[This piece was originally posted at weatherfactory.biz]
Lottie and I are both open production enthusiasts - I did a talk on it at GDC a couple of years ago, and as a reminder, here's our roadmap.
So this is a post about the strategy we've taken at Weather Factory with making Cultist Simulator available for full release, and how it's worked out. It does include sales figures! We talk to a lot of other indies and I know people are always hungry for data. SteamSpy has gone a long way to alleviating this, but it's never the whole story - especially when a game's been Kickstarted and there are an uncertain number of keys in the wild.
Cultist Simulator was Kickstarted, and we have also done a couple of slightly unusual things.
The first unusual thing is this. From a post I made back in July 2017:
So when I run the Kickstarter for Cultist Simulator, I’ll offer just one tier: Perpetual Edition.
Perpetual Edition means: you buy the game before full launch, you get access to early builds and any and all DLC and expansions free forever. If I release the game on Steam Early Access or itch.io’s Refinery or GOG’s Games in Development or anywhere else, that’ll also be Perpetual Edition.
As it happened, we're not releasing CS as Early Access or Games in Development. We'll go straight to full release on May 31st.
Here are the main reasons devs consider Early Access - at least, the ones I'm aware of.
- to raise cash
- for feedback and beta testing
- for buzz and community building
(3a) as a service to backers
[(3a) is a special case of (3). You tend to get people mailing you after a Kickstarter to say 'shit I missed out! can I back?' You can reply to them all and say, sorry, no, wait for full release, which may upset your nascent community, or you can reply to them and say, yes, PayPal me ten quid, which is insanely manual; or you can just provide some sort of slacker backer option. It just removes some pain to add the slacker backer option.]
The thing is these objectives can clash. If you need cash, you'll want to be sure the game is as polished as possible. If you want feedback, you'll want to get it out as early as possible.
If you're running a business, you don't turn down money. But we were okay for cash - we'd just had a well-funded Kickstarter. And I really wanted feedback. Cultist Simulator is a deliberately experimental game, and I wanted player responses ASAP. Plus, we have limited QA resources, and allowing backers to help us squash bugs earlier was a big win.
So I wanted to get the development builds in front of people quickly... and I didn't need that many participants. We already had nearly 5,000 Kickstarter backers. I knew most backers would be happy to leave the game in the oven for the next eight months, but at a conservative estimate we'd probably get 400-500 beta testers. That would be plenty, right? For a single-player game, you don't get 10x better feedback at thousands rather than hundreds of players. You do get a lot more comments and bug reports to process.
But people burn out, and also people become veterans. If 500 people beta-test your game when the first dev build hits, a good many of them will stop playing after the second or third (especially if early builds are buggy - which they will be if you're going in early). And a lot of others will be very familiar with the game, and will stop seeing it as new people will. You'll still get useful feedback, but you can't use them to tune your early game experience, which is the most important five minutes of the game. (This is one of the things I learnt the hard way on Sunless Sea.)
So we wanted a steady trickle of new players. But we only wanted a trickle, not a gush. We wouldn't get much more useful feedback from Early Access on Steam, and we might screw up our chance of a good launch later.
And of the digital store-fronts, Itch is unmistakably the one where people expect to find experiments. I'd already put the pre-Kickstarter alpha on it. Itch has, anecdotally, something like 2% of the market share of Steam (I've heard numbers as high as 5% and low as 1%). So we could soft-launch over in Itch and not many people would notice, but we'd get our trickle of new players.
(Also, did I mention I really like the Itch upload tool? I really like the Itch upload tool.)
So we settled on early purchase / early builds on itch.io.
I had originally planned to leave Perpetual Edition on sale until the day before we hit full launch, but Lottie, when she joined me in December, talked me out of it. A month before launch, we wouldn't be able to do much with the feedback - it would be hard to change direction - and we'd be busy tidying everything away and strapping in. Sure, we'd lose a month's revenue, but we'd already agreed the point of the Itch Early Access was not primarily cash. I tentatively agreed.
After that, we got advice from another source [*this is foreshadowing for a followup post*] that if we encouraged people to wishlist the game on Steam rather than buying it now, our chances of charting on Steam in launch week would be that much higher. This made sense. So we agreed we'd make Perpetual Edition unavailable in February, but make it available again in launch week, as an early incentive. In related news, here is the wishlist link for the game. It would be charming for you to use it.
A few other things that have happened that might be useful info if you're also a dev.
First of all, we've had some lovely streams of the game: SystemChalk, FuzzyFreaks, Mike Laidlaw (stream deleted; cryface), Jessy Quil, to name but a few. This is nice for buzz and great for feedback. Someone playing your beta product through, talking about their experiences, trying out loud to understand it, while the audience discusses it in text chat? Holy shit, I wish I'd had that when I was building Fallen London. (And to my surprise, it streams pretty well: it's slow-paced and abstract, but there's always something just about to happen.)
Secondly, lots of people wanted to be able to play the game on Steam ahead of full release. This surprised us a little - we hadn't realised how much of a lock Valve has on people's playing habits, even though nearly all my gaming is in Steam - but it encouraged us to sort the kinks out of Steam deployment ahead of time, so I'm glad we did it. But!
(i) it makes messaging surprisingly tricky. We've gone from
'you can buy Perpetual Edition on itch now, or wishlist it on Steam and get the full game at launch'
'you can buy Perpetual Edition on Itch now for free lifetime DLC, or wishlist it on Steam and get the full game at launch, but actually if you buy Perpetual Edition now you can also play it on Steam, and yes if you buy it now you will still get free lifetime DLC, but also you can get free lifetime DLC in launch week. On Steam.'
...I mean we generally just say 'buy on Itch if you want preview builds, wait until May if you want the full game'. But it's led to a few confused conversations. This is what happens when you adapt your strategy on the fly.
(ii) fun fact! You can upload Steam keys to Itch, but by default they're only available to people who've purchased the game for money on Itch. If someone redeems an Itch key you give them because they Kickstarted the game, you'll need to send them the Steam key manually. We had to do some extra work to sort this out. Be warned.
Data is valuable. Honing your ability to use that data is super valuable. Always do projections, now matter how ugly or informal. Comparing what you guessed against what you achieved will make you that much better at guesstimating the future next time.
I did top-down and bottom-up, aka 'hand-wavy' and 'based on wrong numbers', informal estimates for itch sales:
- Top-down was: From talking to similar-sized devs and remembering the pre-order numbers on Humble for Sunless Sea, reckon I can manage a half dozen purchases a day. Call it five. 5x30 is 150 a month.
- Bottom-up was: I got 4,788 backers in 30 days for the KS. I was working hard to get the word out, and I'll be working much less hard now. Plus, where Steam is a metropolis, and Kickstarter is a district capital, itch is a pleasant country town. So... say 5% per month of the original total for not banging the the drum and getting press, and for tapping out my core fans (who I bloody love. Thank you.) That takes us down to 240 a month. Cut that by 3/4 for itch being out of the way. That takes us to 60 a month.
We have played fast and loose with maths this far, so we now casually average the two estimates together to get 105 a month.
October, November, December, January, February, 5 months, maybe 300 sales lowball, 750 highball, 525 most likely.
How close is that?
621 purchases. (And 3,367 downloads? Those extra downloads are Kickstarter backers. More about that in a moment.)
I'm posting this on 20th Feb, 8 days before the end of that 5-month period. So this is noticeably better than I expected, but in the ballpark. It is a running joke round these parts that my estimates tend to be remarkably close to 80% of the final total (this goes right back to the Sunless Sea kickstarter I blogged about when I was still at Failbetter), and the longer that keeps up, the better. I would much rather always err on the side of pessimism.
Why was it higher than I expected? I think some of my initial numbers were a little optimistic, but then we had a bunch of good things the last couple of months. We got interest from streamers; we got a feature in PC Gamer and an award nomination; we got a video feature in Eurogamer. This is better than we thought we'd do in terms of publicity at this point. We also saw some spikes every time we reminded people that Perpetual Edition was not being sold after February - time limited offers do that. Nothing dramatic, but enough to show up in the figures.
One thing that I cheerfully averaged out in my original estimates: how much difference would it make when people could actually play builds? It did make a difference, but less than you might think. The game was on sale for pre-order from October, and we put the first dev build out in December. We went from three-four downloads a day to six-eight downloads a day, with an exciting sales spike on build release day of 11 whole purchases!!! The numbers started creeping up as people actually started playing the game.
Anyway, I'm very happy. 621 sales isn't going to break any records, and 9000 USD doesn't come close to paying our costs for five months, but the Kickstarter has us covered for that. The feedback and the buzz we've gained have been invaluable, and we're in a much better place for actual launch.
These are the lessons I think we can draw from this.
(1) I've seen people worry about Kickstarted games doing worse because their backers have already bought the game. I am strongly of the opinion that the reverse is true. If you are an indie, then unless you are incredibly successful, there are always far, far more people who have never heard of your game than there are people who have heard of it. Our problem, as indies, is to get noticed in the first place. Hundreds or thousands of backers who will help get the word out, who will talk about your game on social media, who will show up as 'now playing' on Steam - this is a big deal. If people are playing your game, other people will buy it. I'll say that again: at our level, there are always far, far more people who have never heard of your game than there are people who have heard of it. It's also another good reason, if you needed one, to keep talking to your backers.
(2) Always estimate sales, however vague the numbers and however uncomfortable it makes you feel. We do mentoring for other indies and we often have to push people into actually guesstimating their figures. But the more you do, the better you get at it, and really blunt-edged estimates can be surprisingly accurate because so many of the unknowns cancel out.
(3) We've usually been on the itch top sellers page, popping in and out of the top 10. We went as high as #7 at one point, we're generally sort of #14ish. So if you are planning to put your game on itch, this is the sort of business you might see as a moderate top-seller. (Don't draw too many conclusions about the top 5. The top of a chart tends to do dramatically better.)
[EDIT: the day after I posted this, we jumped up to #4. We sold 24 copies that day, though about 50 on each of the two previous days. Er, okay? Maybe itch works on a 3-day moving average. I'll tweet Leaf Corcoran and ask.]
(4) This is pretty encouraging for full launch. My casual top-down prediction would be that, if itch is 3% of Steam, this will amount to about 20K sales, which would put us firmly in the black for the year and make me very happy. There is more competition on Steam, but we really haven't started our marketing push yet; itch is more forgiving of weird games, but the game isn't actually finished yet. The big unknown variable is that we have a publisher now (yes, that foreshadowing I mentioned above - more soon) and I don't yet know how much their assistance will help.
(5) Make sure you know why you're going into alpha, or Early Access, or paid beta, or any other pre-launch sales-y activity. Most importantly, pick whether you're looking for cash, or feedback: choose a star to steer by, and stick with it.