Gamasutra: The Art & Business of Making Gamesspacer
arrowPress Releases
August 29, 2014
PR Newswire
View All
View All     Submit Event





If you enjoy reading this site, you might also want to check out these UBM Tech sites:


 
Picking the Perfect App Icon
by Greg Pollock on 03/21/13 04:30:00 pm   Expert Blogs   Featured Blogs

The following blog post, unless otherwise noted, was written by a member of Gamasutra’s community.
The thoughts and opinions expressed are those of the writer and not Gamasutra or its parent company.

 

Having a compelling app icon is important to an app’s success, first when potential players are making the decision to download and later when they’re opening or re-opening the app. Creating a good icon is, in my opinion, first of all about expressing what the game is. Presumably you are making a quality game and honestly presenting its fun is your best bet for getting players. But even a great icon for a great game can get lost in the noise if many superficially similar apps are using the same representational strategies. Knowing what icons your app will be situated near can help  you choose one that will stand out from the pack.

I made a spreadsheet where I took a snapshot of overall top grossing overall, top grossing RPG, and top grossing Adventure on 1/25/13, categorizing each icon as one of the following.

  • Cartoon person
  • Realistic person
  • Cartoon animal
  • Realistic animal
  • Relatively realistic inanimate object
  • Iconic/abstract

I am also interested in the use of gender for both identification and sexual attraction (we've all seen those Evony ads, right), and made columns to track when a strongly gendered male or female figured showed. Because of the question I wanted to answer--how does gender identification work in app icons--not all figures who appear to have a gender are tagged in those columns. For example, I omitted figures that, judging by shoulder and hip width, are probably male, because I do not think this kind of representation is using gender identification in the same way that something like Modern War (which has a close up of a man's face) does. This could well be a flaw in my methodology and I would be happy to see a better study focused on gender in app icons.

Part of the reason for doing this study--as well as the reason for the softness of my methodology--is that creative judgments require a human eye. When I have a question that numerical analysis can answer, I trust our marketing and analytics team to give as precise a response as possible. But there’s no command that can read the art style of an icon, and so to answer this question I had to accept a small data set and an imperfect, impressionistic method of categorization.

Part of what I discovered were the limits of my rubric. The distinction between cartoonish and realistic as I use is very imprecise and a little misleading. Here, realistic tends to mean detailed while cartoonish means broad. A lot of the icons that I class as “realistic” are done in an anime style and are not realistic per se, but realistic in that they use techniques that would capture the detail of real things. This was the best terminology I could come up with.

  • I used this distinction because it works pretty well for describing the two major styles of human representation. It also maps pretty well to whether the target audience for a game is male (realistic) or female (cartoonish). This is especially true when women are represented. Very “realistic” women front grindy PvP games with high fantasy themes, which, based on historical app data from Facebook web games, I believe do best with boys. Cartoonish women front games with realistic themes (paradoxically enough) like Diner Dash or Campus Life. Again, this repeats a pattern seen in Facebook web games.
  • The realistic/cartoonish distinction works worst for describing animals, where many icons (including Gaia’s Monster Galaxy franchise) sit in the middle. I think this is because monster training games are less discretely gendered and are more likely to be popular with kids.
  • Cartoonish icons also appear more common in games with the largest audiences (Clash of Clans, Jetpack Joyride, Temple Run). This makes sense: an accessible icon that doesn’t divide by gender has a larger addressable audience (than, say, an anime babe in a metal bikini). The key, of course, is to put such an icon on a game that is also addressed to a similarly large audience.
  • An aside: there are a lot of Minecraft-related apps in the top grossing charts. Minecraft isn’t a game, it’s a platform.
  • The iconic/abstract category was originally two categories--iconic and abstract--attempting to parse the difference between representational (iconic) and non-representational (abstract). I broke down under the weight of slots apps, which are both iconic and abstract. (I would say that as icons and games they represent their own abstraction: they are just random number generators with pictures). I cannot understand how there are still so many slots apps, but there are. Slots aside, I was surprised how many games use iconic or abstract strategies when it seems to me like a bad way to represent a game. This is something I learned.
  • In top grossing, strongly gendered characters split evenly. I take this to mean that if you are making a gender play, either a male or a female character has equal chances of standing out to your audience.
  • If you look at subcategories, there are striking differences from overall top grossing. RPG is a good example of the extreme deviation that happens in a subcategory, where app icons use human much more frequently than in overall top grossing. There is also a marked change in gendering, with women (and most often women represented through a male gaze) more than doubling the number seen in overall. This could either mean that using a babe is a strategy with consistent results or that it is oversaturated.
  • Adventure seems to use realistic inanimate objects more than the other two charts. I don’t take this to mean that a realistic icon is s better or worse, but that you might change your icon to match the goals of a marketing campaign targeting a particular chart.
  • Finally, I looked at the very top of the overall charts (screenshot below), since the top nine or ten apps get the lion’s share of exposure and revenue.
  • Top of the App Store charts
  • The sample size is by definition very low, which is why I wanted to look at the top 100 of three charts in the majority of my study. The rates of representation are about what we see in the top 100. Kingdoms of Camelot and Marvel War of Heroes, for example, are apps I would class as “realistic,” and they have about as much representation in the top 10 as in the top 100 grossing. One thing of note: in the top 10 of grossing, paid, and free, there are no women except for She-Hulk tucked in the corner of the Marvel icon. The same is not true of the subcategories, where two apps in the top 10 of RPG have female characters.

After looking at a few hundred app icons I've been able to refine my initial idea of matching the image to the game. Certain strategies are more common for certain game types, and those tendencies become more exaggerated when you look at the subcategory charts where they are selected for. Knowing where you'll fit can help you play to your audience's expectations, if you're going for a modest improvement on an existing genre, or give them something that will stand out if your game is likewise a radical innovation. 


Related Jobs

Mixamo
Mixamo — San Francisco, California, United States
[08.28.14]

Animation Outsource Manager
Cloud Imperium Games
Cloud Imperium Games — SANTA MONICA, California, United States
[08.27.14]

Senior Producer
Yoh
Yoh — Vancouver, British Columbia, Canada
[08.27.14]

Rendering Engineer Job
Yoh
Yoh — Vancouver, British Columbia, Canada
[08.27.14]

Multiplayer Designer Job






Comments


Zack Wood
profile image
I'd love to see the results for "Top Grossing Action", too. Thanks for doing this research.

Seems like there are lots of other interesting ways to gauge the icons, like by color, or by commonness of the "Something in the center with lines exploding out" design, which seems extremely common.

I'm not sure if it'd be better to match top-grossing icons' design in your category, or do something different to stand out...

Greg Pollock
profile image
Yeah, there are a million ways to categorize icons. Optimization is a real rabbit hole but it's fun to go down now and then. I'd love to see more analyses of art styles in the app store given the large sample size (and comparative grossing rankings) it makes so easy to access.

Our company has done a lot with the sunburst background you mention. I have no idea if there is a rationale for why it works (I guess it is a nice technique for focusing the eye) but I find it pleasing and give it a thumbs up.

Lex Allen
profile image
Our eyes are naturally drawn to human forms, so if you're looking at a painting, your eye will often go straight to anything that looks like a figure.

I think color plays a major role. Most of the apps are blue, and none of them are white (if you look at the top ones). Black and red aren't really as popular either.

However, this could be because of the designer's choices or because the majority of the icons are blue already. It's not necessarily related to user choice, but blue consistently rates best as a marketing color.

Greg Pollock
profile image
Actually I think red is a very good color for app icons. Angry Birds uses a red main character, as does Dragonvale. Candy Crush's foremost piece of candy is red. Clash of Clans' background is red. Slotomania is primarily red, as is Zynga Poker's current icon. Legend of the Cryptids AND Rage of Bahamut are red. CSR Racing uses a red car in its icon. And in our internal tests, red characters have performed best.

You're right that blue is a good color, and in some of those icons (and the Monster Galaxy icons I've run A/B tests on) blue and red is a winning combination (Dragonvale, for example, is red character on blue background). Because you are trying to get potential players to make the quick decision to download, and there is no downside to downloading a free game, I think the urgency of red helps tilt that decision in your favor.

Oh, and I totally agree about the power of the human face to draw our eyes. That was my initial hunch that I wanted to verify with a little quantitative analysis.

Steve Cawood
profile image
I think the biggest selling point is the screen shots not the icon. The icon is like a dvd/bluray cover where the art can look awesome but isn't reflective of the final product. Lots of low budget horror movies do this.

The screen shots are what makes a title sell, and of course the ratings along with the title. I think a really bad icon would make me curious to check out the screenshots more than a really boring icon.

Steven Christian
profile image
The icon is the hook and the screenshots are the sinker.
You need both, but your audience will never see the sinker if they are not hooked first.

Matthew Burns
profile image
Nicely done Gregg.

For us app designers/developers, coming up with the right icon is an essential but difficult undertaking. It seems you only have a second to catch a gamer's eye as he/she peruses over the countless games available.

Matthew Burns
profile image
I hope you will share anymore research you do on this subject.


none
 
Comment: