the map shows is a 50,000 feet view of how these motivations are often related to each other among gamers in general. Or using a food analogy, when you survey all cuisines, bread and butter are often eaten together, but it doesn’t mean this is the only way to eat bread (e.g., bread is almost never eaten with butter in traditional Chinese cuisine).
I am curious what percentage of people who completed the survey fall predominantly in each of these cluster as it might give another interesting take on the primary motivation of players to pick a particular game, and provide a better way to define game genres compared to the mechanics-dynamics-aesthetics system (which gives an arbitrary feeling set of aesthetics but not a motivation for them)
Getting the percentages for each of the 12 categories might actually be less informative than the three clusters, though a chart of pairings, the percentage of players that score high n any of the possible combinations of motivation factors, might be quite valuable to game studios. Especially if there a bridge cases like ’10 percent of the players who score high on the design motivation also score high on competition’ as those would point at a potentially under-served genre.
Hi Marian. This is exactly one of the services and insight reports we can provide to gaming companies. See: https://quanticfoundry.com/audience-profiles/
Because we also have game title data along with the motivation scores, we can provide benchmarks for motivation combinations. So not only can we say there’s an untapped opportunity in “Design + Competition”, we can specify that this intersection is at “the level of Design in Game X/Y/Z with the level of Competition in Game A/B/C”.
[…] Previous […]
[…] A visual map of gaming motivations based on our data from over 140,000 gamers worldwide. […]
[…] data from 220,000 gamers who completed a survey about what motivates them to play computer games. Motivations clustered into 6 groups: action, social, mastery, achievement, immersion, and creativity. Obviously, different games speak […]