How Early Access Can Work for Episodic Narrative Games — A midpoint post-mortem
For those of you who don’t know me, I’m one of the two developers of Scarlet Hollow — a choice-driven horror visual novel releasing episodically through early access. Episode 3 of 7 just released on March 7th, a little over three weeks ago as of the time of this writing.
The games industry as a whole has a pretty negative slant against early access — a lot of publishers don’t want to touch it, a lot of players have publicly stated that they’ll never buy early access titles until they’re fully released, and there’s a strong consensus that releasing narrative focused titles in EA is a death sentence. That hasn’t been the case for us, so I wanted to put together a post about our general marketing strategy as well as some of our key milestones and metrics to serve as a counter-example.
Before I dive into things, a quick definition of early access for those not in the know. Early access is a release mechanism for games that are still in active development. Examples of popular early access titles include Phasmophobia, Valheim, and Baldur’s Gate 3. On paper, this release mechanism serves a few purposes:
- A revenue stream to support studios while they continue active development.
- A means to QA and balance a game with a smaller audience before it’s fully released to the general public (imagine everyone who pre-ordered a game getting full beta access.)
- A means of communicating with potential buyers that your game still is a work in progress, since there’s nothing stopping you from releasing an unfinished game through a normal Steam release.
There’s a few sticking points against early access that I’ve seen come up pretty often:
- Some bad actors have permanently damaged the reputation of early access by never fully releasing their games, or transitioning an unfinished product to “full release.”
- Early access titles can have their visibility penalized by Steam. They notably don’t get full launch visibility until they transition to full release. Users also have the ability to specifically filter our early access titles from what they’re shown on Steam.
- “Early access is bad for narrative games!” Most successful early access titles are multiplayer or sandbox games, genres that have an easier time bringing returning players back into the fold, and that have high replay value. For a narrative game with a significant enough amount of branching, this can also be the case, especially if the story is released over a period of time before release.
- “You only get one release.” This refers to both press coverage as well as general user hype. While this can be a helpful rule of thumb, I strongly disagree with this sentiment, but more on that later.
A couple last things to keep in mind:
- We’re still in early access, and will continue to be there for quite a while, so we have no idea what sort of bump we’ll eventually get when we do our full release.
- What’s worked for us won’t necessarily work for you. This post isn’t intended to argue that EA is a universally effective strategy. Rather, it’s intended to illustrate that EA isn’t a universally bad strategy, even when used for genres that the community at large wouldn’t expect to perform well in EA — don’t rule it out as an effective strategy without considering what makes for a successful EA release!
My wife Abby and I started working on our game in the spring of 2020, by complete coincidence right around the same time the pandemic hit. She’s been a full time web cartoonist and graphic novelist for over 10 years, and we thought it might be fun and interesting to make a game together.
In terms of pre-release marketing, we kept things almost entirely under wraps, wanting to capitalize on our announcement with an immediately playable first episode (which we released for free), and riding the hype from that into a Kickstarter launching the following week.
Part of this strategy honestly stemmed from not knowing a thing about the specifics of game marketing (my background was in building audiences for media companies) and by extension, not realizing the importance of Steam wishlists. That being said, the launch went well enough — in Episode 1’s first week, we had about 2,000 unique players, and our Kickstarter hit a little over 300% of its funding goal.
There were some major factors that contributed to that initial success:
- We have a large email list and a pre-existing audience we could market to from Abby’s years as a cartoonist.
- We put a lot of marketing energy into a scrappy campaign to get a bunch of press coverage and streamers, which wound up working out really well for us.
It’s possible that, had we announced earlier, these numbers would have been a lot stronger, just as much as it’s possible that they would have been weaker. If there’s an argument to be made here, though, it’s that wishlists aren’t the end-all-be-all of games marketing, and that focusing on more traditional levers can be an effective strategy on its own.
Following the end of Episode 1’s Kickstarter, we had a lot to consider when it came to the strategy for selling and releasing the rest of the episodes. Our original plan was to release them as paid DLC, with the option to buy them all bundled as a single season pass. There were a few issues with this approach:
- The complexity of the game’s narrative would have made selling individual episodes as DLC difficult. Steam doesn’t allow you to set prerequisites for purchasing DLC (i.e. you can’t require someone to own Episode 2 before purchasing Episode 3.) And Steam does require that purchased DLC be playable, so we would have had to spend extra development time picking an “average” world state as well as the ability to start at any chapter, which would be a poor user experience for our players.
- This left us with selling a ‘season pass’ as our clean option, which confused people.
- Releasing other episodes as DLC also would have made it a little more difficult to get increased visibility from participating in sale events, since the base game would be free.
- To address the above, we considered changing our page from free to paid (and charging a small amount, say, $1–3 for Episode 1.) This would have wiped our reviews, however, which were sitting at 98% positive at the time with a total count of around 250.
After a lot of hemming and hawing, we decided the best way to move forward would be to launch a new page for the full game, and use early access as our mechanism for releasing episodically. If you want a full deep dive of our rationale on that decision, we’ve detailed it in our devlog. With this pivot, our original free-to-play page effectively became a prologue, similar to what you might see from Tiny Bunny, Backbone, or Sir Brante.
This pivot let us focus on getting wishlists for a new page using everything we learned from our original launch, including thoroughly optimized marketing copy and screenshot selections. By the time Episode 2 was ready, we had an “okay” number of wishlists (in the mid four-figure range in the two months our page had been up.) While we could have delayed our release waiting for more, we decided it was more important to deliver new content to our players in a timely manner, so that’s what we launched with.
Our launch week went pretty well— we moved about 20% of our wishlists in new sales, and had a similar number of key redemptions from Kickstarter backers. Our outstanding wishlists also started to balloon. Within a week that number was already 50% higher than when we launched, and it continued to grow from there. These newer wishlists are likely skewed towards players who are interested in our game, but who are waiting for us to leave early access before purchasing, especially given the scope with which that number has grown.
This left some lingering questions, though. Our episode 2 launch was fine, but what would the next episode look like? After all, while episode 2 was our second launch, it was also the paid launch of our game, so it wasn’t clear if we would see similar spikes moving forward.
We got a hint of an answer in October of next year, when we released a large content patch for our first two episodes and ran our first discount outside of launch week. Because we’re going to be in early access for quite a while, we’re intentionally being quite conservative with our discounts. As a rule, we run 20% sales alongside major new content releases (mostly to remind our wishlisters that we exist, and to use updated capsule art to show them what progress we’ve made since they wishlisted us.) Since there isn’t as much value to email reminders going out during major sale events, we cut smaller discounts (10–15% depending on the event.) Email reminders during these events have less value because:
- People are already being emailed about items on their wishlists going on sale for major sale events. Contributing to that doesn’t really help us.
- Most folks shopping during those sales are going to check their wishlists anyways (and see a discount for Scarlet Hollow.)
The weeklong deal we ran in October did quite well for us — generating about 60% of our launch week in terms of revenue. We also saw a decent percentage of our players returning for new content and achievements — our all time peak concurrent players when Episode 2 released was 82, and the October content patch saw that number make its way back up to 22.
This was followed by pretty solid Halloween and Winter sale numbers, and a pretty meh fall sale (probably clustered too close to Halloween.)
We launched Episode 3 on March 7th with another 20% discount to hit up our wishlist, and so far the new episode has exceeded our expectations. Episode 3’s launch week was better than our October patch for sales, generating about 2/3 of our Episode 2 launch’s revenue. Our all time current player record also shot up considerably, reaching a peak of 199 concurrents on release day (242% of our previous all time high). A week and a half later, it’s still peaking at 40+ concurrents every day, which remains significantly higher than any period other than the first couple of days after Episode 2’s launch.
All in all, our experience with early access has already defied some of the pessimism surrounding it.
- We’ve already had two strong launch events for our game, three if you want to count Episode 1 (though that was a separate Steam page)
- People are still buying the game. We’re only 9 months in from launch and Steam sales are currently at 1.25x our launch wishlists, excluding key redemptions from copies we’ve sold outside of Steam. Including those copies, we’re at 1.66x our launch wishlists in sales.
- We’ve been able to use our new content to continually balloon our wishlist numbers, which will hopefully give us a huge bump when we eventually leave early access. Since entering early access, our outstanding wishlists have increased by 5.68x.
- Despite the fact that we’re not staircasing discounts below 20%, we’re still seeing those small sales move units without diminishing returns.
- It’s probably worth stressing that despite multiple launch bumps, we haven’t really gone viral yet — we’ve had one large youtuber (Gab Smolders) pick up the game, but the bulk of this growth is from community management, which means there’s still untapped potential for even bigger spikes down the road.
With sales and marketing out of the way, there are quite a few other ways early access has been indispensable to us.
- Free QA from players who are passionate about our product. Our game has an incomprehensible amount of branching in it — so much that it simply isn’t feasible to test everything at scale. Early access means that we’re able to engage a large, passionate core fanbase that’s happy to help catch bugs and inconsistencies and make our final product that much better.
- It holds us accountable — needing to deliver new chapters of our game to our supporters forces us to sit down and make difficult decisions on what to cut and how to best hit deadlines.
- It’s grown our community. Episode 2’s launch lifted our discord up from 300 to about 600 members. Since Episode 3’s launch, our discord has grown again from about 850 to about 1250 members. Our studio’s Twitter following has increased by a little over 10%, and our subreddit has grown about 25% in membership.
- It’s created a self-sustaining cycle of hype. We actually did a lot *less* marketing for Episode 3 than we did for Episode 1 or Episode 2 (limited bandwidth from being a 2 person team), but our community has done a huge amount of heavy lifting for us, from spreading the word to posting fanart.
- We’re able to respond to user feedback as we go. This means we’re able to do things like release balance patches to make given traits and choices feel more impactful, and we’re able to make small retcons or dialogue changes to change things that were missed or misunderstood by our players.
- It gives us leverage to make the game we want to make. Our early access sales aren’t making us wealthy, but combined with Abby’s comics, they’ve been enough to keep us afloat, which means we don’t need a publisher. And if we decided we wanted one, we’d have considerable leverage in negotiations.
So, all of this is to say that early access can work. Like so many other components of marketing, it’s a tool, and for some people, it’s going to be the right tool for the right job, much like some games are much better suited to TikTok or Twitter or Tumblr or Reddit than others. Much like for other folks, it’s going to be the wrong tool for the job. A unique challenge with early access though is that unlike other tools, it’s not one you can easily put back down, so it should be approached with caution. It just shouldn’t be dismissed out of hand.
In general, when you’re thinking about your marketing strategy, it’s important not just to look at what’s worked for other games, but to think about why a given strategy worked for those other games. Why did a post blow up on TikTok? Is your game capable of blowing up in the same way, or is it better to focus on another platform instead?
Here’s some of the reasons early access was a good idea for us:
- We’re not writing our game episodically for its own sake, we’re writing it because that’s the best format for the story we’re trying to tell.
- Scarlet Hollow is a mystery game that requires multiple playthroughs to get a full set of information from a given episode. This creates an ideal environment for our players to pool information and work together to solve things.
- At the same time, our episodes are written to give enough information that they can be fully enjoyed and mostly understood on a single playthrough.
- We had our entire game planned out before we started writing, so we’re able to avoid making a Lost-style mystery that doesn’t have real answers. I think this is part of why the mystery resonates with our players — they’re able to see that there is a conclusion on the horizon, and it makes them want to piece it together.
Anyway, I just wanted to share our experiences here so folks could get another perspective, since the discourse has felt so lopsided on this subject! I hope this is helpful in figuring out what strategy you want to use when releasing your own game.
If you’ve enjoyed what you’ve read here, consider supporting our studio on Patreon to get early access to all of our new devlogs! Likewise, if you want to dissect Scarlet Hollow and figure out for yourself why early access has worked for us, the game is only $20 on Steam!
Since we’re between episodes right now, I’ve got a few more blog posts in works — next up is a write-up of the GDC talk I gave this year about narrative design strategies to make the choices in your narrative game feel like they actually matter. Follow us either here on Medium, on Patreon, or on Twitter to be the first to know when that goes live!
