Dec 18, 2017
Why You Probably Shouldn’t Make the Hottest Character in Your Visual Novel a Secret
Inverness Nights — A Post-Mortem
Prologue
Let’s Talk Characters
The thing about visual novels, as a medium still oriented around romance games, is that established marketing strategy for them is likewise oriented around who you get to interact with. I wouldn’t have played Hustle Cat if I hadn’t known about Graves, I wouldn’t have played Dream Daddy if I hadn’t known about Damien, and I wouldn’t have played The Niflheim if I hadn’t known about Skeletiano — among many, many other examples that come to mind less readily. The big draw for a visual novel usually comes from the characters you present to your prospective audience and whether they appeal to them or not, and this tends to extend to VNs which aren’t romance-oriented — probably because at the end of the day, they’re a lot of talking, and if you’re not talking to someone you find interesting (visually or conceptually) you’re going to get bored.
A good example of non-romance games using character-lead advertising can be seen in Dangan Ronpa, a thriller VN series. Instead of asking you to care about its core murder-mystery premise — classmates in a locked school putting each other on trial — the series usually runs promotions based on its suspects-and-victims to-be, asking audiences to invest in an archetypical character and hope that character will make it through the deaths promised by the story’s genre. In my case, I waded into Dangan Ronpa to see whether cool, buff wrestler girl Sakura Oogami would survive, not to see teens pretending to be lawyers. This isn’t to say the rest of the premise doesn’t matter, just that Spike Chunsoft and other developers working with similar properties know that it’s easier to make people invest in a character than in a concept.
To that end, an important part of marketing visual novels is selling your characters, perhaps moreso than selling the rest of the game. In an ensemble cast, that tends to mean selling each character to a specific niche of your wider audience. Returning to Dream Daddy for a moment, we can describe the game’s wider audience as ‘people who are romantically interested in paternal men’, and some of the niches it targets as ‘people who like athletic men (Craig)’, ‘people who like artistic men (Mat)’, and — hi — ‘people who like gothic men (Damien)’. By selling on the characteristics those niches want or need to see to care about each character (a healthy exercise routine, a specific and critically-acclaimed taste in music, a sprawling manor home), Dream Daddy and similar games are able to gain a wide audience made up of several discrete types of player; seven particular tastes bringing in seven separate groups to make one whole audience.
When your VN has a focused narrative with a conventional protagonist, deuteragonist, so forth, rather than pitching your entire cast it’s common to use one or two of the major characters as a representative for each major section of the game’s story. For instance, the marketing for recent cult-success VN The House in Fata Morgana focused on the game’s deuteragonist, the Maid, a mysterious, cold, and wistful woman in a classic Victorian maid uniform, whose design and personality clearly conveyed to prospective players that the game was a tragic, historical romance which would have huge plot twists throughout the story (who is the Maid? Why is she sad? etc). Instead of appealing to several niche audiences with a character each, this approach requires having one or two characters who can speak to your broadest possible audience in sweeping, thematic terms; telling prospective players they’ll like the big picture of your game rather than the details of it.
The point of all this preamble is that, unless you’re running a particularly imaginative and wild promotional campaign, the critical part of marketing a VN is being able to point at your characters and say ‘these are what you’re here for’. Whether your players are going to kiss them, kill them, or go on a tearjerker quest to undo a curse on them, they are the main event, and the worst thing you can do is put yourself in a scenario where a key representative of the game can’t actually represent it. To that end, I made an exceptionally bad choice in creating my recent VN, Inverness Nights, which has haunted me through release and into the great beyond, which I’d like to discuss in this postmortem.
Reader, I made my best representative character a secret.
This is an analysis of that not-good decision, a few others, and maybe some good decisions too. Hopefully, it’ll help someone out there learn from my mistakes. Of course, this comes with a spoiler advisory for Inverness Nights in general — but honestly, please keep reading even if you haven’t played. There’s a chance you might be more interested by the end of the piece.
Chapter 1
How I Got Here
I’ve covered how the non-spoiler premise of Inverness Nights turned out the way it did in an interview, so to save some repetition, here’s that discussion for your perusal. The short version is that I wanted to make something about queer community and how difficult it can be to find and rely on your people when you’ve come to expect you’ll be alone forever. Within the advertised part of the story, that mindset is represented by the protagonist-apparent, Tristram Rose, an immortal gay man who could quite literally be alone forever if he doesn’t learn to form connections with people other than the men he dates.
But there’s a second part of the story which isn’t advertised, with a different narrator facing the same problems from another angle: Shahvir. Where Tris’ section of the story is about him breaking up with his boyfriend and trying to find an emotional outlet in the aftermath, Shahvir is a ghost who’s been isolated for hundreds of years after his boyfriend died. Unlike Tris, who struggles to find community because he sucks at open communication, Shahvir struggles to find community because he devalues himself and doesn’t think he’s worth anyone’s time. This is emphasized by the fact Shahvir possesses other people’s dead bodies to interact with the living world, removing him from it entirely. As a character who doesn’t physically appear until around ¾ of the way into the game, it’s hard to discuss him without spoiling who he’s pretending to be and why, even though he has almost as much screentime as Tris.
Which brings me to my problem: a character representative of a full half of Inverness Nights is a character who can’t be shown, because his whole point is that he’s not meant to be seen until late in the story.
In addition to him being a key representative character, he’s also the cutest character, at least in conventional terms. As a general rule Inverness Nights’ character designs were inspired by Renaissance Disney and Miyazaki movie designs, where characters are colourful, aesthetically pleasing, and might be inconsequentially attractive but are rarely designed with attractiveness as a priority. Shahvir was the exception to that rule within the cast, as he initially appears in the story as someone’s seemingly-insignificant eye-candy boyfriend. Being the token hot guy, he was designed to be handsome in a way the others weren’t — a variation on conventional prince designs, to be specific, hence the regal purple/gold. Since handsomeness/beauty/choose-your-sexy-poison is a critical part of pitching characters in most visual novels, we can mark that down as another aspect of the Shahvir-can’t-be-shown mistake.
All this left me in an awkward place with my marketing, because I could either opt to spoil the biggest twist in the game upfront and bring in people who would like the game for Shahvir, or I could keep my mouth shut and pretend Tris was the sole lead. I decided to opt for the latter, because I figured spoiling a story-based game would have people asking why they were playing it to begin with, and crossed my fingers that word would get out about Shahvir’s existence in lieu of the creator themself spoiling that he was there. Trouble is, for word to get out, people have to know you exist in the first place, and unable to sell on that key character I’d have to scramble for something else to pitch with.
Chapter 2
What I Did
Tristram Rose is not a likable protagonist, which means he was not an easy protagonist to market around. He’s interesting, but he’s not likable. Between his poor communication skills, his abusive dynamic with his ex-boyfriend Alasdair, and his domineering attitude, it’s hard to pitch him the way you might pitch a traditional VN protagonist, with reference to lots of choice or being relatable to the layperson. This lead me to pitch on what he does rather than who he is, a decision which helped with making the tagline of the game snappier at the expense of capturing the game’s tone.
For the most part, my elevator pitch for Inverness Nights was ‘break up with your boyfriend’, with additional elements tagged on depending on the setting — usually expanded to something like ‘you’re a gay man in 18th century Scotland, and you’re going to break up with your boyfriend’. This pitch has obvious hooks, but I think makes the game come off more caustic than it is; Inverness Nights proper is contemplative, intimate, and low-concept in a way that made it difficult to describe within game industry pitch norms. The trouble was that anything I tried that was evocative of the game as a whole, rather than what Tris does in it, was too meandering to work with. As soon as I broke out something explanatory, it felt like I needed to go further into justifying my choices for the cast and the setting and the subject matter. ‘An immortal gay man in 18th century Scotland tries to save a relationship he thought would last his lifetime’. Okay, but what do players do? Why Scotland? So on. (Though retroactively, I like the idea of testing this pitch on some platforms, because what it lacks in agency it makes up for in tone).
Moving on from the pitch, finding a visual way to market Inverness Nights was another challenge, due to a mix of budget and — again — the ‘Tris is a difficult character’ issue. Because Tris’ section of Inverness Nights is a break-up sim rather than a dating sim, so to speak, selling it with a full cast picture seemed misleading, which meant that either Tris or Tris and his ex Alas would have to be the focus of the promotional art. As I was unemployed during the second half of Inverness Nights’ development and art with more characters was more expensive, I opted for Tris only.
Both of the main promotional images I used (a piece with Tris by a river, and a piece with Tris in his tailoring workshop flanked by mannequins) were supposed to emphasize the melancholic and thoughtful atmosphere of the game, and on that level, I think they succeeded. A memorable anecdote;when I posted the demo with the Tris-and-mannequins art, someone linked me to a 4chan thread where people were speculating ‘gay or European?’ with his design, before a third commenter burst in with ‘this man is gay AND European’ — thanks, Legally Blonde. The downside of featuring Tris but none of the other characters is that he doesn’t fit cleanly into either of the usual MLM (men-loving-men, ie gay and bi men) VN visual trends. He’s not buff or heavy-set enough to fit the bara or bear aesthetic of MLM games for MLM people, but he’s also way too conventionally masculine to fit the femme or cute aesthetic of MLM games for women. I’d tried to go for a midpoint to make it clear it wasn’t meant to be a titillating title like the majority of MLM games, but in the process I’d muddied the water of who the audience was.
So with a pitch that conveyed the action of the game but not the style of it and promotional art that was true to the style in a way that didn’t really appeal to a particular audience, I set out to try and sell my game. Something I’d like to note at this stage — in case anyone’s wondering ‘why did you make these terrible decisions with a commercial product’ — is that when I began Inverness Nights, it was as a passion project, while I had a job, with no intent to sell it. Making it a commercial work was a decision that came later in development, and it was a decision made out of a mix of me wanting some minor form of income while unemployed, along with not wanting to devalue visual novels.
For those who aren’t in the VN community, there’s periodically discussion about how lengthy, free VNs such as Katawa Shoujo can set unreasonable expectations about how much work goes into a VN and what can be achieved in a VN with no or minimal budget. Because a ‘full’ VN is normally the lengthy of a regular novel (ie. 50–200k words) with additional art, sound, and programming work, it is a substantial undertaking, but the fact players might only see ¼ of that in an average run of the game — and thus spend maybe 1–3 hours with it — can give the impression it’s less effort than other genres and mediums. Consequently, there’s a general consensus that if you can charge for a VN, it’s a good idea to do so, to help give the genre greater value overall and encourage players to consider that they’re effectively buying illustrated novels which in any other format would be worth at least a few dollars a piece.
With this, and the marketing compromises I’d made in mind, let’s discuss how the release worked out.
Chapter 3
Let’s Sell a Game
Going through the actual process of release, the problem that I ran into — unsurprisingly — was figuring out what kind of audience to angle my work at.
During production the intent had been to sell to queer people in general, but despite the overlap in smaller social groups and in some general entertainment tastes, online queer-oriented communities tend to be split along gender lines. In addition, most MLM communities I looked at were only interested in having work by and for MLM shared in them; ditto for WLW (women loving women, ie. lesbian and bi women) communities. The exceptions to the pushback against creators from other demographics were cases where those creators were still clearly treating MLM as their primary audience, such as in Coming Out on Top, which features bara character designs. Being an NB person who has mostly dated other NB people, who made an MLM game as a distancing tactic from a failed real-life relationship, this meant that I couldn’t really pitch the game to those MLM spaces without coming off as opportunistic. Then there was that old issue of Tris’ design not having a clear audience to appeal to. I’m happy with his design, but because it was ‘Tris’ first and foremost instead of angling for broader ‘cool’ or ‘sexy’ or ‘mysterious’ appeal, I didn’t have a plan for leveraging it to attract people. Who was meant to be interested in Tris? People who like grumpy old-fashioned tailors, I guess, which is probably an audience (Phantom Thread sure is a movie) but not an easily accessed one.
If I had to assess the problems with Inverness Nights as a commercial visual novel — and recommend a strategy for people looking to avoid this situation — it’d be to consider how your characters will appeal and lean into it from the design phase, while making sure that the audience you’re appealing to are an accessible group. I don’t want this to sound like I’m advocating for every VN character being blandly handsome or cute or sexy — what I mean is that you should follow a visual design strategy that pairs with an audience (or perhaps several audiences across several characters) and adhere to it as carefully as you can. Some good examples in the queer VN space include We Know the Devil (which appeals to witchy queer women by having monochrome designs that pair quite plain costumes with alternative hair tropes like feathered+swept fringes and ombre dye) and Ladykiller in a Bind (which knows butch lesbians in suits will always be popular, and that more suits is better, so its protagonist The Beast wears five or so through the game). This is a pretty old-hat argument for most art director-types, I imagine, but it’s a point that doesn’t get drummed so hard when you’re someone from another discipline looking through game design concepts, which is why I wanted to present a case study that shows why it matters.
As for my title; why do I think that using Shahvir as the poster boy would’ve avoided the problems I had with Tris as the poster boy? Because the princely archetype — especially with a dash of tragedy thrown in — is, always has, and will ever be a popular archetype for people who’re sexually or romantically interested in men or masculine characters. Whatever variation you make on it, as long as you have a well-dressed, well-spoken character with fine features and some sense of chivalry, you can safely assume a lot of people will dream of kissing them. To that cynical end, telling people upfront that Inverness Nights is a game where you can see a sad, princely ghost man kiss people probably would have helped it reach a wider audience than telling them it’s a game where you can see a grumpy tailor ruin his love-life.
Before I wrap this up, though, I’ll toss in a caveat; while I’m sure changing my marketing to focus on Shahvir in addition to Tris would’ve helped bring a wider audience and thus, maybe more people discovering they like the game’s core themes, the strategy I chose gave me an audience who were interested in the challenging aspects of the story. If people came for sweet, beautiful Shahvir and got abusive, plain Tris, I imagine there would’ve been some unhappy players who felt they’d been suckered in by an uncomplicated premise and then blindsided by some very confronting themes. Asking people to be sad about a guy with a dead boyfriend is easy, asking people to engage with a guy who mistreats his boyfriend is hard. There’s also the fact that Shahvir doesn’t truly appear, as mentioned, until ¾ through the story. Anyone who starts playing wanting to see him without being aware which character he’s disguised as would probably get tired of waiting for the reveal. Reworking the story to avoid this would mean an almost complete overhaul, something that wasn’t feasible by the time I was worrying about the marketing (so, hey, when marketing professionals say ‘worry about marketing from day 1’… listen to them).
I think it’s worth remembering that the wider the net you cast with any kind of promotional work, the more likely you are to catch disappointed players, and if you’re working on a project like mine where maximising profit isn’t so important you should ask yourself whether a genuinely interested niche audience or a broader, less invested audience is what you’d prefer. If you can spare the money, consider that while recognition is cool, making a deep connection with a small handful of people can be a valuable experience as well, and it’s probably much less stressful.
Chapter 4
Stray Notes
The marketing side of Inverness Nights felt like the aspect with the most useful takeaway for other devs, so that’s where I wanted to focus the meat of this post-mortem, but I came by a lot of useful knowledge in production that’s worth discussing, so let’s bulletpoint through it.
THE GOOD
--While I was doing my editing I made a rule that every textbox should either reveal new information, be critical to a scene’s pace, or advance character development/plot — if a line didn’t serve any of these three purposes it got axed or folded in somewhere else. This gave the player a reason to keep advancing the text, and avoided the sense they were clicking mindlessly until a conversation went somewhere worthwhile. A lot of people have complimented this approach to editing and said it kept the story rolling where a lot of VNs are slowed by a padded wordcount, so it might be a useful editing technique for other writers.
--When I was doing my last-minute usage rights check to ensure I wasn’t taking advantage of any contributors, I discovered that the musician who’d done the game’s original soundtrack had publically said he didn’t want to be associated with VNs anymore. We hadn’t spoken in several years, so I panicked a bit and opted to remove his work for safety, then scrambled to add a royalty free Baroque soundtrack at the last minute. This move got the game some compliments for authenticity and a sense of time/place — there’s a bit of 18th c Scottish fiddle in there, primarily for Alas’ scenes — so I wanted to bring it up and remind developers that if you’re creative with your choices, royalty free soundtracks can still impress people! Don’t rule them out :D
--Likewise, the sound design in general got some notice due to the fact every narrated movement (except normal footsteps — that would’ve gotten tedious) had an accompanying noise, some of which were balanced in accordance to where the sprites were on-screen or filtered to help place them inside/outside/on certain terrain. I’d worried I’d overdone the soundscape because it has more in it than an average VN, but it got a positive response; in particular, I think, because it makes up for the limited action allowed by the small number of sprite poses.
--A lot of VNs have realistic art styles for their backgrounds regardless of the style the sprites are in, but because Inverness Nights had flat-colour sprites, I opted for a geometric, pastel style for the backgrounds to match (largely inspired by the background work in Steven Universe (colour-wise) and Scooby Doo Mystery Incorporated (shape-wise)). Though the execution might’ve been amateurish, I’m happy with how they paired with the sprites, and I think that going off the beaten track for BG art is something VN developers should consider more frequently.
THE BAD
-Outside of big productions which can afford to pay salaries or contracts for a team versus per-piece/per-hour commissions, there’s an attitude that it’s best to bring in artists once your first script draft is done. I’d actually argue that if you’re running a smaller project, it’s better to try one of the following approaches:
Separate artists for sprites, BGs, and illustrations
-First draft — sprites/BGs
-Second-to-last draft — illustrations
Single artist for sprites, BGs, and illustrations
-First draft — BGs
-Once the plot structure is locked — sprites
-Second-to-last draft — illustrations
When I started Inverness Nights, I wrote a first draft, then commissioned Alyssa (@leedalangin) for sprites and illustrations all at once to ensure they’d be done by the end of the game’s production. The trouble that arose is that first drafts are slimy, changeable things, and what was a 30k game with no ending or climax was soon revised into a 60k game with higher stakes, a larger cast, and consequently, more art assets.
At that point Alyssa’s art school work took precedence over commissions, which meant getting ahold of assets for the expanded game wasn’t always possible. The critical things — a sprite for Astor (the villain), and some additional outfits and arm poses for the other characters — got sorted, but for the remainder of the illustrations I had to mix between Alyssa and two other artists, Toasty (@emptyfeet) and Crista (@spoiledchestnut). Toasty and Crista both did a great job tweaking their styles to compliment Alyssa’s existing art for the game, but it is noticeable that the illustrations were done by multiple people, and the mix doesn’t look intentional like the multiple artist approach does in games like Dream Daddy.
Sprites having a different but complimentary style to illustrations is an accepted convention among VN fans, as it’s understood that asking an artist to draw all your character art and a few dozen illustrations is a lot of work for a single person. Different styles between illustrations is less accepted, unless it’s done with intent. To that end, I’d recommend setting up your art pipeline so that when you hire someone for illustrations you know exactly what you want and how much work they’ll be needed for, so you don’t need to unexpectedly switch between people like I did. Wait until you have your scenes locked in and you’re just tweaking prose and other minutiae to organise your illustrations, unless you’re working at a scope where that’ll delay release substantially. In that scenario, you could try a comics pipeline where you have separate artists for lines, colour, etc — some friends of mine have been trialling this in the VN space recently and it seems like a good solution for getting consistency without placing too much burden on a single team member.
--If you’re wanting to exhibit your work at shows, age ratings are an important thing to consider within your VN. A problem I ran into with Inverness Nights was that the themes and content were too mature for me to showcase it at our family-friendly local events, but larger international events with adult game showcases were angled for 18+ sex games. When building demos, trying to either adjust your work to be appropriate for all-ages or punch it up for erotic game audiences might be a good idea, so you don’t end up in the void between like I did; too mature for kids, too tame to be ‘adult’ in the genre sense. Of course, whether it’s worth including sex just to get the attention-grabbing (but platform-complicating) tag of ‘18+ game’ depends on whether you want to deal with the other hassles it adds. It’s just a pitfall I wanted to warn others of :)
--Solo devs, please invest money in a reliable video editing program when you’re able to, because a recurring problem I had once the game was done was being unable to afford the cost of putting together a trailer, which is a prerequisite for a lot of exhibitions and awards, as well as kind of important for marketing. It sucks not getting a shot at showcases because of a fairly basic equipment oversight, so I urge you to put it on your expenses and sort it out as soon as you can.
--Unless you’re the kind of person who loathes outlines, outlining for VNs is pretty important, and I recommend doing it thoroughly and thoughtfully before you touch your first draft. When I started Inverness Nights I wrote the story without branches and then tacked them on as I realised what needed expanding and who wasn’t getting attention, which was a nice, organic process but also a terrible way of putting together a game where finding new branches needed to be a clear, easy-to-follow process. What I ended up with were a lot of half-baked resolutions to different story threads, because I’d put them in to answer a particular question about a character or fill a particular hole in the backstory without caring how they’d resolve in a satisfying way. Reworking those threads into satisfying routes with a beginning/middle/end tripled the writing time for the game, and caused the illustration runaround discussed earlier.
Do your narrative design work: break out some charts, figure out what people want from selecting each option, whether they should get what they want, and have an ending to each branch that intentionally satisfies or denies that desire. Otherwise, you’ll have a lot of players hitting endings and asking the dreaded ‘so what?’ question.
Chapter 5
In Conclusion
Having spent three years of my life on Inverness Nights, I’m happy to say that I’m proud of the work I did on the game, and that it was a valuable learning experience on a lot of fronts — I’m better informed about hiring team members and planning my marketing than I was when I started in 2014, and I’ve gained a lot of skill in coding, audio editing, game design, art, and writing.
As far as return-on-investment goes, Inverness Nights cost me approx $4000USD out of pocket, and probably 3000–4000 hours of my own time; while I won’t quote the exact figures, I’ve made maybe a fifth of the monetary expenditure back. Some key points to remember are that I’ve never released or been credited on a game before, my social media following prior to now was mostly focused around voice-acting for a webcomic dub group, I’ve only released on itch.io, and I had a marketing budget of about $120USD with no media contacts or pre-release media features. While it would be amazing if more people played the game (if you tried it and liked it, please encourage others to give it a go!) I started off in pretty much the worst possible position for trying to sell a commercial project, so I’m not surprised that this is where I’m at with regards to sales. In early 2018 I have a month gap between jobs where I’ll be doing the paperwork for a Steam release. My hope is that I’ll gain a few new players there, but I’m not anticipating overnight success.
If I had to give a final verdict on Inverness Nights, it’d be that I’m happy or at peace with everything except how long I spent on it. A lot of people told me around a year into development that it wasn’t worth spending more than a couple of months on a first game, because a lot of mistakes get baked into the project and you’ll be stuck on it forever if you try to fix them as you realise they’re there. I’ve got the hindsight now to know that’s totally true, but I think most of the people who could use this advice are gonna be the people that don’t want to hear it. What I’ll say instead is that if you spend a very long time on a first project that no-one’s really invested in but you, yes, try and make something that you’re proud of — but remember that the reason you should be proud of it is that chances are, you’ll get it done and still be its most important audience.
Provided you can afford it there’s nothing wrong with a game that exists because you wanted to make it, but you have to remember that’s why it‘s here: for you. Find joy in that. Then use what you learned making it to decide whether you want to do it again.
On that note, I’ll be seeing you in 2018 when my second game, Captain Dracula, sets sail.
Happy holidays, all.
Inverness Nights can be purchased on itch.io for PC/Mac. From 12/18/17–12/31/17 it’s on 30% sale for $8.40USD, after which point it will return to its usual price of $12USD.
