From Rational to Emotional: Designs that Increase Player Retention

I'm going to go ahead and get started. So, um everyone, welcome. Uh thank you for showing up today. Uh good morning, good afternoon, depending on your state of jet-laggedness. Is that a word? Can I use jet-laggedness? Um please remember to turn off your cell phones and all that other fun stuff that we typically tell you before um one of these talks. My name is Jim Brown. I am from Epic Games. And my talk today is rational to emotional designs that increase retention. And the word rational generally has very positive connotations, which might lead people to believe that if they're making rational decisions, they are therefore making good decisions. Uh and that's not always the case because we're actually very often influenced by powerful emotions first that we then try to rationalize the why after the fact. And as I talk about emotions today, I want to be very clear that I'm not talking about ways to make players happy

or sad. Those are feelings. Um emotions are those core parts of us that are physical and instinctive. They're programmed into us after years and years of evolution. And that makes them more universal to us as humans as opposed to the feelings which make it which are more unique to us as individuals. And emotional design therefore gives us the capability to reach a broader audience in a deeper way. So, I'll start today with a brief introduction to what emotional design is cuz not a lot of people are really familiar with it, I believe. And then roll into some more concrete examples of how you can use it in your designs to help improve re- retention. While he likely doesn't need too much introduction with this audience, I hope. Um I do want to cover my bases. So, this dapper fellow is of course Don Norman, uh UX pioneer and icon. And he actually has a book about emotional design where he describes the

specific design elements that help us form attachments to things and how those attachments can help or can influence our experience using them. He says that good emotional design has three major components: the visceral component, the behavioral component, and a reflective component. Now, the visceral, of course, is your gut reactions or how you respond to a stimulus. Uh the behavioral asks, is something actually useful? Does it perform in the way it's supposed to? And can it, through its use, somehow influence or change your behavior? And then the reflective component is does it make me think about it, especially after it's gone? Do I think about it, uh the object that I was using when it's no longer there? Now, looking at each of these individually, uh that visceral component is really, really easy for us to get as a designer. But visceral reactions tend to be very unpredictable. And as an example, I give you this.

I can almost guarantee that everyone here has some sort of visceral reaction to seeing this image. But the problem is I can't tell whether you're going to respond positively or negatively because those reactions tend to be very personal based on your own context and experience. Visceral reactions are also very transient. So, even though I've left this up here while I've continued to talk, it's probably even in that short time any emotional response that you had has probably started to fade away. So, as a designer, you need to be very careful with your use of visceral design. Monster closets stop being fun. Large explosions stop uh stop having the impact that they had. And the the transient and unpredictable nature of these visceral reactions means that alone they aren't very good for retention except that they are incredibly important for making first impressions.

So, visceral designs can be a crutch that you can use uh to get people interested early or to obtain very short-term retention goals like, "Hey, come back the next day." Um while you rely on other methods for longer-term retention. Uh by contrast, game systems are specifically designed to affect the player over long periods. As time goes on, the systems can affect player emotion and that can in turn affect their behavior. And the really cool thing about game systems is that they can then respond to those changes, which resets or continues the cycle. And this sets up a feedback loop, so the game system actually has the ability to respond to both behavioral and emotional state changes. As an example, uh spree systems tend to have very good emotional design. In many games, including Epic's own uh

Paragon and Unreal Tournament that I have up here, an announcer will come on and say things like, "Oh, killing spree." or "Your power grows." These messages are feedback that communicate how well you're doing in a match, but they also serve a much larger purpose. When you're on a spree, your heart rate rises, that adrenaline starts to pump. And this is your visceral emotional reaction. And since you're doing well, you might push the attack on an enemy that you would not have otherwise. And that is a behavioral reaction. In fact, by broadcasting your spree to everyone in this in the match, this system can also affect teammates, maybe by inspiring them, or enemies by intimidating them. Meaning that your killing spree has the potential to affect everyone's behavior in the entire match. The system can then respond to that spree by making you a more high-value target or various other systems that people use, which encourages further

behavioral and emotional change. And the design of these systems is so effective that this ultra kill t-shirt has become one of the top selling items in our online store, showing that the spree system is also very reflective, playing into the last of Norman's components for good emotional design. And while these components are all interesting individually, it really gets interesting when they start to come together all at the same time in one single design. And that's what Norman calls providing a emotional context. If you think about your typical game interaction, the you start with this kind of physical influence that's imposed by the input device, mouse, keyboard, VR helmet, touch screen, whatever it is that you're using. And that determines the boundaries of what you can do. But then that logical or rational side of our brain kicks in and tries to figure out what we should be doing in

the face of those physical limitations. But at an even higher level, what we can do, the physical, and even what we should do, the rational, often play second fiddle to what we want to be doing, even if we can't explain why. And this is why I think emotional design is so important. By designing for emotion, we can reach that part of us that is universal to us as human beings and create a lasting impression that people want to come back for over and over again. But what does emotional design look like in actual practice and how does that lead to retention? Retention can actually manifest manifest itself in a myriad of different ways, whether it's um number of matches played, hours played, or days played in a row. It could be is the user spending money or engaging with the game systems,

combinations amongst all those things. The one thing that they really have in common is that sense of attachment. And if you can recall, the purpose of uh emotional design is to give the user an emotional context that causes them to form attachments. So, I want to go over some of the factors that might influence attachment today and discuss them alongside some examples from the games that I've worked on. And the categories uh categories I want to talk about in general are memory, patterns, feedback, and mental models. First on that list, of course, is memory. If we can design systems or games that stay in someone's memory, they're more likely to to be reflective and therefore more likely to retain in the longer term. But the brain isn't a computer that stores and retrieves information.

So, memories can't just be programmed into them. Uh we sometimes talk about brains that way. I even do it a lot in this presentation, but that's more just a shorthand that we use. The reality is that the brain is actually more akin to a system that is affected by and responds to stimuli. And we are constantly awash in an avalanche of feedback and input from our surroundings, from from everywhere. So, the way that we compensate for this is that we filter. And memory is therefore tied to attention. Where we choose to focus our attention affects uh what we will remember. But attention is, of course, a very limited resource. And when you're focused on playing a game, your brain is locked in this battle to prioritize what's happening in the moment, what's happening in the story, what are your short-term goals, what are your long-term goals.

And so, the peak-end rule is a very good way uh to help you determine when is a good time to try and focus the player's attention so that it has the most lasting potential. When we look at something that's happened to us, our brains kind of look at how we felt at the peak and the end of that experience. And over time, those two points are much more likely to be the brain's reference for our overall emotional experience. So, if you can focus on making the peak and the end the most vivid, the most meaningful to the player, then they become the most available and lasting to long-term memory. And this is one of the reasons that people can suffer through extremely difficult encounters or long levels or difficult quest chains and still come away from that experience feeling very good. Quick example from Gears of War.

The sequence shown in this shot was, for many of our players, the peak experience of the game. There's this cool dramatic intro that led to an extremely difficult combat sequence where we introduced a key new enemy for the first time. And the line at the end of that intro, which you can see up here, "Look at all that juice," was funny, it was memorable, and it actually did become a meme that we saw resonating through the internet for a while. It showed up in other games in forms of achievements and all this other craziness. And during development, we knew that this was going to be a key moment in the game. And we didn't want players to miss that funny, we didn't want them to miss the drama, and we definitely didn't want them to miss the introduction to this new enemy cuz it was a key part of the game. So, we decided to put a checkpoint right here. But that was a very rational decision. On paper, it sounded smart, it sounded reasonable. In the end, this actually turned out to be one of the biggest drop points in the entire game.

Even years later, there is a thread on NeoGAF that discusses games with the worst checkpoints of all time. And you can guess what that thread is called. Look at all that juice. This sequence had a very strong rational design, but poor emotional design. It showed no empathy for players stuck in a very difficult part of the game and no respect for their time as they had to repeat this over and over and walk through this long sequence and watch that same cinematic and go through that same fight every single time they died. And my favorite uh post in that thread says, "It's amazing that game designers still don't get things like that." And they're right, because it's so easy for us to rationalize a poor decision and forget the power of emotional design. Compare that to a more recent example that we're working on in our game Fortnite. Instead of generic loot crates, Fortnite

uses llama piñatas. And as one does with a piñata, the way you open it is to beat it with a stick. So, there's this really strong affordance for hitting it, which uh is a very strong visceral reaction. And when you pull back and you're about to unleash your attack, at the peak of this experience, the llama reacts, inciting a behavioral response from you, because now you want to hit it even harder. Or maybe you feel sorry for it, whatever your your your deal is there. Um the piñata theme, the the actions, and of course, the end result getting all the loot, help players remember that it's more than just a loot crate, making an experience that players reflect on. But opening this thing is a really long sequence, and you have to use this broom to clean up all the pieces at the end. And you do it pretty regularly every time you earn something. But then up in the top right, there's also a fast-forward button that shows empathy and respect for the player's time.

Now, each of these sequences from the two games are memorable, but for opposite reasons. Both incite an emotional response from the player, but Gears does it in a way that pushes the player out of the game experience and it takes their attention away from what they're actually trying to do. While Fortnite grabs the player's attention at the peak and the end of the experience, focuses on the positive emotional context and creates a lasting memory that you want to do this over and over again. And that's important because memory is biased toward these emotional events. Now, the easiest and most obvious way to capture someone's attention is through the use of simplicity. Simplicity means focus. Simplicity means elegance. Simplicity means clarity. And when done well, simplicity has a

real emotional power, and that alone can be quite memorable. And I know that talking about simplicity with a group of designers, which I saw many of you were when we started this, it's like this basic design concept that you're all like, "Yes, I understand that." But again, that's very rational thinking. In actual practice, you know there's always that executive who wants you to add a monster to the door cuz it'll look scarier, right? And then the art director comes in and wants you to add some better shaders and some other elements to really show off their cool art techniques. And even just these small differences are striking. I mean, that image went from something that you were staring at silently to I hear people giggling, right? Simplicity has become complexity. And that really detracts from that overall visceral component that you originally had with the simple image. So, as a designer, you need to be able to find a way to compromise uh in a way that doesn't sacrifice that

important component of your design. Without that emotional pull, the player isn't likely to notice the things that you want them to focus on, which means they're less likely to remember, so they're less likely to retain. Stories are also a very uh important component to memory. This doll was part of a research study called the Significant Objects Project. Researchers bought it just randomly off of eBay for $3. And then they made up a background story about it, about how it was infused with the spirit of the doll maker, and on moonless nights, it would come come alive and dance in the streets. It was a complete fantasy. But then they put the doll back on eBay. But instead of selling for $3, the original purchase price, it now sold for almost 200. The story made the buyers think about the doll. It became significantly more reflective. And stories lead to attachment because

they create actual connections and associations in our brain, actual measurable physical changes. And this is why in Paragon, we include very simple, very vague stories, background stories, about each of our characters. The story encourages players to think about the hero, to form an association with them. And then as they start to play, their own personal story gets all mixed up with the character's background story, and the two start to merge. And because those stories are physically part of our player's memories, they're more likely therefore to stick with playing them and to become attached to them, therefore retaining over longer periods of time. In Fortnite, bacon is used for healing. And sometimes it can be found in toilets. Now, this decision to create toilet bacon was actually pretty deliberate.

First off, bacon heals you. It's also delicious. So, it serves a useful purpose that benefits you, therefore it fulfills that behavioral component. And when a a new uh item comes in your inventory like, "Hey, this this thing equals health," it lights up uh pathways in your brain that are associated with cognition and learning. But when that item is also something like coffee, or in this case bacon, it'll also light up the pathways that are associated with your olfactory response. And that is something that is will give you a very instant visceral reaction as well. So, in addition to bacon, you have another and hopefully different, but just as strong association to the word toilet. And we now have these two concepts where you have an attachment to both of them,

and we've created a link between them. And your brain knows them, your body knows them, so they're really easy targets for our emotional design. And lastly, bacon is actually a very rare find in Fortnite. So, when it pops out of a toilet, it's surprising, it's funny, it grabs your attention at a peak moment, and it's memorable. So, it creates a story that you want to share. It's therefore become reflective. And all of these components combine to create a small strong emotional design that is more likely to retain players. And lastly, memory is important to retention because the emotional uh connections that it creates can be very, very enduring. If you think about the most recent Doom game that came out, when you look at it in isolation, that multiplayer beta that came out first, it was a pretty cool experience, but critically, it was panned and the community started to get out their pitchforks

because it didn't feel like the Doom of their memory. And then a few weeks later, the single-player campaign comes out and all of a sudden, bam, this game is a major hit because that single-player component of the game matched very well with the players' memories of the past Doom games. That original Doom is now ancient by video game standards. But the memories that you have from playing it will last you a lifetime. And whether or not that game actually holds up by modern standards is irrelevant as long as the new game that is built on those modern standard measures up to your memories. And if it does, that means that players are more likely to keep playing. So, if we can grab the players' attention, focus it at the right time and place, and let it develop into something that they can actually associate their own personal stories and personal emotional context to,

then we can create a design that stays in someone's memory and they're more likely to retain. Okay. Next uh component I'd like to focus on is patterns. Patterns help us establish a sense of continuity. Instead of action leading to reward, you can also get reward that leads to further action. Every end point becomes a new starting point for the continuity of the pattern. And this helps us form habits, and habits lead to retention. Take a moment and think about social media that you use, whether it's Facebook, Twitter, Instagram, whatever. Now, I highly doubt that you can recall every single time you've ever logged into Facebook and checked a post. I bet you can barely even remember what you did yesterday. But just me mentioning it, I'd also be willing to bet that several of you have itchy hands that want to reach for your phones.

Because that pattern of daily use has turned into a habit or ritual behavior. Now, Rain Man is a great reminder that experiences that develop from patterns are also much more memorable. I just finished talking about how memory is important to retention. And this is because repetition is a key component of the process that helps us encode information and move it from short-term to long-term memory. Yes, Ben, wherever you are, I know that's shorthand for not what actually what happens. But this is why games offer daily rewards, timed rewards, or notification systems that invite you back as often as possible. Some even have significant components of their game that require ritual visitation in order for the player to be successful. These types of patterns set up strong uh uh short and long-term goals and help players track their progress. Log in tomorrow and get a reward. Gain

one more level and you'll get a custom skin. The more you come back, the better you become. And the more you're building more of those physical connections in your brain that make it easier for you to access the memories associated with the game. Streamers also rely on ritual behavior to en- enhance their viewer retention rates. If you watch Hearthstone or Paragon streams, they'll take the simple process of opening a card pack and draw it out. There's a whole button there that says reveal all and they all ignore it. They'll hover over each of the cards individually to find out their rarity and then they'll start with the most common card and work their way up slowly to the big reveal at the end. There's no rules that say you need to open card packs this way. And yet, in almost every single instance on YouTube and Twitch that you'll find, people do it in this exact same way. They follow these patterns because they know it's likely to grab your attention

and burn into your memory. They're betting on the fact that their pattern behavior will lead to better viewer retention. And if you want to develop this kind of ritual behavior, you need to give your player repeatable actions. This is the entire foundation for Paragon. You load into a game, you select an ability, you select a deck, you add a card, and you set off to lane. It's the same sequence and it repeats every time you play. The game is separated into a pre-game phase, laning phase, a mid-game, a late-game. Every phase is a pattern of its own and repetition builds memory. After the match, you see the the reward screens, you upgrade your deck, and then you head back in and start all over again. Every aspect of the game is a pattern that induces ritual behavior that makes players want to come back for more.

I was too big of a chicken to get video, but here's a character that's fighting some AI minions. And when you see them in game, the speed of the animations and the sound that it creates is a very strong pattern. It establishes a rhythm that lets the player know exactly how long they can stand in this one spot before they're going to die. It also allows them to pattern match against the countdown on their ability timer, so they know when those are going to be available and how long until another enemy is going to spawn in that location, and all sorts of various other components all fit into the same pattern. And by intentionally creating patterns like this, we help players understand the game better. And they they start to develop these ritual behaviors of their own. The constant feedback repeating loop of action and reward is visceral. It drives us players to think, "Oh, if I stay here, I'll get just one more point. Just one more try. Hey, just one more

match." So, it's become behavioral. Ritual behaviors become ingrained in our memory, so you tend about to think about them even when you're not playing the game. They're reflective. Each of these components drives you to play more, to play again, leading to better retention over time. Patterns also help us to establish muscle memory. Paragon uses a series of key presses for quick communication in game. And here, you hit left, right, right and it says, "Hey, good job." Now, I would really have to strain and think about it to tell you how to say in game, "I'm on my way to left lane" or "Hey, I need help in mid lane." But the point is that when I'm playing, I don't actually need to think about it anymore. My fingers know these patterns automatically after hours and hours of patterning. Now, of course, just going through the motions and doing a pattern doesn't mean that they're going to stick. That

pattern also has to be meaningful. And significantly, these messages are more than just a semantic pattern. They're used, again, as a reward mechanism. Good job pops up across the server when anyone on your team gets a kill or takes down a tower. I send these messages to other people that I'm playing with even if I'm in the same room with them and could have easily just said it out loud because it's become an automatic motion. And when complex actions become second nature, like these messages, it can reduce the overall cognitive load that helps with enjoyment and, of course, retention. Now, patterns don't have to be a collection of movements or a collection of moments, and they aren't limited just to player actions and behaviors. Uh here's an example from Unreal Tournament, which is a game that has tons of different weapons. Each weapon has two, sometimes three, sometimes even four different firing modes. It's a lot to learn. And in a game that's really

fast-paced, it's also really difficult to keep up with and understand what's happening. So, to get over this, we originally tried to give every firing mode for every weapon a unique rhythm. And then assign each of those to a specific time signature and musical key. So, the link gun you see is actually a 4/4 beat. And then the sniper rifle is a 6/8 beat, so it does a longer and then a reload. Uh flat cannon is 2/4, rocket launcher was 3/4. I mean, these these pattern through all of our weapons. This made them easier to use because the recognizable patterns allowed players to predict the firing state of an enemy, to understand the length of their reloads, and time the button presses on their shots. And best of all, this requires no effort on the part of the player because it's affecting them at an unconscious, emotional level. These patterns exist purely to help the

player, but from their perspective, they're not being helped, they're just playing the game better. And since they feel more successful, they experience growth and progression, which means they're more likely to come back and try for more. Social patterns are very important to retention. This could be an entire talk of its own. I don't have time to go into everything today, but just as a brief summary, I wanted to go over them because they can be so incredibly compelling. Whether people are in a guild or a clan or whatever your game's social group is, or even if they're just playing as part of a team, that social structure itself leads to increased engagement. Kongregate is a mobile publisher. They released this data a while back that shows that more than half of their revenue comes from 0.5% of the population that has played their game for more than 500 hours. Every single one of their top 10 games

has a guild, and guild players spend 10 times more than non-guild players. Among the people who do spend money, guild players spend three times more than people who aren't in a guild. Most relevant to today's talk, social ties led to an increase increase in engagement. Players in social groups spend more money, they play longer, and they play deeper. They retain much, much better. We saw the same effect when we introduced co-op to Gears of War and our Unreal series, uh especially when we hit uh horde mode. Led to huge gains in player retention for us because we created a social pattern. Co-op game types encourage repeat play because every time you go to play, you're with a different group, and you're going to have a different experience. And if you didn't do so well this time, try again with a different group. Cooperation is also more motivating and

rewarding than competition because the social connections create a sense of belonging and a sense of duty. They give you a reason to stick around and keep playing cuz you don't want to leave and hurt your team. Another really important social pattern in games is gifting. Gifts help us establish positive emotional responses from players, and thanks to our lovely human nature, feelings of reciprocity. A gift serves as a reminder of the person who gave it to you. It starts setting up a relationship with that other person, and it encourages you to respond. This reciprocal pattern of behavior leads to engagement. In our new mobile game, which was just announced this morning, Battle Breakers, go play it. Uh the more engaged you are, the more you participate in daily activities, the better your gifts are for your friends. And this fosters a sense of social engagement and rewards

players and their friends for the investment of their time, which just encourages them to come back and return that favor again for you. You can also create social patterns by adding friends list to your game. I mean, have you ever been compelled to accept a friend request from someone who you've never actually met? And when you send a friends request, isn't there an automatic expectation that whoever you send it to is going to reply? Social networking is all about that expectation that there is a pattern there and that it will continue. So, we carry on every day trying to build relationships and collect friends and likes and followers on social media, and we report our status with people who, in reality, we have no actual relationship with. But we tend to think that we do. And that ever-important emotional context motivates us to continue to come

back again and again and again. And the key takeaway I want people to get here is that when you're designing patterns, consider not only how they affect player behavior, but also how they could impact player relationships because relationships provide a much larger emotional context that can dramatically improve retention. One final uh one word of caution uh before moving on from patterns, once a pattern is established, you have to be very careful about breaking it or interrupting it. There were times in Gears where we wanted to create a very impactful moment like this scene here where Dom visits the grave of his wife who he had to euthanize after she had been tortured and lobotomized. Gears is a fun place. We really wanted this scene to have weight, to build up those feels in the player, and watching the sequence develop in our daily stand-ups, we were really happy with the way it was coming

together. We finally get it in game, and there's Dom kneeling next to the grave saying his final goodbye. And then he stands up right afterwards, runs into combat, and starts yelling, "This is GOING TO BE AWESOME! OH, YEAH!" NOW, these were his dynamically generated combat scenes, which were playing as designed by our combat system, but contextually were really, really wrong. And the inclusion of this emotional scene broke that established pattern of combat, and it was so obvious that it made this scene with Dom comical instead of impactful. It was really bad emotional design. So, of course, we ended up modifying the system before we shipped so that we had a series of switches that would filter out specific lines in areas where we wanted to maintain an emotional momentum and to do it in a way that was so

seamless to the player that it didn't interrupt that pattern of their emotional context. Patterns have a significant impact on the player's perception of your game. They determine that sense of pacing and flow. Repetition helps us build memories, and the consistency of action, feedback, and reward, and expectation help players set goals and find meaning in your game, which can dramatically improve retention. All right, signs and feedback. Good feedback is obviously essential uh to UX, especially in game design. This is one of the biggest do not miss opportunities for any designer. Feedback is vital for clarity, communication, and successful pattern recognition. The player has to know what state they're in if you expect them to

respond. Can they attack? Should they move? How should they react? Now, we experience emotion as a direct and immediate result of our interaction with the world. So, feedback needs to be just as immediate and crystal clear. We took this to the extreme in Gears. We focused on explosions with deep audio, visual effects that really kind of shook you to your core. We spent weeks working on just how much the screen would shake while you were running and the feedback that you would get when you were locked in a chainsaw duel. In Gears, you don't just get into cover, you slam into cover, and dust crumbles off the wall, and you hear the the character grimace and you can see the expression on their face. And that feedback helps the player feel the game experience in a very visceral way. In Paragon, we do very much the same thing. Whether it's launching a meteor or the impact against a shield, we do our best

to ensure that this conflict is felt, not just seen or heard. And a lot of our marketing, like this image here, Gideon pulls uh a meteor out of a dimensional rift and slams it down into the enemy. But in the actual game, he actually launches a series of small meteors consecutively instead of just the one big one. And we tried to rationalize that decision by telling our players that the small meteors actually did more damage over a larger area, so it was actually a more effective way to attack. But from that visceral, emotional point of view, our players were still very disappointed. So, we switched it to be a a meteor, one big meteor, just like in that image. And sure enough, right when we did that, the use of Gideon in Paragon spiked. He almost instantly became one of the most popular heroes in the game.

It's also important to note that at the same time we did this, we actually nerfed his abilities. So, if you look at his win rate over that same period, it actually stays pretty consistent. He didn't actually get any better by any metric, but we improved retention by changing the game's effects in a way that provided better emotional feedback for our players. Feedback doesn't always need to be physical or even really obvious in order to be effective. Paragon uses these absolutely insane shaders for skin and hair. We have eyes that can actually dilate in response to light. It's it's completely over the top, but it makes our characters feel that much more human. And as humans, we're hardwired to react to another human differently than we would to a game character.

And if the player's brain reacts to this character as it would to another human instead of reacting to it like it would a collection of painted polygons, there's a much higher chance that they'll form that emotional connection. And while your standard visceral reactions to game characters tend to be very transient, these types of influences tend to stick around because we're cementing them at that emotional level. Now, of course, you don't need over-the-top visual effects to achieve those same goals. Sometimes even just a simple color shift can affect someone's retention. As I demonstrated earlier, Gears had a very, very dark storyline and it got darker with every single iteration of the game. The first game was about the loss of home and family. The second game transitioned from small battles to these huge wars and instead of losing friends, we were losing entire

populations and cities. By the third game, we're dealing with the extinction of the human race and the potential destruction of the entire planet. Every game was more depressing than the last and we really feared that this oppressive tone might overwhelm or turn away a lot of our players. So, to counter that, we very intentionally tried to contrast the visual and emotional feedback channels. When the story became darker, the visuals became brighter, more colorful, and more saturated. This was less emotionally oppressive to people, so they stuck around longer. And while there are obviously a bajillion other factors involved, we actually saw a 10% increase in retention, in campaign completion rates, and a 14% increase in co-op completion rates across titles. And this type of tactic is actually a lot more common than you think. It's actually referred to as developing the

color script for your story. We were heavily influenced by the color script for Pixar's Up during the development of Gears, which follows a similar type of progression. It's completely desaturated in the intro, cooler in the middle, and then warm and hyper-saturated near the end. Schindler's List shows how color can serve as a way to really focus someone's attention. But more importantly, recall those memories at the height of emotional impact to ensure that they stick around. Our struggles with game tone in the Gears franchise heavily influenced the development of our later games. We really asked ourselves what changes we could make to the game without changing the actual gameplay that would hold on to people, more people, for longer periods of time. And when we were at that point in our discussions, this was Fortnite.

It was very gritty and very violent, just like Gears. This is Fortnite now. It's whimsical, it's light-hearted. They look completely different. But the gameplay and the core loop hasn't actually changed. So, if you were going to play Fortnite and retain for days, hours, weeks, months, even years, which of those two worlds would you want to spend 1,000 hours in? At the time, this barren landscape was Paragon. This lush environment is Paragon now. Again, where would you want to spend 1,000 hours? It's not to say that you can't make a compelling experience that's dark or moody. What I'm trying to say is that with a directed effort, you can use all of those various feedback channels to influence retention. And for us, the intent was to pair the tactile

feedback of the game with the more visceral feedback of tone and mood so that we could create an experience that you look forward to returning to day after day after day. So, good feedback that helps player understand their feeling or understand their surroundings at a physical, sensory, and emotional level can make them feel comfortable, and powerful, and successful, leading to the perception that your game is better, and therefore fostering repeat play. All right, last category for the day is mental models. Uh if you understand your player's frame of reference, you can use it to enhance your design rather than fight against it and risk player confusion or even abandonment. Uh in Fortnite, there are dozens of different objects that you can collect. Some are meant to be weapons uh used in combat. Others are meant as tools that

are used for scavenging. And on paper, this makes perfect sense. Someone with a broom is unlikely to use it to try and chop down a tree. Someone with a samurai sword is unlikely to use it to smash rocks. These are all very rational concepts, right? But our UX testing actually proved the unreliability of those rational design decisions. This is a player from one of our UX tests who tried for several minutes to smash rocks with a sledgehammer. But in the world of Fortnite, hammers are weapons, not tools, so they don't work for harvesting. In another test, players were totally confused when a stick was actually three times faster at chopping down a tree than an axe was. The axe is a dedicated weapon, not a tool. The design intent didn't match with the players' mental model of how these objects should function. This caused frustration. It made the game feel very

much like a long grind, which pushed players away. Eventually, we separated the weapon and harvesting systems so that each progresses and uh operates independently. And after that change, we saw significantly less confusion about harvesting in our UX tests. And while again, while there's a million other things that went into the two builds, we actually saw a 10% increase in repeat play amongst new players. Conventions like using an axe to chop wood can be important in other ways as well. Diablo has a red globe of liquid health on the left and a blue globe of liquid mana on the right. So, when you find a red potion in game, there's a pretty strong affordance for what it'll do. It matches player expectations, making it easy to learn. But that's not Diablo. That's actually Path of Exile. And here is Arbiters Judgment.

And here is Torchlight. I could go on. Similar games, similar genres, similar gameplay, similar UI. Why reinvent the wheel? People use known conventions as a means to learn new things. So, you can make their experience better by taking advantage of that. The less that players have to struggle to learn new things, the more likely they are to stick around. This means that affordances are actually pretty fantastic for retention. You see a pinata, you want to hit the pinata. How do you hit it? You swing back your weapon. And sometimes when you hit them, the pinata switches from paper to silver, or from silver to diamond-studded gold. And you get a pretty good sense that it's actually been upgraded and that the items inside will be more valuable. We never taught players this because what's intuitive doesn't need to be learned. It hits us at that visceral level and positively affects our

behavior. So, gameplay is more natural and with fewer hurdles, people are more likely to stick around, to think about the game more, to be reflective, and to come back and play more. And I know that sounds like a totally obvious, pointless statement to make. But when you're caught up in the complexities of development with interacting uh goals and personalities and deadlines, these sort of things really often tend to slip through. Take the tower design in Paragon. This giant thing right here is a tower. It's several stories tall, arguably the most important thing in the game, pretty hard to miss. It's right there in the middle of the screen. It's the focal point for all team play. You destroy towers to advance and win the game. So, we put tons of resources into their design. And yet, in our original implementation, we totally over-rationalized. When you're in a game, the actual physical tower is completely off screen and what you attack is the gem that

powers it. The gem doesn't look dangerous or protected. It's completely amorphous. And when it does attack, the tower hits you from somewhere completely disconnected from your view. So, it's difficult to make that association between cause and effect. And in our UX tests, we saw people tower diving over and over and dying over and over. Rather than attack the gem, some of them even turned to attack the actual tower that was shooting them, which makes perfect sense, but that has no gameplay effect whatsoever. All it does is get you killed. This poor participant from one of our UX tests played on easy mode versus AI, which is, in theory, nearly impossible to lose. And after a brutal hour and a half match, lost the game and came away with zero kills, 35 deaths, almost every single one of them due to repeatedly running under a tower and dying. And we saw this type of behavior very

regularly with new players. It was a critical issue in our UX reports because it was frustrating and it pushed players away from the game. So, we changed tower designs. Now, the gems actually look more like towers. They have shields over them to protect them and the shields fall when they're vulnerable. And when that tower attacks you, the beam comes straight from the thing that you're attacking. So, you can better associate cause and effect. And they now have rings around the outside that are red when it's dangerous to approach and green when it's safe. Sorry if you're colorblind, still a work in progress. But as of our most important our most recent UX test, tower deaths have actually significantly decreased and players seem to understand towers much better and are less frustrated overall. This goes to show that affordances run deeper than just looking at an object in the world. They apply to game systems as well. So, if you can design situations and

systems that so that it's easy and intuitive to do the right thing and hard to do the wrong thing, people will find success, feel empowered, and stay in your game. Now, of course, these are only a few of the factors that could enhance retention in games. And when I was putting together this talk, I couldn't decide whether I wanted to talk about the way that emotion affects design or the way that design affects emotion. And in the end, I really couldn't easily separate the two. And that's one of the reasons that I think emotional design is so important. It's universal. So, if you can address the visceral, behavioral, and reflective components all in one design, you're more likely to hit on something that is applicable to a broad audience of humans rather than satisfy one small group or niche audience. That's all I have for today. Thank you

very much for your time. Thank you. I think we still have about 5 minutes or so, yes, for questions. And please do fill out surveys. We'd love to hear what you thought. Tell me how I screwed it up or how I was wrong. Yeah, go ahead. Hi Jim. This is Ryan Miller. I'm from Deep Silver Volition. When you were discussing the social aspect, you were referring primarily to the multiplayer aspect. I do remember that Gears of War, you know, you had squads with you. And I know that players respond relatively well to like the mentor effect of seeing an AI ally do something. Did that have any effect you think on retention? Yeah, absolutely. As I mentioned, it's not just guilds or or clans or whatever, just being on a team has that same kind of impact. And in Gears, we actually specifically designed Dom to be more like your conscious. Okay. So, he would say the obvious

things that we wanted players to notice, but they didn't necessarily. And that helped players understand kind of puzzles and situations. And then each of the characters in your squad had kind of a specific personality that would offset different things that would really would basically existed to help you understand the game or get more into the story or more attached to them as people. And each of them had a very specific role that we were hoping they would fill out. Okay, thank you for that information, Jim. Hi, I'm Hannah Murphy. I'm a new user researcher. I do freelance game user research. Uh-huh. Um as you're talking about character design and becoming more realistic, um what are your thoughts on uncanny valley and do you see that being an issue? It's it is. It's a huge issue, right? And it's something you have to be very, very careful with. Uh generally speaking, it's a lot easier, cheaper, faster, all of those good things to embrace kind of the cartoony

side of things because if you don't make it all the way across the uncanny valley, you're going to fall into it. And so, you have to be very, very careful with that kind of stuff. Um as time has gone on, we have actually realized the high production cost of these kind of characters and just recently have kind of started to pull back from that a little bit more. Um so, you'll probably start seeing the effects of that later as we as we get into game, but I was just talking to one of our animators this week who is their entire design philosophy has shifted because of that, just knowing how much effort is involved and there's no guarantee that you're always going to make it across that uncanny valley, so it can be very dangerous to even try. Thanks. Mhm. Hi Jim. Gina Smith from 2K. So, I really appreciate that you used a lot of different games as examples throughout this from different platforms and different genres. Um and I guess my question is have you found that the platform on which a game stands or or even the genre that it

comes from affects these design principles at all? Uh it definitely can, but I would say it's more about how well you reach your target audience. You know, if doesn't matter if you're making a platformer versus a shooter. It what matters is if you're making a platformer for shooter people, right? That like that's going to affect it more than just I think the general platform. Absolutely. And access to whatever platform that is, you know, mobile maybe higher cuz there's more of them, sort of thing. Thank you. Hi Jim. Thanks for your talk. Nikki Crenshaw, I'm a PhD candidate at UC Irvine. Um so, you were talking about kind of like if it if it looks like Diablo, we expect it to play like Diablo. That kind of like establishing player patterns and stuff from the way a game looks. And this is something I've been thinking about for a while where like does this limit kind of what devs are able to do in terms of creativity? Um James Berg, I don't know if he's in here still, but he gave a talk a few years ago about they were doing some testing and it was a character that had a spear, but shot fireballs and all of these players were dying because they weren't expecting it to have magic abilities if it was holding a spear. Um so, is there a way

for us to get out of kind of like like breaking player expectations in a way that helps them learn or helps them break out of patterns that we've already like really deeply established? Right. I think so. It's not about I think limiting our creativity in terms of like, "Oh, if I'm making a Diablo-like game, it should look like Diablo." It's more about finding those particular aspects of your game that you don't need to reinvent the wheel on so that you can focus on those new elements that actually enhance learning and creativity and you don't spend your dev cycles working on the things that you don't need to worry about. So, you can make them easier by making them intuitive or by using known conventions, matching mental models, all those things so that the player gets those things intuitively and then your dev effort can go into the more creative and new elements that are unique to your game, whatever that is, and really you know, shining the spotlight on those because that's what's going to make your game stand out. Cool. Thank you.

Hey Jim. Mustafa Talji from Xbox. So, my question is what are the various ways and methods you use to effectively and accurately measure and gauge emotional involvements of user? So, one quick example you gave us in Paragon when the character became popular. So, you're looking at probably in-game in-game data how it rose of the that character's playability and the count. But for some trickier ones like the bacon out of the toilet, how do you kind of get that data and and get effectively and accurately measure like, "Oh, the user's engagement rose." Is it just looking at Twitter data, forum data, collating and correlating it? Are there any other effective ways you have? There's getting that data is incredibly difficult. It's one of those things that you have to Okay, so back in the days of Gears of War, we collected data on everything. And we actually ended up with so much

data coming back, we were doing a denial-of-service attacks on ourselves because we had more coming back than was going out. So, we had this huge amount of data that we had for no purpose and nothing to do with it. And so, we ended up throwing a lot of it away. So, now we try and be a lot more targeted in where you collect data on a specific thing that we are trying to, you know, research and find more information of. But if you don't think about that in advance, that means you won't have the data available. So, So, a lot of it has to do with collaboration between our teams and the dev teams where we say, "We're looking at this specific thing for X. Will you put the the hook the analytics hooks in the game so that we can collect that data?" And then over the course of several builds, we'll send out surveys or, you know, do UX tests and see how that information changes. Um that's not always easy to do because you need the dev team on to be able to do that for you, to hook put the hooks in. You need a data analyst to go through

the data and all, you know, all those various aspects. this is for predetermined uh emotional engagement like Right. I know that you want to measure this event because you know it may create that emotional engagement. There's other a lot of others where it's aha, like I just discovered this emotional. And these are probably the trickier ones of how do you actually in a lot of those cases, you won't have the data. You just kind of have to to look at it and you might be able to find some sort of data that supports that. That's dangerous cuz it may not always be accurate. So, I wouldn't I wouldn't recommend that, but it's possible. Um and then some of it is just as a designer, you kind of have to think about these things in advance so you don't get into that state where you're needing to make the change. And if you can think about it beforehand, chances are better that you at the end, you'll come out ahead. Thank you. Hi. Joyce Wang from WB Games. So, you're talking about patterns and creating rituals for players by making them repeat the things over and over again. So, how do you combat repetitiveness

that can come out of that? Uh so, that's where I was saying at that one point where just doing a pattern can start to feel like a grind. It doesn't it in and of itself doesn't become important or significant unless that pattern is meaningful. Um so, like those those reward messages that I the example I used where you go left right right, good job, those are actual rewards. Therefore, they become meaningful to the player. And so, they're more than just a semantic pattern of pressing buttons. And you have to you do have to be careful about things becoming a grind or things becoming repetitive. One way to address that I would say is start by establishing your pattern and then once it's established start building it and layering in more and more difficult things or new things to kind of keep that experience fresh and allow players to feel mastery of the thing that they're doing repeatedly while also being challenged by the new thing so that you kind of have help them experience growth in the long term. Cool. Thanks. Mhm.

So my question is kind of a follow-up in the other direction. What advice do you have in terms of making sure your patterns aren't so good that it leads to addictive behavior? That you know dark patterns and all that is a very it's a slippery slope. I specifically tried to avoid it today. I didn't want to talk about ways to get players to spend more money or ways to manipulate them. I think you have to go in with clear goals and a clear conscience to actually know that you're doing the right thing for the right reasons. You know you'll know if you're not, right? But you do have to be careful because there there still can be times where you want a pattern for the right reason but like you said people become addicted. You know people play hours and hours of MMOs. You get addicted to them and can't break the loop all that stuff. Um I don't know that that that's a very sticky ethical area that I'm not sure I

want to get into right now. Sure. Thank you. Uh-huh. Hi. Dana Ludwig from Xbox. You have mentioned a couple of times about UX testing. How is that different from user research testing? Can you just kind of describe what the goals are? I don't know that I could address that one. I would guess I have my two I have my my UX people down here. I'm going to guess they're just different terms for the same sort of thing really. Yeah. I mean I say our UX test but we have our researchers come in set up the test. You know we do the observation all that stuff. So it's pretty much the same thing. Just curious. Thanks. Yeah. Cool. Thank you everyone for your time. Appreciate it.