I write a lot about games that succeed. For example, you can read about the millions made by Peglin, Tape 2 Tape, Cosmoteer, Battlebit, and Dwarves here. But there is always that one guy that yells “SURVIVORSHIP BIAS!” because I never cover the games that don’t succeed.
One of the reasons I don’t normally cover games that miss the mark is when a game doesn’t find traction on Steam it is emotional for the developers. I don’t want to pick some game to death while the developer is still licking their wounds over it. Also, the HTMAG community is pretty big now so I am basically directing a big spotlight on a small team and inviting thousands of people to also pick it apart like a kettle of Vultures. No developer is prepared to be picked to death by some random community. (So loyal HTMAG reader, please don’t pick people apart).
But, I made an exception after I watched this great post mortem for the city builder, automation game named Forge Industry. The development team did an honest, clear-eyed assessment of their own game. Their video got a lot of views which means the Vultures have already descended so the team probably has a thick skin by now.
I also thought Forge Industry was an interesting test case because the game is in one of the genres that I always say are perfect for Steam: a “crafty-buildy-game.” Steam players love crafty-buildy-games! So why didn’t it catch on? What went wrong?
After watching their postmortem I reached out to the team to make sure they were in a good place and to see if the were ok with me shining my spotlight on them. They were ok with it! They also shared additional numbers and context with me.
So let’s get to it. But first, I beg you, please, do not go to my Discord or take to twitter to pick apart this game thinking you know better. Let’s approach this with grace and be constructive.
About the game
Forge Industry is an automation city builder where you optimize supply chains for a local blacksmith.
Here is the basics of their campaign:
- The game launched Jul 21, 2023 and sold 104 units the first week for a gross revenue of $1824.
- At launch they had 2500 wishlists
- They had a 3.5% wishlist conversion rate.
- The game currently sits at 15 positive reviews.
- The sales per review ratio is about 8:1.
- They participated in Steam Next Fest and Dreamhack Beyond.
- They did not show up in Popular Upcoming, they did not show up in New & Trending.
- Development time: 18 months
- Team: 4 programmers who met at University while studying computer science.
The team was formed when they would play games at night and one of them said “Instead of always playing games, why don’t we just make them ourselves.”
What I think they did right
Let’s start with the good things first.
Good #1: Good genre
The genre you pick is the biggest marketing decision of your life. Building/crafting/city builder/factory games are huge for Steam and these games have an avid fanbase. I think indies should be making more of these types of games and fewer platformers, puzzle games, RPGs, and zelda-action-adventure-likes.
But be aware that just having the right genre isn’t a guarantee.
The dev team is made up of 4 coders which means their art skills were a limiting factor. The smart thing is that a city builder game is a great choice for a team that is more technical. Compared to something like a linear adventure platformer, city builders can be created with fewer art assets, less animation, and benefit from a deep well structured code base.
If art isn’t your strong suit, don’t try to make a 3D action adventure game.
Good #2: Good capsule
The biggest marketing mistake you can make is drawing your own capsule. Unless you are a full time, professional illustrator (not just a graphic designer), hire an artist. This team did that:
We reached out to a fiverr Artist, who we gave a stick figure drawing of a person on a mountain, hammering on an anvil. She had only a few other art pieces.. But we figured we would just give it a go, as it was only €120 at the time of ordering. Later on we also asked for a vertical extension of the main dwarf, which cost us another €30.
Good #3: Quick development time
The team came together and launched their game after 18 months of development.
We didn’t work 18 months full time on this. We did start development in January 2022, and released in July 2023. But as we all did this on the side of a regular 40h workweek, the amount of hours put into was limited, and there was a period of 1-2 months where we didn’t do much substantial work.
If we would have gone into it fulltime from the start, like we are now, I think we could have reached a similar result in 9 months instead.
Why is this short dev time good considering they failed? Your first game is most likely going to fail and the faster you can move past it the better. I really worry when first time game developers spend years and years working on a game.
The team is also already planning their second game. It is good to get back out there. As I have shown most teams make 1 and only 1 game. That is so sad.
Good #4: No Sunk Cost
This might sound like a negative but also a positive so bear with me.
The Forge Industry Coming Soon Steam page was launched about 1 year before release. Which is good! But in that time they only earned 2500 wishlists. That is bad!
Now I have always recommended that you try to get to 7000+ wishlists to ensure popular upcoming placement. However, if the game just isn’t hitting with the public and you can forecast that it will take years to get to that magic 7000, it’s time to button up the game, ship it, learn from the experience, and move on to the second game. I have seen too many game developers waste years of their lives trying to market and iterate on a game that just isn’t that interesting to the public.
“Sunk Cost Fallacy is the phenomenon whereby a person is reluctant to abandon a strategy or course of action because they have invested heavily in it, even when it is clear that abandonment would be more beneficial.”
— Oxford Languages definition
Because the team was so willing to move on from this project I think it shows they really understand priorities and long term they will find something that connects. This is healthy.
Good #5: High price
The game’s launch price is $16.99. In a blog post I wrote last year I found that the median indie game costs $7.30 which is actually $1 less than the median indie game in 2011. The only way we are going to beat this devaluation of our art is by raising our prices. Even Steam is trying to force us to increase prices. See this latest minimum pricing rule change from Steamworks
“The minimum base price that can be entered for any and all products on Steam must be at least equivalent to Steam’s recommended currency conversion for the $0.99 USD tier. For details, you can see a breakdown of minimums per-currency… If your product’s base price is below Steam’s lowest possible transaction price in a currency, it is no longer available to purchase in countries that use that currency.“ – Steamworks
I actually think this is a good rule change by Valve because it will force indies to push their prices up.
Now you might say “Well the game sold terribly, shouldn’t they have dropped their prices?” We will get into this below but I think the price was not the reason the game didn’t sell well. I don’t think dropping the price would have significantly increased the number of units sold. So good for them for not dropping the price to some $5 game.
What I think went wrong
Now, it is a miracle that any game is completed and I think it is a fantastic accomplishment by this team. Also I don’t like the term “failure” because every single creative endeavor must start somewhere. Your first game is supposed to fail.
HOWEVER, here are some things that were done or were not done that I think negatively impacted their chances. These are ordered by biggest impact first.
Improvement point #1: Graphics
A game’s graphics are very very important to a game’s success. They shouldn’t be, but they are. With 4 coders, there wasn’t someone on the team who specializes in the visual arts and I think the game shows that. The game uses asset store art, which actually ok. I don’t think that is a problem. However, a game’s artistic style isn’t just the models used in it. It is how the models and overall style work together that determine if the game “has good art.”
I am no environment artist or technical artist so I can’t be specific but there are definitely some things that can be done to just make it look more professional. Lighting, shaders, particle effects, all make a game feel more lived in. Even if you can’t hire an artist part time, see if you can find a freelance art director who you can hire for a few hours per week to consult and point out some areas where you can stretch to make the art look more professional.
Now, there are games that can get away with programmer art but they usually have to make up for it with infinitely deep gameplay. Caves of Qud and Dwarf Fortress are excellent examples of this. Based on browsing reviews I don’t think Forge Industry reached the necessary levels of depth to spread by word of mouth and overcome the limitations of the graphics.
Art direction and style is a holistic experience. This is why I don’t think AI art will magically make every game look good. It requires finesse to get things to work together and still look good.
Improvement point #2: UX and Playtesting
This is a comment from a user on the Forge Industry Forums forums. One of the first things that new game devs need to learn is how much UX hand holding your game needs to do. You don’t realize how much modern games do for us behind the scenes to make our experience feel smooth. When you are making your first game, the amount of edge cases, helpful affordances, etc really add up and can hurt your game if you don’t implement them. This is why so many first games “fail.”
I played the game and I got stuck. I wasn’t sure if I was stuck because I didn’t click the right worker and gave them the right job or if the game was just operating as it should have. These factory colony sims are very complex and rely on easy to understand menus and commands. A lot of UX playtesting, and rework are needed for any game to appropriately communicate with players. Mess it up and players will bounce off your game and try another one.
To discover and fix UX bugs, your team MUST do 1×1 playtesting and constant iteration on the first 30 minutes of game play. Here is what the game’s developers had to say about Playtesting:
“The playtesting was linked to when we went to events to showcase our game mostly. We then also started gathering a small following in Discord, who also did some playtesting when we were approaching Next Fest, and had a new playable build. But for the period between November 2022 and April 2023, we didn’t do any external playtesting at all.”
While it is good to watch people playing your game during conventions, they are not a substitute for 1×1 closed door playtesting. I have written about how to playtest using Discord here and other cheap feedback methods here. The reason (especially for complex builder games like Forge Industry) live conventions are not a good place to playtest is that show floors are loud, distracting, and people are usually with their friends who are just sampling little nibbles of games from a bunch of booths. If you are making a deep game like Forge Industry you need to test it in a quiet, neutral environment where a player can devote 100% of their attention to your game because that is how they will play it at home.
Improvement point #3: Finding the fun first
This person from the forums is the reason I love player feedback and doing open development. He is really good at articulating perceived problems:
Striking the balance between tutorialization and fun is haaard. I have no advice other than to fix it. Sorry, this is not a game design blog. Look at someone else for that lesson. But it needs to be done, tested, and improved for every game.
Improvement point #4: Low median play time
This is a bit of a combination of point #2 and #3 but If people bounce off your game because it is confusing, or the fun isn’t obvious right away, they won’t play more and they won’t recommend the game to their friends.
This is why I think median play time is secretly the most important metric you need to pay attention to in your demo. In my survey of hundreds of games and how they sold, I found that median play times less than 25 minutes often portends weak sales. Builder games typically have median play times that are HOURS long for a DEMO!
The Forge Industry demo had a median play time of 9 minutes.
My theory is that the slow start of the game, confusing and unclear user interface, and tricky tutorial just turned players away before they could really get into the game.
I know people think marketing is just Tweets and Facebook ads, but if people don’t want to play your game, they won’t share it with their friends, they won’t buy it on day 1, they won’t join your community. Your game must be exciting and easy to immerse themselves in. There are just too many games out there that can take their attention.
Improvement point #5: Late Breaking Demo
The Forge Industry team only released the demo at Steam Next Fest and then released the full game shortly after.
As mentioned above, Forge Industry doesn’t have the best graphics. HOWEVER, that isn’t always a deal breaker if the game has a very, very deep gameplay loop. Steam players will overlook bad graphics if the game is fun.
But the only way that people will know if it is fun is if they can play it. You can’t tweet “Our game is fun!” and people will just believe you and wishlist the game from that. Instead, you have to give them a sample and let them figure that out themselves.
If you have a game where looks aren’t the primary selling point, get that demo out as soon as you can. A perpetually live demo allows streamers to find the game and play it across several streams which can cause the game to go viral among the streamer community. Read this blog post I wrote on how demos help your marketing.
Demos also allow you to work out lots of little bugs and suggestions from the community playing it.
Steam Next Fest should NOT be the debut of your demo, it should be the grand goodbye to it because you should do the last Steam Next Fest before you launch.
Improvement point #6: No online shows or Streamers
Online festivals like Tiny Teams or Steam Next Fest are the best, most consistent way to earn wishlists. However because the demo was delayed, the team did not apply for enough and therefore didn’t get in to any.
You can find a full list of festivals here that is constantly updated by the HTMAG community: Howtomarketagame.com/festivals
Streamers are also one of the best ways to get wishlists. But you can only really get streamer coverage when you have a Demo. Because the team only tried to get streamer coverage around Steam Next Fest, there weren’t too many who were open to making videos for them.
Once you have a demo, keep it up constantly and you should be constantly reaching out to new streamers to see if any will play your game.
Improvement point #7: Steam strategy
If your graphics aren’t appealing (improvement #1) and your game is not retaining users (improvement #2, #3, #4) then your game is essentially a bucket with a giant hole in the bottom. So any promotion you do is going to be like water through the proverbial porous bucket.
So once you get the holes all stopped up, marketing on Steam starts to become more effective. At this point, you need to think of Steam as a ladder where each rung of the ladder is another visibility tier. If you skip the first rung, it becomes nearly impossible to grab the second rung, and on and on. You can learn more about this here in this talk:
So here is the visibility tiers:
Rung 2: Launch and get 10 reviews ASAP.
Rung 3: Survive the Discovery Queue as long as possible.
Rung 4: Get enough sales to trigger New and trending (Blog from Game Discover Co).
Rung 5: Months later, update your game and get into Daily Deals.
Improvement point #8: Low wishlists at launch
Forge Industry launched with just 2500 wishlists so was very unlikely to get into Popular Upcoming.
As you can see from the ladder, if you miss that first chance to appear in Popular Upcoming, you don’t get that extra visibility, you don’t get those extra wishlists, and that makes getting 10 reviews and appearing in the DQ harder.
You also don’t have as many wishlists converting to sales which lowers your chances of hitting the threshold to get into New & Trending.
Improvement point #9: Discovery queue
When you launch, Steam throws you in the Discovery Queue (DQ). The more your game sells in the DQ, the more DQ traffic they give you.
In my own consulting experience I have found that Steam typically gives games 24,000 views right at the start. If the game does well with those 24,000 and sells a lot of units, Steam will give them more views, and more and more until the game stops converting. Games that are poor sellers typically peak at 24,000 and go down from there.
If I look at Forge Industry’s DQ peak at launch (they did get into the DQ before earning 10 reviews) but the number of DQ views was just over 1700 before turning downward the next few days. Once they earned 10 reviews in August, Steam gave them another shot at the DQ. But it was only about 6,000 views.
Based on my comparison to other games, I believe Steam tested Forge Industry in the DQ and it just didn’t convert well and so they stopped showing them in it. I think this low conversion rate might be due to the Improvement Point #1 Graphics and possibly the game not having a strong enough hook.
The DQ is below with the yellowish green line.
Improvement point #10: No launch discount
Unfortunately due to some miscommunications, the team didn’t set a discount at launch on Steam.
At launch Steam will send out a “Game is now ready for purchase” email to every Follower and person who wishlisted the game regardless of the discount you enter. Even if you do a 0% discount. That’s great.
But, what a discount does is it creates “Fear of Missing Out” (FOMO) in the shopper. When there is a limited time discount, people who are on the fence say “Ah I might as well purchase it now so I don’t miss out on this low price” It creates a firm date where they have to make a decision. If there is no deadline, there is no FOMO.
Some games can get away without doing a discount. However, those games are hotly anticipated, from major studios with hit games. Those hotly anticipated games typically have 100,000+ wishlists, median playtimes >30 minutes, and have Steamers begging to play it. Forge Industry does not have that “hot” status so really they should have done a deeper discount of 20% to induce some sort of pressure on people who were potential buyers.
Improvement point #11: Pre-sale keys
In October of 2022 the team got a booth at Gameforce which is Belgium’s largest gaming conference with about 20k visitors.
To help defray costs they actually had a pretty novel idea which was to sell keys to the game before it was released.
“If they seemed to be enjoying themselves, we offered them the option of getting a pre-order. We had two tiers:
€13 standard package, which just gave them a Steam key at a discount, and Discord access
€20 supporters edition, which gave you a Steam key, a 60x40cm poster, a small wooden tankard keychain, an extra in-game decorative building, and Discord access” – Marnix
When they sold these keys and trinkets at Gameforce, they earned €806 in revenue. That is an awesome return!
At first, this sounds awesome. It is creative. It shows hustle. It is great! I love it!
But the timing is wrong.
When a developer requests a Steam key for their game and sells them outside of Steam (as this team did) Valve does not count that key activation as a real sale and if that player left a review, it does not count towards the magic 10 or towards the number of sales needed to get onto New & Trending.
Your die hard fans are people you meet in person and give you their money for a pre-release copy. By selling keys early, the team took away the potency of their most eager fans. Instead of giving your early die-hard fans keys that will negate their review and purchase, you should explain to them how important buying the game Day 1 is, and how helpful 10 reviews are. You want to activate them the moment your game goes live.
SIDE NOTE THOUGH: I absolutely love the hustle these guys had selling keys at their show. Selling keys post-launch is actually ok because you are not so concerned with “the Steam Visibility Ladder.”
Instead, you should sell keys for your back catalog. This is why it is so helpful to have a back catalog of games that you have worked on. If you are attending a live show, you can defray the costs by selling keys to your old games while hyping them up to buy your upcoming game day 1. Just please don’t sell keys for pre-release games.
Improvement point #12: Demo at launch
The team left the demo up on launch day. Overall, I think demos are great for marketing! But if the game seems to be lacking excitement (which I think Forge Industry was with it’s sub 10 minute median playtime) then I say pull the demo in the weeks leading up to and the weeks after launch. You need to focus everyone you possibly can into buying your game. A demo can take away some of that excitement and hurt sales.
If your game is HOT and your demo is HOT and ends on a cliff hanger, if the median playtime is high, and you have data that shows that after the demo people turn around and immediately buy, then keep it up. But I didn’t see evidence that the Forge Industry’s demo was helping them.
SIDE NOTE:
At this moment in the video:
Thomas mentioned that after launch, when they posted their demo their wishlists went up, when they pulled it they went down. This tells me that it is good to keep your demo PRE LAUNCH, but I see those post-launch wishlists people saying “eeeeh, I will wait for a sale” and that is not great.
Other side notes:
Youtube marketing
The team built a lot of their marketing around youtube. Marnix made this graph which correlates big Yotube videos they released to wishlists.
I am of two minds about Youtube and devlog marketing. Youtube might work. There are definitely developers like Jonas Tyroller who make fun general public videos that support hit games like Thronefall, and developers like Dani who propelled KARLSON to one of the most wishlisted games on Steam.
Youtube is a very very tricky platform where you have to spend HOURS generating content that might work or usually doesn’t. It can be a huge distraction. Basically you have to be a youtuber that makes games for this to work. You can’t just be a game dev who sometimes throws a video onto Youtube.
If the game is good, you can market a game without Youtube.
But, if the team is having fun, they actually like it then sure do it. Just don’t let it be a distraction.
Here is a key quote from Marnix talking about their current Youtube videos
The “problem” is that our core audience is game developers, not people who play our genre of games. If we had profiled ourselves more as “entertainers” instead of “educators”, we would have most likely attracted a different crowd, which is more interested in the game, instead of the process of making a game. On top of that, our devlogs also were consistently the worst performing videos we uploaded each month, and we just did them because I didn’t feel like giving up, and wanted to complete the story.
Demo bots
All demos bring bots. The team mentioned they were excited because when they uploaded their demo they quickly got 9000+ downloads. You can see that with “Lifetime free licenses” here. Unfortunately there are Steam bots that “claim” every single demo that is uploaded to Steam. I don’t know why they do it, but they do.
The actual number of real people actually playing it is “Lifetime unique users”
The lesson here is to not get too excited when you upload a demo and you get a bunch of “licenses”
Summary
Although this game did not sell well, I think this team is quite good at a more crucial skill which is velocity and shipping. They just shipped a game in 18 months. A complicated one at that. I think one of the biggest problems indies have is they never release and stretch their development time on for years and years and years and then burn out and never release another game.
This team made a lot of basic Steam marketing mistakes but that is FINE. Everyone does. I did. You have to make the mistakes before you can make the right choices. The good thing is they made the mistakes and now they can move on with their next game.
What I would do if I were in their shoes
I think Forge Industry is in the right genre for Steam. I think it is the right type of game for a team of their skill set. However, they just made beginner mistakes.
I don’t know what their plans are, but here is what I would do.
- Quickly, make 1 more content update for Forge Industry that tries to address the simplest-to-fix complaints that Steam shoppers had. Don’t spend more than 2 months on this update. The goal here is just to show the Steam community that you are a “Good Dev.”
- End development on Forge Industry. Be up front with the community why you are moving on. Tell them how proud you are, how thankful you are for the community’s support, but also share how little you made. Again, you want to show you are a “good dev” that respects the players but you need to set boundaries and be transparent about how it isn’t sustainable to continue working on the game.
- Assuming the code is decent (they are 4 CS majors so I bet it is), I would build on top of the existing code base for a game in the same genre (automation) or adjacent genres such as city builder, colony-sim, simulator, or tycoon management games.
- I would focus a lot of effort and time playtesting the game 1×1 and dedicate care and attention to testing and improving the onboarding tutorial, UX, and ease of use of the game
- I would hire a freelance artist or art director to give them guidance on making the game look good and visually cohesive and improving the UI graphics. Automation games don’t have to be the most beautiful game on Steam but they do need to look professional. Metaphorically, if releasing a game were like a job interview, you don’t need to show up to your job interview in a $5000 Balenciaga suit but your clothes do need to match, and you need to wear nice shoes, shave, and comb your hair.
- I would leverage the small but dedicated community they already built around Forge Industry to playtest and help them develop their second game. By sticking to the same genre or sister-genre you are more likely to transfer that audience.
- For game 2, I would follow Steam algorithm best practices of getting a Steam page up ASAP, demo up early too, continuous outreach to streamers, and trying to get at least 7,000 wishlists or more.
- Once game 2’s Steam page is up, I would try making 5 new youtube videos and edit them in a way to try to appeal to a more general audience that is not game-dev focused. If after 5 videos there is no evidence that video engagement converts into wishlists, I would ditch youtube. Yes Youtube is fun and all the cool kids are trying it, but, if you really want to be a youtuber, just be a youtuber and give up wasting your time on this game-making business when you could become a youtuber who reviews types of cats or something. If your real passion is game design, then focus 100% on game design and don’t let youtube distract you.
Again this is just my advice. Don’t yell at the team if they pick a different path than what I recommend. It is their life, their artistic vision. I am just a guy on a website.
By the way you should subscribe to their youtube channel, they make more great videos like this: