Search Unity

  1. Welcome to the Unity Forums! Please take the time to read our Code of Conduct to familiarize yourself with the forum rules and how to post constructively.
  2. Have a look at our Games Focus blog post series which will show what Unity is doing for all game developers – now, next year, and in the future.
    Dismiss Notice

Gaia post mortem

Discussion in 'General Discussion' started by Radu392, May 13, 2020.

  1. Radu392

    Radu392

    Joined:
    Jan 6, 2016
    Posts:
    210
    Hi everyone, I'd like to put into writing my experience with my first Steam game, Gaia. There are multiple purposes for why I want to do this. First and foremost, I hope my experience will help future indies getting into this business with issues and tips that are rarely present, or not even at all, in modern indie game-making guides. I've read a few post moterms myself and they've been a great help, so I guess this is me returning the favor. Second, I'm hoping that some of the players who tried out my game read this and get a better understanding of the thought process put into it, because I'll be linking to it. Finally, I think it will be interesting to analyze the situation with input from other, experienced devs.

    So let's get to it! Warning: Long read.

    Background:

    Before I started working on Gaia, I've had some experience with Unity from around ~2016 when I spent a few weeks developing some 2D platformer that only ever got ~150 downloads on the app store. This is back when I was getting started with a Computer Science bachelor. Fast forward 2 years later, in the early days of summer 2018, I've decided to try my hand at making a serious game. I had heard many great stories about game dev, and I was almost done with University, so I also wanted something to show on my CV. I went straight to the Asset Store and downloaded a bunch of different assets, put them in a scene, and started thinking. I then came up with the idea of Gaia. I was trying really hard to find an 'original' idea for a game and I came up with this: A sci-fi base building survival game, featuring raids, agriculture, intelligent AI, trade and natural disasters. I wanted to basically make 3D Rimworld, but in a RPG setting, with no story, colonists or any other characters.

    Before Early Access:

    I thought this was pretty original, so I put a few assets together, made a trailer, took some screenshots, and put up the Steam storepage later that summer. To my surprise, the game got thousands of wishlists within a few weeks of the page going public. By July 2019, when the game came out in Early Access, it had gathered 23k wishlist additions and 1.8k deletions.

    Screenshot of wishlists from the day the store page went public to right before the game came out (day of release not included):

    a4VBSDNlCO.png




    So I put myself to work, wrote a (what I thought was) comprehensive GDD and started programming away. Progress was extremely fast at first. I was spending 12 hours every day during the summer 2018 hours and a bit less during school times. Lots of people were excited.

    I even put up a Kickstarter page in early 2019. Sadly that didn't work out. I recommend checking it out if you're interested in seeing how the game's progressed since then, as it still has the old trailer showing there. At this point, I knew I couldn't do anything about the animations and graphics that were going to be in the game. Not without any funding. Throughout the development cycle, funds were also too low to be able to afford any meaningful investment into hiring people to work on the visual aspect of the game.

    Pre alpha period:

    During this time, before the game was out in EA, I made a few builds and gave them to around 20 excited players. I allowed them to make videos on it and the general response was great, that is, if you count received feedback, which was from around only 4 players... At this time, I also found out that my game was extremely similar to Empyrion, a game I hadn't heard about before. In fact, some of the assets used in my game were the same as in Empyrion and some people pointed that out. This had only one outcome: accusations from some players that I was making a copy of Empyrion. At the time, I didn't take those accusations seriously. In fact, I was taking that as a good sign. I was thinking 'wow people are comparing my game to some super popular game!' Now, to this very date, people are still mentioning that and of course, always with a negative connotation.

    First mistake: Not doing enough research into the genre before starting work on the project. Had I known about Empyrion before starting the project, I wouldn't have gone forward with the idea, seeing as how similar it was. It wasn't just a problem of similar assets, but also of similar idea. Since I had put a year into development already, I wasn't going to just let it go however.

    During this pre alpha period however, the feedback was great. I had made my own website and had people report bugs and post stories there instead of over Steam. I think this might have been another mistake. My own forums have about 285 threads and 1.3k posts, the vast majority of which happened before the game was launched into EA. I think if I had kept the community all in one place, in this case on Steam's community hub, it would've been better. Instead, I chose to have people use my website and upon release, migrate onto Steam. I knew that was going to happen, but I went ahead with that plan anyway.

    Early Access launch:

    I was a bit disappointed with the EA launch. The game received about 20 reviews in the first week and stabilized at around 30 for the majority of the EA period. The score was averaging about 77%. Among the negative reviews, the reasons posted were mostly bugs, performance issues, old graphics and some gameplay mechanics. I'll address these later in the post. For now, let's keep analyzing the release.

    I got almost 2k wishlists on the day of the launch alone. Here is the graph for wishlists:

    X746hPRwBg.png


    If you compare the 2 wishlists graphs, the before launch and after launch, you'll notice that daily wishlists have dropped from around 30 to 50 pre launch to -5 to 10 post launch while the visits to the page increased. Possible causes:

    1. High price. The game came out with a 20$ pricetag and some players complained about it.
    2. Lack of hype. Usually, when a game releases, a very good indicator about the quality of the product is word of mouth. People did not post as much as I thought they would for example. There were no 'WOW THIS GAME IS AMAZING' reactions. The reaction was very mild.

    I'm not sure what else could affect such a drop in wishlists as the store page itself didn't change much. What was showing in the trailer pre launch was almost the same as post launch. The amount of visits, excluding the period around the release, was about the same.

    Sales wise, the game did not perform well. During its entire EA period, the game made around 2k sales. Much of the analysis on an indie game's future success is based off of wishlists. They always say 'the more wishlists you have, the more sales you'll make' and all those articles provide examples and formulas of how many sales you can expect based off of wishlists for certain periods. But it doesn't work like that. I was analyzing other game's launches and many of them made many more sales with a much, much lower wishlist count, as well as vice versa sometimes! I don't have that data anymore, but it is easily provable. Just scour a few dozen titles on Steam and you'll see it. There's also a list out there showcasing all the money made by every game on Steam, which was proven fairly accurate.

    Needless to say, these numbers reduced my desire to keep working on the project. But I kept going anyway, since i made many promises and didn't want to be another one of those devs who abandon their games in EA. So I kept working on it for another year, albeit at a slower rate.

    During this year, I've found out that, no matter how many promises I kept implementing, no matter how many new features, vehicles, maps, items, mechanics and so on I put into the game, players always wanted more. They always wanted something new, and so I obliged. My thinking was that if I kept doing it, eventually it would become a great, complex game. The players would see past the graphics, buy the game and therefore allow me to spend some money into modernizing the models. So I kept going from Alpha to Alpha. Each Alpha, major new additions were implemented and they were happening almost monthly for a year. It went from Alpha 1 to Alpha 7.

    Second mistake: Naivete. As you can see from the wishlist graph post EA launch, none of the updates had any effect whatsoever on either wishlists or sales. There aren't any upticks at all. It's as if those updates never happened. That is because it's mostly always the same players asking for new things, not new potential players. It's important not to lose sight of this.

    The truth is, it doesn't matter how much content your game has or receives. Some exceptions to this might be multiplayer games, or games that are so deep you can sink in thousands of hours into. But for single player indie games with no fame to their name, in general, the amount of content does not matter. What matters most is the quality of the mechanics and their harmony with each other.

    Third mistake: Going wide instead of tall. By that I mean, I've put in all these different mechanics in the game, but they were kind of superficial and didn't interact with each other very much. Gaia has underwater gameplay, raids, fire spread, psychological needs and so on, but all these mechanics have almost nothing in common. The more mechanics you have, the harder it is to link them together and make a deep game.

    Fourth mistake: Not choosing to develop one or two mechanics deep enough. I didn't know much about game dev when I started Gaia. I was kind of just going along. I thought that by just listening to player feedback whenever possible, it would be enough. But it's not. There are many many aspects to game development that are not visible to the naked eye, so to say. A few months ago, I read a book about it, written by Tynan Sylvester, and after reading it, I realized that I didn't actually know anything about game development. So if you're still reading this and plan on making games without having read any books on the subject, this is my most important advice: go read. Read a lot. Articles online, posts by other people and post morterms aren't enough.

    Final release:

    I've finally decided to release the game in April 2020. It's been over a month since release, so there's not much data to show. But I can say that it only got worse. The day of release actually made a lower amount of sales than the day of EA release (About 3/4ths of the cash). For reference, the page has had 150k visits over the last 90 days, with 78k of them on the day of release. By comparison, the day of the EA release had 49k visits and still made more sales back then. The visits to sales conversion ratio was therefore much better.

    There are a few reasons for that. First, I raised the price from 20$ to 30$.

    Fifth mistake: Raising the price. There were 2 reasons why I did it. First, it was to reward those players who supported the development during EA and second, due to the amount of new content that was added over the year. In a normal business environment, this is how it usually works: the more your product offers, the more expensive it should be. Not in game dev. There's a reason Steam recommends not changing a price after release and this is why.

    Second, player expectations increased by a factor of over 9000 due to the loss of the EA tag and the time passed. While in EA most players were forgiving of buggy mechanics and 'bad' graphics, post EA, players had much higher expectations. I suspect that is because the longer your game is in development, the more expectations rise.

    Sixth mistake: Taking too long to release the game

    These 2 factors shot up the refund rate by a few percents and tanked the review score from the 77% it had to 62%. This is most likely another factor why sales have dropped, though visits to the page remained high.

    Pirating was also a big problem. The amount of people who had a pirated copy was 3 times higher than the amount of people who purchased through Steam. When I first heard these news, I was quite devastated and looked into how I could remedy this problem for about a week, before giving up and realizing there's no way to stop piracy.

    Final mistake: Not having planned enough. This is mostly due to lack of experience. When I was working on the project back in Summer 2018 and was making huge progress, I wasn't taking into account edge cases. I couldn't have known that 90% of a mechanic takes 10% of the time to implement, while the remaining 10% of it takes 90% of the time to implement. And it's precisely this 10% part of a mechanic that really matters in the end. It's like having an artist draw a horse and leaving the head to be drawn by a 7 year old. So during that Summer, being so naive and thinking 'Wow it's so easy to make a game, let's put in ALL the features!', I went ahead and did just that and did not foresee what awaited me. Basically, it all came down to complexity. The game became such a monstrous complexity that I couldn't manage it anymore. I have this mentality that, if a mechanic should be in game, then it should not be limiting to the player. Therefore, if you want underwater gameplay and have cars at the same time, then YOU HAVE TO make it possible to drive underwater, what with all the physics restrictions on it. YOU HAVE TO make that transition between the water and surface work smoothly. And don't get me started when you start including realistic fire spread or flying... And so on and so on. And yes, you can drive underwater in Gaia, for a while at least. How cool is that.?

    But anyway, I personally think that having that mentality is an extremely good asset for any dev. However, it can only realistically be applied if your game doesn't contain many mechanics, or if you have large amounts of time (years and years) at your disposal. This amount of complexity will also exponentially increase your hunt for bugs and fixing your performance issues.

    Conclusion:

    There were, of course, other mistakes I made, such as only using publicly available assets. Also mixing unappealing models with appealing ones. However, I just wanted to highlight the biggest ones. As an indie dev, you have to pick your battles. You don't have infinite time or money, (unless you do, which is pretty cool, nice job), so you have to sacrifice some things. You're a good artist? Focus on a game that showcases that skill. You're a bad artist but a somewhat okay generalist like me? Then don't bother about graphics and focus on specific gameplay mechanics. I personally did not bother about graphics, because although I know how to make 3D models, it would take me simply too long to draw anything good.

    Gaia didn't perform as well as I was hoping, financially speaking. Before the EA release, I remember how much I was enjoying comparing its followers count to super popular games like the Heroes of Might and Magic series or some older Fallout games. I was having some grandiose thoughts and although they didn't happen, I did learn a lot from this experience. And this is another advice for you fellow, inexperienced indie dev: No matter how many books/articles/threads you read, you will still be lacking the most important attribute: experience. One thing is to know about it, another is to live through it.

    My future plans include making a Gaia 2.0 eventually. This time however, there'd probably be less features but more engaging and meaningful gameplay as I will be applying what I've learned.

    Thank you for reading! Feel free to post your thoughts or questions!
     
    Last edited: May 14, 2020
    Rich_A, SimDevs, alakoring and 11 others like this.
  2. BIGTIMEMASTER

    BIGTIMEMASTER

    Joined:
    Jun 1, 2017
    Posts:
    4,835
    great lessons learned!
     
    Radu392 likes this.
  3. Socrates

    Socrates

    Joined:
    Mar 29, 2011
    Posts:
    785
    Thanks for the post. It is always helpful to read honest self-evaluations from devs, especially the level of detail you went into.

    I have seen your Steam page before. The thing that really sticks in my mind is the image of the player swimming near that whale-creature-thing. My immediate thought was, "Nope. Not going in the water. Nope. Not happening."
     
    Radu392 likes this.
  4. jamespaterson

    jamespaterson

    Joined:
    Jun 19, 2018
    Posts:
    385
    Thanks for writing this. For what it's worth i think your game looks pretty cool!
     
  5. Zapan15

    Zapan15

    Joined:
    Apr 11, 2011
    Posts:
    162
    Thank you very much for your insights. We wish you the best!