After showing at the Portland Retro Gaming Expo, it became apparent that some of the bosses (I'm looking at you, Dodongo) still had a couple bugs. I took the opportunity to wrap those tweaks up with the bugfixes that I've done since Steamlaunch into a new patch.
I haven't done a boatload of testing with this version, but how bad could it be? You can opt into the "stable" branch if you would like to roll back to 2.12 for whatever reason.
I would like to hear what people think! Probably the most visible change is the doubling of the dialogue + portrait boxes. It was a little jarring for me, but there were enough people that couldn't read the text that I decided it necessitated a change (and thanks to pixel art, it'd have to be 2X). Plus, although it was designed to be somewhat ignorable, people were missing some important information. If this is abhorrent to you, let me know and I may put in a "small dialogue" option.
Fly true, my friends. - Wick
P.S. There is now an unlockable cameo ship from Hard Lander, which was made by my friend Nic Biondi who I've shown at a ton of conventions with. http://soulareus.com/
Highlights # bigger dialogue box/text # edge scrolling in starmap (ability key to re-center self) # hard lander cameo! # databank dropped by bosses, not given at start # improved bulwark boss -- bigger, less damage, more knockback # improved dodongo boss -- bombs do damage more consistently, it sits in the middle of a gravity well to make hitting him easier but more dangerous, minion summon tweaks # option to force it to use OpenGL graphics. # added ability to set arbitrary FPS in options file (semi-unsupported, I think it can do some strange things to gameplay) # alt-enter switches to windowed
Tweaks # beam turrets turn slower # tweaked unlock costs and information # allies aren't auto-repaired for the final battle # pakiceph : bigger, front armour, no shield! # "how do flamethrowers work in space?" description # slightly buffed default zukhov # "must have *rescued* at least one cryopod..." in stores # shotgun+cloak too powerful, increase reload times, nerfed cloak slightly # for multiple abilities in the statbox, show an X2 instead of multiple entries # revolver bullets more of a trail # moved store button to make it easier to notice # realhuman doesn't suck up points anymore. it was too bad an effect # press P at any time to print the current framerate to the log # reduced collision damage for allies # pilots start with 1000 points (so they can reset off the bat) # option to offset joy axis # bumped up starting points for sol to 350
Bugs # lias is too cool to hang out with you # fixed bug with final boss death # kickstarter shouldn't show up as an ability # FIX ALLY MULTIPLE BUMP (esp. dovell) # WHAT THE HECK IS THIS GRID EFFECT??? >>>> if you have it, just delete space_tile2.png from gfx/effectGraphics # bubble drone orbits too far # flushkeys on credit start # on death in hardcore mode, delete the save # nodes still target cloaked things # petey shouldn't target buffered things # cloak + death blossom stops most of the missiles from launching # options menu is slow again... -> it was the constant joycount() call # lias doesn't allow you to choose colors # a little smarter graphical compatibility tweaks
(Note: Valve has asked us to not share our exact sales numbers. I feel like that takes a bit of the teeth out of a postmortem, but there are still plenty of lessons to be shared)
I was going to write a standard postmortem -- explain what we tried, show some stats, and pull out some lessons. I'll do some of that later, but I think I'm starting to realize that any specific tips I can share is less important than something else I learned.
Steam traffic is a gigantic morass -- having my game on the front page that first morning felt like timidly standing in the empty floor of the stock exchange moments before it opens. Suddenly, before I can take a breath, the wave of humanity hits. One million (1,000,000) views on the front page, Steam promises. It only took a couple hours. Store page clicks were two orders of magnitude lower than that.
Final sales? Two orders of magnitude lower.
Not gonna lie, I was disappointed. I'd seen the ocean and only felt a drop. I felt that the store page must have failed somehow -- should have used more graphics, should have rewritten the summary again, should have included a demo. The game mechanics were good, but it needed a better hook. It's just another clicky-explody spaceship game; nobody’s going to say, "ooh, tell me more!"
"As a book editor at Big Five publishers, it never failed to astonish me when enthusiastic tweets to 7-figure "followings" failed to sell a single book. Unlike Soylent Green, Twitter is not made of people."
I'm starting to realize what a messed up set of expectations I had. It’s easy to get sucked in and blinded by the numbers above all.
> The #1 lesson to myself is to calm the heck down. Yes, there are a lot of people in the world. Your game is not going to speak to all of them. Concentrate on the people who *do* engage with whatever mess you've hung out to dry. A single enthusiastic fan is worth fifty purchases. If your audience turns out to be small, well, you have to weigh your economic ability to continue versus how much you believe in what you're doing.
> Frantically grasping for reviews, exposure, any way to get the word out! is a dead-end shout into the void. Coverage isn't something to accumulate, the let's players and reviewers who take a look at your game aren't (or shouldn't) just be auxiliary megaphones. I think it's a lot more fascinating to hear where *they* are coming from and why they'd want to spend their limited time on this earth playing a clicky-explody spaceship game.
> I want having an online "presence" to be something that's sustainable. I don't want to feel like an advertisement bot. I was watching robotloveskitty's Twitch (http://www.twitch.tv/robotloveskitty) the other night and she mentioned that streaming held three attractions: 1) she got to play indie games 2) she got to meet and talk to indie devs 3) it let people know that robotloveskitty is a thing
Tying promotion into things that you enjoy doing anyway seems like the way to go. If anybody has more suggestions or good examples of people doing this, feel free to let me know.
Overall, I’m happy with Starship Rubicon. I made enough to buy all my friends sushi (which was the real goal of this three-year project). We’re sitting on Steam with a tentative 100% thumbs-up. It’s been an experiment in making a game with conservative design (woo spaceships) and great execution (“hey, this is actually pretty fun!”). Maybe the best thing has been to have met people doing the Kickstarter and watch them support me all the way to Steam. I can’t describe how great it feels to have people believe in you like that.
I’ve only just begun to work. Great things are ahead.
*** OK, /EMOTIONAL REVELATIONS. HERE'S THE STEAM REPORT: *** (probably more relevant to soon-to-be-Steam-devs than the general public, but here goes for anybody who is interested:
Steam Integration + API
Steam is strange as far as stores go because it has all these metagame layers it wants to slather on top of everything. Achievements, the Steam Workshop, leaderboards, and trading cards are the big ones. If you want to include these, it requires you to actually modify the code of your game (instead of just handing them a zip file and calling it a day).
I didn't even know what an API was when I started, and since I use a sorta-obscure language (Blitzmax! It's great.) there weren't a lot of out-of-the-box solutions for using it. It was close, and if I'd given myself more than a weekend, I probably could've figured it out. However, I ended up shipping without using anything from Steam's API (so no achievements/workshop/leaderboards). Honestly, I think it worked out fine; if the game picks up later, I’ll spend more time adding them in.
Steam trading cards just required sitting down with the site for a day and doing a lot of graphic design work. It has a useful checklist and there are some opportunities to be playful.
There's a non-API method of doing cloud saves, which is what I ended up using. You just specify in the browser which files to save (/pilots/*_save.sav) and POW, you're golden.
Once you get the hang of it, Steam's backend is pretty nifty. It allows you to manage what files go in which packages/builds on what platforms and for updates only pushes out the small chunks of files that changed. The documentation is sparse but there are a couple of great video tutorials.
The summary is super-important, I think. It's what people will see when they hover their mouse over your game. Look at some other games for how do it (Team Fortress 2's is the best I've seen - short, descriptive, funny). Similarly, there are a lot of miscellaneous graphical assets Steam wants for displaying the game that are very important as they are the only thing that you're guaranteed 1M views of.
Good screenshots take a lot of time! A good trailer takes even more time! Spend that time. I tried to show off different systems (shooting, node travel, ship customization) and cool set-pieces or enemy setups.
I, ah, made and *educated guess* about the minimum system requirements by looking at similar games (will I get in trouble for admitting that?). I ran it on an old netbook once. So far nobody has mentioned them.
I wish we had a demo when we opened the doors. I don’t know if it would have made a difference in the long run, but getting that initial wave of visits who didn't engage at all to do SOMETHING would have been nice.
We decided early on about the $9.99 price point. Sales feel substantial, it seems to be the market standard for this scope of game, and high enough to generate some actual revenue (I don't understand how somebody *could* afford to sell a game for less than $5).
WE LAUNCHED THE SHIP INTO THE STEAM!!!
Valve has to manually approve your store page before you can launch, which only happens on weekdays. We waited til the last minute, but it'd have been better to get it ahead of time so we'd have had more control over when we went live.
Apparently, you're guaranteed 1M views on the front page in the "new releases" box. Don't get too excited and think that it meant that many people were looking at *your* game. This guarantee happened in about four hours -- from 8:36 AM to 12:26 PM. They only bump it up into additional visibility if it sells a lot of copies. Starship Rubicon... did not. Traffic, as expected, plummeted once we were off the main page & the first page in the action/indie tabs.
I did notice that over the course of the 1M views, the click % (from the main page to the store) went from around 2.5% to less than 1% (the global average was around .3%). My theory is that people are more willing to look at something new early in the morning than later in the day.
We also didn’t realize how important Steam reviews are for a new game. The page was felt barren without an indicator for what the “Wisdom Of The Masses” thinks. Thankfully, a couple of super-awesome Steamers played the game, enjoyed it, and posted their reviews within a couple hours. To solve this, next time we will run a longer beta and ask people to pretty please review the game before launch. That should help people make an informed choice as nobody wants to buy a game no one has rated.
I looked for any interesting stats to share with y’all, but I think it’s a pretty typical story. Spike of views at launch, exponentially decays, stays at flatline until disturbed by any singular bursts of exposure.
There were a couple of technical issues that came up in the first couple days, which I took seriously and handled promptly. I assume that for every person who says something, there are a bunch more that don't. I think it paid off; several reviews mentioned the technical problems, but included that they were quickly resolved and ended on a positive note.
The Steam community hub is probably my favorite thing about the platform so far, since it gives people an easy centralized place to look for tech support and ask questions about the game. I also made one of the steam trading cards out of one of the first screenshots that were posted (because I thought it was funny).
jdodson: I’ve been happy with how well the game has been received and plan on promoting the game as long as we can. We are getting a healthy stream of Let’s Play and review requests and I like responding to them and seeing what kinds of videos come out of that. This summer we will be bringing Starship Rubicon to a few local video game conferences and Wick and I are working on a talk about all our experiences with launching the game. We’ve talked about showing at bigger cons, but the massive cost of showing at larger events like PAX doesn’t justify the costs with our current sales but I wish it did. I plan on writing a post at some point with some tips for YouTubers and review sites to increase your chances of getting a review copy. We usually error on the side of being generous but there are some scams out there for reselling Steam keys and a badly written request can look like a scam pretty quickly. Let’s Plays and reviews are fantastic for our game so it’s better for everyone if we can get review codes to every legitimate reviewer that wants them.
Wick: I’m also going to add in a demo and maybe retool the store page again before we run any sales on Steam (which is usually a major traffic bump at the cost of devaluing your product, so this’d be way down the line). I see it as an opportunity to mess around and try some different things to hook people.
Who knows what else the far future might hold? (spoiler: probably bundles)
jdodson: We will include the game in some kind of future bundle but the trick is to include it in the right bundle. Since Wick and I didn’t bet the farm on Starship Rubicon selling enough to afford us both mansions and a bling-mobile we don’t need to act rashly. A few bundles have reached out to us but we are holding off on pulling the trigger on that until it’s the right time and bundle.
Summary The biggest surprise for me is still the dropoff from store views to actual purchases. I consider it a hard lesson in the need for a hook - a single element that you can rely on to make your game instantly pop. Art style, compelling concept or scene, neon novel mechanic, whatever. The thing that they’ll see in the four seconds they’re on the page and will stick with them after they leave. Rubicon has super solid gameplay, but I’m sorry to say is sorely lacking a hook. The NASA backgrounds are sorta cool but I didn’t pump enough juice into them.
So: consider this a saga in a hella indie game getting onto Steam and not necessarily making it big. And you know what? That’s OK.
jdodson: When I talked to Wick about having Cheerful Ghost publish Rubicon I pitched him the idea of bringing the game to a larger audience. I noted that this would be an experiment as both of us had never sold a game before. If you go to YouTube and search for “Starship Rubicon” and scroll through pages of videos and then stop by Steam and check out the reviews and the community activity we accomplished our goal. Hell yeah!
Wick: I think our traffic graph (at the top of the post) speaks for itself. We had a huge initial surge (I assume from being on the front page of Greenlight for being “recently submitted”) which QUICKLY dropped off over the week as we moved further and further away from the front page. The later minor spikes came from exposure through Let’s Players and Twitch streamers.*
Then came what is only known as the Dark Times, where we weren’t actively promoting it and were more-or-less sitting pretty at a steady ~40% of the way to the top 100. We were banking on getting more bumps down the line from doing some conventions and maybe a bundle. Then BAM! Out of the blue, Greenlit! So what on Earth happened??? Why did we get Greenlit when we did? Read on for my best guess:
* (huge shout out to Stream Friends, Maris from GamerQuest, and many others)
If you notice one thing in our post mortem it should be that I am mostly going to agree with Wick on just about everything. He is not only smarter than I am but he also is really good at spinning fire so it’s best to avoid major disagreements.
Like Wick said, the first large traffic spike is pretty apparent. Front page Greenlight means votes. Some people have said having an animated GIF means more people will check you out, unsure if it’s true but I found other Greenlight games animated GIF’s distracting.
The other upticks in traffic was from our December 12th 2014 launch. As you can see the numbers start to head upward as the game generated more buzz. We launched a PR campaign on Cheerful Ghost, social sites as well sending out the usual press kits and review keys to game sites and the Let’s Play community. All press coverage we received, which wasn’t much, was all very positive. That said, the majority of our buzz was from from the Let’s Play community. As much as I think I can craft a good press kit nothing beats seeing someone play the game and have fun with it. The night Wick and watched Stream Friends play the game on Twitch was magical.
Wick: Here you can see how our project was faring against the average top 50 Greenlight projects. The most obvious discrepancy is the traffic volume: about 3,000 visitors for us versus 20,000 for them (i.e. we had 14.4% the traffic they did) - an entire order of magnitude difference! D:
It gets a lot more encouraging once you start looking at the actual response we were getting, though: with only 14% of the traffic that they did, we had 34% of their favorites, 40% of their followers and 21% of their “yes” votes! That is, even though we had much less traffic, each visitor to Starship Rubicon was around twice as likely to respond positively than a visitor to a top 50 project.
Interestingly, we still had 13% of the “no” votes that they did -- basically no per-visitor difference. I guess that there’s just a standard “I don’t care about this game” level that is applied to all projects.
(per-visitor) rubicon avg top 50 any vote: 75% 70% yes votes: 30% 20% no votes: 45% 49% favorites: 3% 1% (followers is basically the same)
jdodson: It’s nice to see that Rubicon did better than the top 50 when our numbers are averaged out. I have some guesses as to why we were able to do this but none of them may actually be accurate.
Our Greenlight was basically a Kickstarter pitch. From the start of working with Wick I knew I wanted to focus on him and his story and talk about the game secondly. When you hit the Greenlight page the video automatically plays and you get a few seconds of gameplay and it’s Wick talking about the game. My only regret is the “bouncing on the moon camera work” which is totally my fault. I was trying to keep the camera moving to give his talk some momentum but I went a bit off the rails. After looking at a ton of Greenlight pages and really enjoying some recent video game documentaries I asked Wick if we could focus on his story as the first push for marketing the game. I think this personal touch at first contact with the game helped soften the audience up to click yes. One thing I told Wick early on is people might not give a shit about the game but people are generally interested in hearing someone's story.
People that came to our Greenlight page were mostly from Indie / Greenlight friendly sources. We received no coverage on Kotaku, IGN, Eurogamer, Gamespot or PCGamer. Our initial spike was from the Greenlight front page, social sharing & Cheerful Ghost. Our later bump was from the games launch, Let’s Plays and Twitch streaming. I postulate these sources are more friendly than larger sites. “Indie sources” have readership that might not always look for the next shooter or sport title. As such Rubicon is something they may enjoy whereas a more mainstream audience would never consider due to its retro charm.
Wick: Eh, I didn’t find this graph very useful. It doesn’t say anything new (besides it looks like top other items all get giant spikes rather than gradual climbs, probably from specific exposure sources).
Why is there an “ask again later” option? Literally nobody I can see uses it.
Wick: Now for the Analytics data! ... Huh. Make sure you have screenshots and a video that gets to the point, I guess. I wasn’t expecting the peak to be so drastic. The dwarf peak at 61-180 seconds are the people who actually watched the trailer, is my guess (it’s also pretty close to the # of people who came here through facebook…)
Wick: Seeing that our traffic was dominated by Steam Greenlight campers, I’m guessing that this is just the readout of that population.
I’m not showing the map for space’s sake, but I was surprised that only 44% of our traffic was from the US! Next was Russia at 8%, followed by Germany/UK/Canada/France/Austrailia, but it’s a long tail. Apparently Steam is friggin’ global.
Two more minor notes: -- Social (fb / reddit) traffic was only 8%. I wish I could compare their responses versus people who found it through Steam. My money’s on them being waay more likely to give it a thumbs-up, but it would be nice to be able to see how large an effect it had. Is our better-than-average positive response due to a small population of supporters? -- Analytics saw a coupla significant spikes in visitors in the week or so before it was Greenlit that don’t show up at all on the Steam graph. Bots that got filtered out? Valve internal traffic that somehow doesn’t count? Who knows.
Whew. Still with me? My takeaways from the whole thing:
Steam is massive.
Your pitiful-but-valient efforts at promotion for months will be completely dwarfed in volume by the sliver of people who happen stumble onto you through Steam in a single day (see point 1.)
You should still be upbeat and talk genuinely and politely to everyone who will listen anyway. Again, I can’t back this up with numbers since Steam doesn’t show detailed traffic data. BUT! The couple hundred (i.e. dozen) supporters who enthusiastically follow and support you are way more valuable in the long run than two thousand people just passing through.
The selection process is opaque and sorta hit-and-run.
I wanted to wrap this into another post about the process of technically integrating into Steam, but people were asking about the process and it had already gotten sort of long, so expect another postmortem sometime down the road!