XING Postmortem (2019)

NOTE: I wrote this post all the way back in 2019 but never posted it. I’ll append with some “6 years later” comments.

The Good, the Bad, and the Ugly

Well, it’s been almost a year since we put XING: The Land Beyond on PlayStation 4 and PlayStation VR, and over 2 years since we initially launched on PC and PCVR. I think it is finally time to take a step back and look over what we have accomplished, where we are now, and where we are headed! That’s right, it’s the XING postmortem!

I’d like this piece to be interesting and accessible to anyone interested in our story as well as information for future developers. I know that I have personally read many postmortems looking for advice and guidance, especially when we were starting out, so if possible I hope this will be useful for future devs. I’m going to go ahead and speak personally about my experiences and takeaways from the project.

We set out to make XING seven years ago, all for various reasons. For me, it started as simply something to work on as a capstone for my senior year of university. We’ve already written a blog about the early years of XING (or of course you could just simply look back at our earliest posts) so if you are interested, check it out.)

Let’s go ahead and get some of the juicy stuff out right away:

HOW WELL DID XING SELL?

Pretty well, at least so far! Actually, as far as I can tell, better than ~90% of indie games out there, which is crazy. It’s still a little unclear how transparent we are legally allowed to be with our sales data so on the safe side I’ll euphemize it a little.

XING has earned around 200k net since launch. This estimate basically represents money we have actually received from Valve, Sony, and Oculus, and doesn’t include any expenses or other income, like Kickstarter and grants.

That translates to around 20k units spread across our three major storefronts, which is incredible.

Some interesting immediate takeaways:

• Sales from PS4 / PS VR have already eclipsed Steam and Oculus Home individually, and are on track to beat them combined within the next year. Based on what we have seen from other developers, this really isn’t that surprising, but I think there are a couple of factors at play here other than the simple “less competition” argument that gets thrown around.
  • Since the PS4 version came out over a year after the initial PC version, it launched with a dozen or so patches already built in, avoiding some (but not all!) post launch issues.
  • We actually partnered with Sony officially - this helped fund our development, and perhaps it gave us a few extra opportunities we would not have had otherwise.
  • For whatever reason, XING resonated with the PS VR community far better than it ever did with PCVR. Honestly I’m not entirely sure why that is - maybe the game is just a better fit on console…? More on that later.
• We’ve sold more copies on Steam than Oculus Home, but have earned more on Oculus. Likely this is simply a result of Oculus customers being less numerous but more willing to pay full price / less dependent on sales.
• If we included pre-sale / kickstarter sales, this number would of course be more skewed. Take a look at this graph I quickly made with our sales data. Again I’ve omitted the exact amounts, but you can probably figure out the magnitude here:

INSERT GRAPH HERE

I’ve crudely labeled some notable events: A: Our initial launch in September of 2017 on PC and PCVR. Notice that this is not our best month! Pretty much every chart we’ve seen online follows the “Stegosaurus” pattern - one huge spike at launch, then smaller spikes over the next few years. While we certainly see spikes here, our first month was not spectacular. In fact, I recall being quite concerned with even being able to fund PS4 development based off of our first 3 months. B: If nothing else, this singular data point shows just how at the mercy indie devs are to our storefronts. In March of 2018, out of the blue Oculus reached out to us asking if we would like to participate in a bundle with an upcoming title: Apex Construct. For a week or so we were plastered on the front page of Oculus Home along with Apex, netting us tons of sales. C: Likewise, in October 2018, Steam asked us if we would like to “participate” in their startup popup ads. Yes. Yes is the answer. That single day the popup ran we made something like 20x the sales of the entire month prior! D: Finally, we launched on PS4, and likely due to some lessons learned about building up hype, and with likely a little help from Sony, we were able to get that launch day spike we predicted. Keep in mind “combined units” here means sales from both NA and EU - they are technically two different products according to sales. E: While we have had a few sales in PS4 NA, it is a little harder to get approved for EU sales. August of 2019 marked our first EU sale, to great success.

Ultimately I’ve been fascinated with all the data we’ve collected from releasing our first title, and have learned a ton along the way. While I’ve personally think XING has surpassed my expectations critically and commercially, there are some other things to consider:

• We had three people, plus several hired contractors, working for roughly 5 years on this title.
  • Needless to say, we barely paid ourselves over those 5 years. We each had various draws over the years but for the most part the money we had earned from presales went towards marketing, contractors, and equipment. Essentially we were only able to do this due to us being young and living with our folks - a common trend I’ve seen with other young devs.
  • Admittedly, a few of those years we were still students, so one could consider that time as more of an extension on our education (honestly, I’d consider the entire production to be!)
  • Even if we were able to take all of our profits and push them directly into our salaries, we would have been making ~10k a year each. Yikes.
• We are our own publisher
  • While we still owe the standard 30% to the distributors, we don’t have a publisher to pay. Of course, this means we have to take care of everything a publisher does ourselves, which turns out to be a lot!
  • Then again, we could have sold quite a few more copies if we had publisher support, as well as potentially have released earlier (more time dedicated to development)

ALRIGHT, WHAT WENT RIGHT AND WHAT WENT WRONG

For kicks, I’ll use the classic “The Good, the Bad, and the Ugly” format to outline my thoughts from development. I’ll list out some development categories and give some thoughts about what went right and wrong. ART: The Good:

The Bad: Our artstyle attempted to blur the line between realistic and impressionistic. Unfortunately, this style demands rendering power, something at a premium especially for VR. Adding on this our game design demanded dynamic lighting, which was excellent for iteration time but terrible for performance. Finish it up with the promise of PS VR support on a base PS4, and what do you get? A hell of a time getting things to look passable (or run at all!) on low end hardware, and of course a hard limit to which platforms we could conceivably support. Ultimately I’m proud of what I was able to achieve on PS4, and the game does run on modern integrated GPUs, but a modified style may be in the cards for the future. The Ugly: We started on level art REALLY early - in fact, we had entire levels more or less art designed in 2013! Of course, there were many, many iterations between those version and what ended up in the final game, due to engine update, game design changes, new hardware requirements, and new expectations. We ended up redoing level art far too many times. Similarly, we all effectively learned how to content create with this game, meaning that our skill with making assets improved vastly over development. Since we did not have an established target, a general lack of QC meant lots of redone assets. This also cascaded into a poorly organized workspace, and many fragmented art pipelines. All of this could have been solved with more experience and preplanning.

SOUND: The Good: I think, and according to a lot of reviewers, we really nailed the relaxed “Zen” feeling while playing XING. It would be my guess that the sound design played a large role in creating this feeling. I personally spent a lot of time making sure XING sounded “pleasant”, and I think the result works. This was a huge design goal of ours from the beginning - we founded White Lotus on creating peaceful, immersive worlds, and XING is a strong first example of that. Additionally, Koriel was the one to suggest adding voice acting to the game, which was ultimately a fantastic idea. I was initially a bit hesitant, worried that it would be difficult and impede on the relaxed nature of the game, but judging by the reviews, reception to the voice acting has been mostly positive, which is great. The Bad: Sound takes a lot of time - time I could have been spending getting technical details right or adding more puzzles / content. With more capital I could have outsourced a great deal of the more mundane effects. The Ugly: This has to do with more than just sound, but localization - or rather, the lack of. XING is not localized, nor will it likely ever be, simply because we did not consider it early enough in development. Technically speaking, some reasons why:

• Poetry made the built in Unreal text objects worthless due to hardcoded line breaks (translators would receive the text completely out of order with no context)
• No organization / centralization of text assets
• Text objects being saved on MAPS (terrible idea)
• No json / offline storage of text
• Just the simple fact that the entire game is POETRY - one of the hardest forms of communication to translate. Not only that, our game design has strict rules for how large a line of text can be on the physical surfaces. There are no traditional subtitles!
• Many of the fonts used in XING have no extended character support
• One gameplay element of XING has the player discovering letters to reveal hidden messages - I don’t even want to think about how that would get localized!
• Oh, and of course, we would need to hire out dozens of new voice actors if we wanted to fully localize! To put it simply, these issues really bum me out. I would love to see XING be played by everyone, not just English speakers, but our lack of forethought has really limited our options here. We would essentially have to rewrite a huge portion of our game, text, code, and design, to even begin to support localization.

MUSIC: The Good: As the composer, I am pleased and humbled to hear that people generally enjoy the music in the game. Aside from one particularly funny review I found that called the music “stupid and opera-like” people really seem to like the soundtrack. I remember the moments where we finally put in music in some of the areas and things started to “click”. The Bad: The dynamic music system - IE seamless music variations when the player changes the weather state - seems to mostly go unnoticed, which could be seen as good, but took me a long time to implement and maintain. It also didn’t seem to help the occasional complaint about the music looping being too obvious. In the future, I’d like to look into some more ambient ways of keeping the music fresh. I attempted to use motive writing techniques, inspired by Koji Kondo and more recently Toby Fox, but I’ve yet to have anyone point them out to me. Either the game doesn’t have enough of a following for people to write about it, or the motifs are too obscure (or are simply uninteresting!) Either way, I don’t think it matters too much. I’m just a music nerd and am simply happy I gave it a shot. Additionally, the game never really created any “hummable” tunes, something that would be nice, but more of a longshot. I’m able to sit down and play the literal main theme on my piano to my friends and chances are they won’t recognize it. The Ugly: I simply wrote way too much music. Here’s an example: I like to listen to game music while I’m working. I just listened to Martin O’Donnell’s amazing OST to the original Halo game - the OST is just over an hour long. XING’s OST is two hours and 56 minutes. I spent SO much time working on music when I likely could have written a third as much with double the quality.