Magfest Postmortem

I just got back from Magfest, where we exhibited Steel Carnelian publicly for the first time. This is our second time exhibiting there, and we came in with a lot of ideas about how to make our booth work well. I’d like to go over how the convention went, both as a reflection for myself and also as a resource for other indie devs considering tabling at similar conventions.

Before getting into things, I’d like to provide some context: We’re a three person dev team. All three of us attended the convention and spent time managing the booth. At the time of the convention, we’d been developing Steel Carnelian for about a year and a half.  We plan to release a larger, public demo in a few months.

Preparation and Setup

Our first point of concern was getting working 24/7 setups for the event. Exhibiting games can be pretty scary for indie devs, as most of the PCs or setups available to a dev will be their personal property - often their work computers. When exhibiting our last game (Splatter) at Magfest ‘22, we had brought two desktop towers from our homes, and out of concern for their safety we had stashed them away every night rather than letting the demos run on their own. We didn’t have access to “expendable” PCs this time either, but we still wanted to run our setups 24//7. Our solution was a wire dog crate. The desktop towers were placed inside the dog crate, which was stored under a table and locked securely. We could run wires into and out of the crate, but players couldn’t jostle or steal the PCs without significant effort. This worked out excellently, and both towers ran the game for the full duration of the convention with no hiccups.

During core hours, we also ran a third setup from my personal laptop. This gave us additional throughput during peak traffic (which we did end up needing!) and was easy to set up and take down - when we left the hall for the night, I’d just unplug the laptop and carry it up to our hotel room.

We planned two of our setups (the desktops) as KBM setups, and the third as a gamepad setup. None of us wanted hundreds or thousands of gamers getting their grubby little paws on our personal gear, so we picked up a pair of Amazon wired keyboard / mouse combo packs, which did their job and only cost like $10 each. Similarly, we bought two pairs of cheap headphones to avoid sacrificing our own. The only place where we caved and used personal gear was with the gamepad (shoutout to our environment artist Mattie’s roommate for letting us use your PS5 controller, it worked wonderfully).

Two of our setups used standard PC monitors, while the laptop setup used a much heavier TV with built-in audio. The TV setup functioned both as a way to avoid having to buy another pair of headphones and as an arcade-like attract video as, after a period of inactivity, our demo would load a loop of our game’s trailer. Quick note - if you’re using FMOD and Unity, syncing audio to a video player is a major pain in the ass. I decided it was too much of a pain to handle and just stuck a loop of one of our boss themes over the trailer video. This worked fine, but if we go again I’m setting up some kind of soundtrack medley instead because holy shit listening to the same 2 minute music loop gets so old so fast.

Our decor was fairly limited. Mattie works at a print shop, fortunately, so we were able to print out a standing banner and a large backdrop for extremely cheap. Aside from that, though, we just stuck with some B&W posters on printer paper, tablecloths, a rug, and a pair of beanbags for additional seating. We certainly weren’t the most refined booth setup at the convention, but I wouldn’t have gone much further with our look - our stuff was plenty of work to pack in as-is.

We arrived on Wednesday night and set up our booth before going to bed. This cost us an extra $200 or so (hotels aren’t cheap!!!) but it was much less stressful than our 2022 attempt at morning-of setup. Given the option, I will absolutely never attempt to set up a booth on the morning of a convention again.

Our actual setup process was mostly hindered by the fact that nobody on this team owns a large car. Mattie and I both filled our cars and barely had enough space for everything. This not only meant a stressful packing experience, it meant that our load-in process involved walking all the supplies from one car into the hotel by hand and then using the loading dock with the other car (as well as additional charges for parking, as our hotel reservation only gave us one parking space). If we exhibit again in the future I am absolutely looking at renting a minivan for the weekend. My only other note on setup issues was that we didn’t pack any USB or aux extension cables, limiting our ability to place our setups far from the dog crate or laptop. I’d recommend this to anybody who doesn’t have time to test-build their setups before the convention - they’re cheap and they’ll give you a lot of flexibility, which can be a real lifesaver.

The Convention Experience

The Steel Carnelian convention demo was about ~20 minutes long, and contained two bosses. Again, this was our first time bringing the game to any kind of public audience, so we learned a lot of things about the game and about how to exhibit it to our audience.

We did make a few modifications to the demo so that it would be appropriate for a convention. We brought in a limited-lives system to keep players rotating off the setups, and an attract screen that would kick in after a period of inactivity to prevent memory leaks and to give new players an easier startup experience. Both of these features ended up being a bit flawed: the difficulty spikes in the demo caused players to burn through their lives faster than expected, and the attract screen kicked in too quickly (after 45 seconds of inactivity) so players would often get distracted and turn back to their screen, only to find that their progress had been reset. While both these features did have benefits (we did need the rotation during peak traffic, and the attract screen kept our demos functioning for a full 86 hours) I’m definitely going to tune them further for future demo exhibitions.

Our difficulty curve and tutorial was also something I’d like to refine for future demos. While we were able to convey most of the game’s mechanics, players consistently struggled with one of our final tutorials (targeting weak points) and a mechanic we forgot to effectively tutorialize (parrying purple projectiles). This resulted in a major difficulty spike during the first boss of the demo, which wants players to focus down its weak points in its first phase and parry its projectiles in its second phase. Not all players were stumped by this, but it did filter many people and that’s really the opposite of what we want to do when exhibiting a game to the public.

Even among players who did figure out its mechanics, the first boss was a major difficulty spike. This not only turned some players off, it created a false “end-of-demo” impression for players who struggled to beat it and sat through its ending cinematic only to find out there was a whole second level afterwards.  Some of those players chose to get up after realizing that there was more, which isn’t always avoidable (not everybody’s going to like every game enough to play through a full 20 minutes of it at a convention) but we probably could have kept some of them longer if they hadn’t worn themselves out on the first boss fight.

Almost every player who made it through the first boss made it to the second boss, which is once again a strong indicator that we need to adjust our curve.  The second boss did burn through most of those players’ remaining lives, but I don’t think that’s a real problem, as she wasn’t blocking any content.

On a technical level, the game mostly performed fine. We had a few recurring bugs, but nothing that actually locked progress for players. Our only real issue was with the controller support.  While the controller’s aim was incredibly sensitive in the editor, it somehow locked to a very low sensitivity level in the build and we didn’t notice until day 1 of the convention. Test your builds, folks! Fortunately, our UI designer Ethan had brought a copy of the project on his laptop and we were able to push out a revised build with better controller support on Thursday night. I wouldn’t recommend planning to fix your game during the convention, but it certainly was nice to have the option!

So far this has all sounded pretty negative, but reception to our demo was actually very enthusiastic overall. Magfest was in many ways the perfect audience for Steel Carnelian - a ton of old-school arcade game fans attend the convention each year, and many of those players loved the idea of a fixed-camera rail-shooter run-and-gun sort of game. It was a wonderful experience getting to chat with those folks about the types of games we both liked. They’d often guess specific games that influenced us, and most of the time they were correct!

Even outside of our core sicko audience, players generally meshed well with the game’s format and style. I was worried the fixed camera would be a turnoff for many people, but even players who weren’t immediately enthusiastic about it could usually play fine with it after the first few rooms. The cartoony pixel-art style was also a hit, even among players who found the demo tough. At the moment, I’d say we have strong appeal within our niche, and with further refinements to onboarding I think we can reach a much wider audience.

One thing I wish I’d had time for is some kind of automated stats detection from our demo. I’d like to see things like average time in a level, rooms where players died, etc. This had to be cut thanks to our relatively limited time to prep the demo, but it would have given me a lot of actionable information for future development.

It’s worth noting that the convention didn’t score us a lot of direct metric growth (in wishlists, social media followers, discord members, etc).  We absolutely expected this: demo players at conventions don’t convert to permanent followers at a very high rate no matter how good their play experience is.  Our primary goal for this convention was seeing general-audience reception to the game before our full pilot release, so that when we’re ready for a serious marketing push our demo and supplementary materials will be as refined as possible. I’d recommend this mindset to any indie looking to table at a convention. Don’t go with the assumption that the convention will sell your game, go with the assumption that the convention will teach you how to sell your game.

Teardown

Part of the convention is leaving it in a timely manner. Magfest requires all booths at MIVS to remain open until Sunday at 2 PM. We all wanted to be gone as soon as possible, though, and figured out some solutions to make our teardown process after closing quick and easy.

First, we chose to only run two setups during Sunday’s core hours, replacing one of our two desktop towers with my laptop. This let us pack away one of the towers, the beanbags, and the large TV early. If we had planned a little more cleanly, we could have replaced both desktop setups with laptops on Sunday and cleaned up most of our heavy gear before closing.

The other key point of our strategy was that we ran our supplies back to our cars manually instead of using the loading dock.  This was extra manual labor, but we could do it before the loading dock opened (at 4 PM, a full two hours after closing). This was only possible because friend of the team Erin lent us her collapsible cart. Thanks, Erin!

This type of teardown strategy varies intensely depending on your con’s location and schedule, but no matter what it’s important to have a strategy in advance.  By day 4 of a convention your whole team is going to be exhausted and ready to go home, and the longer it takes you to pack up the worse everybody’s going to feel about the whole experience.

Conclusion

Personally, I had a great time exhibiting the game. On a professional level, I felt it was also extremely useful for us.  That said, I don’t think Magfest is necessarily an every-year experience: we’ll probably still be working on the game next year, but unless we have something new to put in front of a convention audience I don’t know if it will be worth the time and money to table.
That said, I probably will attend the convention even if I don’t table. Getting a chance to chat with other indie devs is invaluable! I learned a lot and met some folks who I’d previously only known online. Also, where else am I gonna get a chance to play Virtual-On on the original hardware?
For context and self-promotion, here’s the game we were exhibiting.  You can wishlist it now if you want.