Case Study: One Game's Amazing Click-Rate on Steam

(Vagueness: The take-away points of this will be intact, but if I act vague about exact numbers or certain practices, it's because I don't want to violate the non-disclosure agreement we (and other developers) have with Valve regarding their store.)

TL:DR

That little, tiny strip of your game logo on Steam is really, really useful. Take a second look at it, and try to improve it to grab more eyes. Also, consider tweaking your Tags.
Unless you've released on Steam before, you probably don't know how extremely powerful that slot is.

We had 4 games released in one day, and we compare and contrast why one game got a lot more clicks than others.

The Steam Front Page

This article is about examining a tiny sliver of art which very much influences how much traffic ends up on your game's Steam page. As we can see by visiting the main page on Steam, there are several categories where games are shown.

Front page on Steam, September 29th, 2015

We can see the large, top part--currently it shows 80 Days--which cycles through many games and the lower, smaller section that reads "New On Steam" just above.

The valuable, New On Steam section.

This "New On Steam" part is where we'll focus on our time and thoughts, but let's first look at a few points.

Front Page Flow

Now, without running afoul of any agreements we have with Steam, we can still determine and talk about some interesting points. Anyone that spends some time analyzing the front page for a few days can see a pattern. If you keep track of games that show up on the Upcoming tab near the bottom, and then compare them to games in the New On Steam section above, you can see that it seems that games that went through the Upcoming section get some time in the New On Steam section.

The Upcoming Games reasonably seem to show up in the New On Steam section.

So we have the idea that, 'Upcoming Games' move to 'New On Steam'.

Lots of Eyes

Without saying how many potential times gamers could be shown your game banner, let me just assure you it's a substantial amount, that any developer would be happy to have.
It's a very nice setup, since independent of your marketing budget the New On Steam section affords (apparently?) every game an opportunity for a lot of people to see it.
So, let's accept the assumption that the New On Steam section is very valuable and powerful.

High Traffic Analytics

Once you have your game on Steam, you get access to some nice traffic analytics, breaking down how many people saw your game banner in the New On Steam section but also how many clicked on it. You get to see total views and total clicks.

I'm personally in an uncommon situation, in that we recently ported 4 of our popular XBLIG games to Steam, and released them all on the same day. From that, I now have 4 data points from the same time period. While anyone could simply watch the Steam front page for a while and track how long games show up in the New On Steam section, I'll tell you, for us all 4 games stayed on there about 10 hours each.

Mommy's Best Action Pack promo

Steam also shares with you the *average click-through rate of other games*.
One of our games matched this average.
Two of our games were nearly twice the average.
And our final game had over *three and half* times the average number of clicks.

The Click-Through Puzzle

All of our games in the "Mommy's Best Action Pack" were critically praised over the years. The popularity of the games goes in this order: Shoot 1UP, Weapon of Choice, Explosionade, and then Game Type. Game Type is a fun but very strange game, plus it's an extremely niche 'caravan' style shoot 'em up. It never sold well, but it was respected.

Of course one has to factor in 'brand awareness' when considering why someone would click on your game banner, after they see it. Maybe they've heard of it before and are interested! (That's of course exactly how brand awareness is supposed to work!)
Instead, in this case, the click-through breakdown went like this: Game Type had 3.5x the average number of click-throughs. Shoot 1UP and Weapon of Choice both were over the average at about 1.8x times. And Explosionade simply had an average number of click-throughs.

What does this mean? For me it shows what I'd expect, that Shoot 1UP and Weapon of Choice had some brand recognition, and brought some people through to their game pages. This is good.

But the puzzle for you to consider is: What happened to Explosionade (average click-through) and Game Type (amazing click through)?

Compare and Contrast Clicks

Obviously we all work hard to create the game logo and the title treatment. But the wide, small game banner used in the New On Steam section is tough. It's horizontal and it's small. You don't have a lot of space to show off your game... or do you?
I took this as an opportunity to consider the art used for each of my 4 games and you should look at yours as well.

Here's what the banners for each game looked like at launch, during the traffic period in question.

Original versions

Here are my notes from looking at the art.
The Explosionade title takes up more room than any other game art here. Game Type text takes up the least space. Weapon of Choice and Shoot 1UP are very colorful by comparison, and bright. Game Type has an identifiable human girl, kicking, which is strange. You can kind of make out a cat in Game Type. For Explosionade, I can make out two bullets sort of, but I can't tell what's happening on the lower left, it's mostly a wasted opportunity. The monster on the bottom right of Explosionade shows his eye, but loses a lot of context, which could mean for some people that the coolness of seeing the glowering monster is lost. The Weapon of Choice 'font' is weird and non-standard and has strange organic shapes all over it. Shoot 1UP has a girl who is peaking over the edge, this is titillating.

Here is the new version of the game art.

New versions (only Game Type and Explosionade were changed)

The main changes to Explosionade are to give things more context, and reduce the text size. Keep in mind players will see your game name, just to the right, in plain text. They will know the game's name, so can reduce your title art somewhat if you have more interesting things to show. I've increased the space the mech on the lower left takes up, and increased the space the monster on the right takes up to indicate the cool fighting in the game. Game Type was made brighter and the other two games were not changed.

Let Tags Do Their Thing

The other thing to consider is that I labelled all my games "Action", and "Indie". There is a *broad* selection of  games on Steam that fall into those categories. I wasted my chance by not labeling Shoot 1UP and Game Type as "shoot 'em up" (much fewer games) and I wasted my chance to label Explosionade as "Mech" (very few games).
Players could have quickly seen those words just to the right and had another reason to click!

What's the Point Now? The Moment Is Not Gone!

Why improve the art, after the games are out of the New On Steam part? Because that same art is used all over Steam, in the 'More Like This' section and the 'Recently Updated'. You get more chances to catch player's eyes!
Redemption! If your game had a poor performing click-through rate on launch, you can do better it!

If you have not released your game consider:
As I said, if I could do it over, I should have added more useful tags to my games (the "User-Defined tags", not the search words, which are also important). Adding custom tags that make sense to your game, but some that are also specific.

In any case, take another look at your game banner, especially the small one, and make sure you're making most of it. There are a lot of eyes on that tiny piece of art and you can get more clicks with it!

----

(And since we're sort of talking about designing box art (but smaller, and wider) I found this to be an inspiring article on how others made some very nice art: http://www.polygon.com/features/2015/8/27/9207581/video-game-box-art)

(And finally, please chime in as to why you think Game Type scored so well and Explosionade only scored 'average'. I'd love to hear different points of view, thank you!)