Greenlight Forever

An in-depth look at the Greenlight process, from the perspective of a project that has been there since day one, had a successful Kickstarter since then, currently has almost a 90% approval rating, but hasn't yet been Greenlit.

Since the beginning

Valve's Greenlight program has been a constant presence on the life of our project. Full-time development began two weeks before the release of Greenlight - I already had a working prototype of the game, and used those two weeks to make it look better and make it presentable and more easily understood to put it up on Greenlight.

According to statements from Valve before the launch of Greenlight, developers could post projects in early stages of development, "giving them a chance to raise awareness and generate excitement". It seemed like a good idea at the time - use our Greenlight page as a hub for all things related to the project - posting announcements as some sort of dev blog, using it as a forum, and getting important feedback to improve the project. And putting it up on the first day might mean that we got less competition for views, and/or that getting votes early on might mean even more views as it grew.

We got around 25% of 'yes' votes on that first month - I considered that to be an ok number for such an early build, since the question is 'would you buy this thing if it were for sale', and not a simple 'like' button. Most of the feedback was positive and there was a surprisingly complete lack of trolls. At the end, it still seemed like a good idea to have put it up early, even if it would take several months to get into the Top 100, we were racking up votes little by little. Boy would that prove to be completely wrong later, as we shall see.

Valve's initial queuing system

All credit to Valve, the system is a lot more fair when it comes to giving every project a fair share of views than I expected - and also that most developers and users give it credit for. It seems to me that the system tries to ensure X thousand views for every project will get inserted into users queues in the project's first weeks, which ensures all projects get their fair share of views regardless of how popular they are. Which is a pretty good thing indeed in my book. But I might be wrong on how that works.

After those guaranteed views, your project is mostly left to it's own devices, and views after the first months drop sharply. At the end of the first months, we were still getting about 100 views per day, but after a couple more months, it fell to below 10 per day. Which still didn't look that scary - after we got some proper cool trailer and pushed for Greenlight visibility from the outside, things should pickup again, right? Maybe so, were it not for The Dreaded Browser Barrier.

The Dreaded Browser Barrier

There are three ways a Greenlight entry can get views: through the 'Most Recent' button, through the 'Your Queue' button, or through a direct link.

The problem with getting views through direct links is that it has too much attrition, and there are way too many ways to lose that vote before it gets cast. First you have to get your project in front of people eyes somehow, be it email, social media, videos, or press. But these people aren't actively looking for stuff to vote like the ones that get in via the other methods, they are just reading - a huge portion of them won't bother to click the link. Of those that click, they will get directed to the project's page inside the browser, not inside Steam. Many people leave Steam running in the background all the time, but they might not be logged in their browsers, so even if they get interested, they might not go through the trouble of logging in to vote. Even when they do click to login, you better hope this is not their first access from this system/browser, or they'd have to go get the access code in their email, which is a sure fire way to risk losing yet another vote.

As proof that this is not all just hyperbole and a way to try to justify poor performance, here's the views and votes graph for the month we ran our kickstarter, where we had a Greenlight button right there at the very top of our Kickstarter description:

While we had a noticeable bump in votes during this period, it still paled in comparison to the early months. There was only one day, only our very very best day with tons of press mention, that brought in more views than our worst day on the first month. But the most astonishing find when I went in to analyze this data was this: WE GOT LESS VOTES ON GREENLIGHT THAN BACKERS ON OUR KICKSTARTER!! That needed to be in all caps and bold, because it simply boggles the mind - it can be almost as hard to get a vote on Greenlight as getting a backer on Kickstarter!

The takeaway message

All this means that getting in early on Greenlight is a huge, HUGE, mistake. The 'Concepts' area of Greenlight didn't exist when we got in. It sure was the place our project should be before we started our kickstarter. My advice to new projects is to stay there as long as possible, polish your trailer and your screenshots with very good stuff, get a good voice-actor for the trailer, make an awesome (and animated) logo/thumbnail, and keep licking that entry until you have a very nice ratio of likes-to-dislikes, as those views you get via the Greenlight system itself in the initial period will probably be the bulk of what you get, and it will only get a lot harder after that.

Here's what our votes graph looks like for our last month, with the trailers from our Kickstarter days, with very nice voice-overs, and updated and great-looking screenshots, a huge improvement of 'Yes' Votes % compared to the first month:

And we still need and updated trailer (in the works).

To drive that point home, if I could somehow reboot my entry and start from scratch with zero votes, dumping all the votes we got during our kickstarter, and during the last month when we switched into full Greenlight gears, but having the approval rate we have now instead of the one we had on our first month, here's how our 'Cumulative Yes Votes' graph would probably look like on the first month, compared to what it looks now:

Just looking at that drives me nuts. And I still have an uphill battle to get the next couple thousand votes I'll likely need. Writing this article is just one of the many things I plan on doing to help us get there. Hopefully it will help other devs avoid the pitfalls I fell into.

Do I regreat what I did? I do not, since I did what I thought was best given the incomplete information I had. If someone offers you 5-to-1 to draw to a flush on the river, you should take that bet every time as you are close to 4-to-1 to hit it, and not regret when it doesn't hit. Would I take it back if I could? Sure I would, just like I would take that flush bet back if I could when the river card doesn't hit me.

But in life, as in poker, we don't play with foresight, and we have to do the best we can with the hand we've got. I've got a shitty hand, but you can help me improve it by checking our Greenlight out:

http://steamcommunity.com/sharedfiles/filedetails/?id=92960968

And make sure you are logged in :)

Oh, and I almost forgot - here's your gift of entertainment for having read this far:

_________________________________________________________________________

Update: Commenter Joseph Mirabello pointed out a way to create links that open directly in the Steam client, which should make it easier for people to vote using your links. I thought it was too important not to add an update to the article itself

Original tip was posted in reddit by the developer of Black Annex.

In order to create a link that opens the Steam client in a Greenlight page, there's an undocumented CommunityFilePage url feature in the steam:// protocol. Wrap a php file around that for external stuff like twitter and facebook to properly link to it, and voila! This is the end php file contents:

<?php header('Location: steam://url/CommunityFilePage/92960968'); ?>

Add a redirect to make a nice-looking and clean link to this php file, and there you go:

greenlight.adventurezator.com

Ahhh! Much better! Thanks Joseph!