I’d be lying through my teeth if I told you that my decision to re-write Nintendo’s Fire Emblem: Three Houses (FE3H) as a novel-length, golden route, romantasy action story was fully intentional.

Back in 2024, I hit a challenging roadblock in my work on The Xenthian Cycle and needed a break. Enter Into the Fishpond, my seven-month obsession.

Registered members can read Into the Fishpond on AO3. I’ve also tagged all my reflections on what writing a novel-length project on AO3 taught me.

If you don’t know the game, FE3H pits the heirs of three nations against each other, first as academic rivals and then as opposing generals.

You play as Byleth, a nonbinary mercenary turned reluctant professor with their own secrets to unravel. Whichever leader Byleth champions will conquer the other houses and impose their vision on the continent.

The character writing is excellent, and all the endings are heartbreaking in some way. The game design forces you to make hard choices across the board, which makes it compulsively replayable (e.g., I’ve finished three out of four routes.)

As I began work on Fishpond, I knew a few things:

• My adaptation would be a romance. Since our setting is a secondary fantasy world, it’s technically romantasy.
• The principal character is from Earth and enters the game world aware of the story’s cannon events. In fanfiction circles, this structure makes Fishpond an Isekai story.
• Seteth, the school’s administrator, is the male protagonist.
• Zara, my thirty-something lead, would body swap into Byleth’s unoccupied female body (the player chooses which version of Byleth to play when the game starts).

Beyond that, everything else was up for grabs. Or so I thought.

In Chapter 1, the Goddess Sothis wandered into my adaptation and handed down the terms for a classic golden route.

With a year of hindsight, it was a wonderful decision. Rewriting FE3H as a golden route added richness to my adaptation.

Okay, what is a golden route?

In video games, a golden route maximizes positive resolutions across the board. For example, the main characters all live, peace is restored to the land, terrible choices are avoided, secrets are safely shared, and everyone gets what they want (at least in part).

Unlike Triangle Strategy, another popular video game, FE3H has no golden route. No matter what you do, at least one of the house leaders ends up dead, often alongside key supporters who won’t join any of the opposing armies.

Adding a golden route to the romance plot raised the stakes for all the characters and dramatically changed storylines for the house leaders who play supporting roles in my version of the story (e.g., Claude, Edelgard and Dimitri).

Writing a Golden Route also meant I could:

• Adapt the story’s overall ending.
• Avoid the five-year time skip written into the game’s original story, keeping us in the more idyllic, academic setting that all the routes share in the first half of play (“White Clouds”).
• Write new battles with the story’s best antagonists in ways the game doesn’t allow.
• Change all the endings for the major house lords, supporting players, and non-combatants.

Didn’t that torpedo the story’s conflict?

In short: no.

FE3H has layers of misunderstanding and betrayal baked into the narrative. Despite the wide-ranging success conditions, my adaptation had plenty of tension.

For example:

• Lady Rhea, the archbishop of Fódlan, is a character with deep secrets, a flawed psyche, and a decidedly amoral approach to statecraft. She remained a morally gray character in my work.
• The Agarthans, the true villains of FE3H, rarely get to fight the primary heroes (the game’s Verdant Wind route where you side with Claude is the exception). My adaptation sets them directly in opposition to the united school in ways that were new to readers. I used the game’s existing Agarthan spies and turncoats to foster suspicion and fear.
• Macuil, one of Seteth’s mysterious brothers, plays a minimal, off-screen role in the canonical story. Bringing him into the main plot and putting him at odds with Seteth allowed me to wrap lore drops in provocation. Macuil also let me build a unique supporting character since he’s absent for so much of the canonical experience.

Given the fraught, interconnected character histories, Zara’s and Byleth’s efforts to unite the house leaders still required care and discretion. They’re always on the edge of being found out, often in explosive ways.

What I learned

Fishpond follows the canonical story to a point. Reshuffling the possibilities around an outcome that readers had never seen in the game added extra intrigue.

After spending so much time on this thought experiment, I understand why FE3H doesn’t include a golden route:

• The extra scenes, maps and battles would mean a lot of complexity and development for a game that was already four months late when it was published in 2019.
• Fans complain that some routes aren’t as well developed as others (consensus seems to be that Azure Moon is strongest and Silver Snow weakest, though the debate continues).
• Adding the extra scenes required to support a new route would also mean more character dialogue. FE3H already has four routes + 1 partial DLC route, along with a full and excellent cast, so that’s a big additional cost.

Thanks to the comments, I know that some readers gave Fishpond a shot because it was tilting at this particular windmill, not the romance premise (Seteth is not as popular as the students for obvious reasons).

Conversely, writing an original female character, even one who spends much of the story wearing a body that is part of the game, meant some fans wouldn’t read my story. Some readers only want original pairings; others dislike the Isekai structure.

Either way, writing a golden route gave me another hook to Fishpond’s pitch. It’s an impulse that I’m glad that I embraced.

Other posts in this series:

• My AO3 stats model
• Overturning my Internet Bullshit: That time I wrote romantasy action on AO3

The post What Writing on AO3 Taught Me: Golden route = Powerful premise appeared first on E.M. Williams - Fantasy Author.

While writing Into the Fishpond, a novel-length fan fiction story set in the “Fire Emblem: Three Houses” universe (FE3H), I built a growth model for tracking my readers’ engagement.

The story is long: forty-seven chapters, and over 200,000 words.

When I finished the initial story in August 2024, I copied my model and populated it with a fake/sample data set for other writers with a growth marketing mindset to use.

This post explains the model and my thinking, and contains a link to the Google Sheet for those who wish to apply it to their own projects.

Why I built a growth marketing AO3 stats model

In my day job, I’ve been the head of marketing for several tech companies and worked with many talented growth marketers. If you haven’t heard the term, it’s the field of marketing most concerned with how you grow and sustain a customer base.

Or, in this case, a group of engaged readers.

On whim, I started writing Fishpond on February 9, 2024. I didn’t know what I was doing or why, but I decided to post chapters beginning on February 22.

As a writer, I was completely new to AO3. I didn’t have any presence or following when I started. I had read some FE3H work by other writers dealing with this particular character pairing, but only as a guest. In the beginning, it felt like I was publishing the story for myself. I was skeptical that it would find readers. 

Case in point about how little I knew — I didn’t discover the “Statistics” tab in the home screen until after I had published Chapter 14.

By mid-April, Fishpond had reached 1,000 hits. I told a growth marketing colleague what I was doing. 

My colleague: “What’re you doing for Analytics?” 

Me: “Not much. All the site data is in the moment. There’s no time-based record of the metrics at all.”

My colleague: “How about we fix that?”

We built out a Google Sheet to track what was happening with the story. My colleague also encouraged me to adopt a regular posting schedule given my steady output. 

My use case and assumptions

I started applying my data model when I published Chapter 17 of Fishpond on April 26, 2024. Thanks to back-and-forth texts with my colleague, I had some historical data from screenshots dating from my first weeks of publishing.

My numbers were and are limited to what appears on AO3’s summary page:

Beginning with Chapter 17, I posted twice a week on Monday and Thursday between 4 and 5:30 pm EST. That pattern continued until I reached Chapter 38. After that, my writing process got complicated. I needed more time between instalments.

By July, I’d switched to a weekly model (Thursdays) with Chapter 39. At this posting tempo, I saw significant growth between with each chapter. I’d now recommend going weekly or monthly, although it requires more patience.

True experiments, where marketers control everything except one variable to determine which option wins, are harder to do with fiction. The stories themselves vary. Knowing whether audiences are responding to the frequency or the content would be next-to-impossible to parse out (unless AO3 itself did a study).

As you’ll see, my AO3 Stats Model has two primary tabs: 

• General Growth by chapter, which tracks the baseline AO3 stats — I updated this sheet once a week in the 15 minutes before the next chapter went live. It contains a trend line that changes trajectory as you feed it more data, and a table for tracking overall audience engagement by chapter (kudos, comments, bookmarks and subscriptions). 
• Chapter Growth by key metrics, which tracks activity within the first 24 hours and until the next chapter is published — I track hits, kudos and subscriptions (capped 15 minutes before publication). 

You can’t beat email subscriptions for reach

Growth marketing focuses on actions that reduce friction between user and product. That’s why brands ask you to subscribe to their marketing newsletters when you buy something or to provide an email address before you see a demonstration.

Growth marketers love email because it partially solves the tug of war between social platforms building for self-serving needs and content creators trying to gain an audience. It’s a medium we control end-to-end. 

AO3 is more like a library archive (with a beautiful tagging system) than it is like a social platform. It doesn’t have or need an algorithm with filtration power like you’d find on sites run by Meta or TikTok.

The subscription feature that AO3 provides is similar to email, but not the same. I can’t see the names addresses of anyone who subscribes; AO3 handles the actual emailing of updates I make.

But in principal, it’s close.  

What I tracked

Hits are a key AO3 metric because they’re the way to know how many visits a work has had. Long-time AO3 users know the site counts a hit as one visit from an IP address within a 24-hour period. Multiple hits are not recorded, nor are hits from writers to their own work when logged in.

You can read more about how it works in the AO3 FAQ.

Kudos, the site’s equivalent to likes, indicate a work’s popularity and act as social proof that a story is engaging. Some writers particularly value them from registered accounts. Most platforms have a like/heart/Kudos function. I’m grateful to every single person who gave me one, whether from a registered user or a guest account.

Subscriptions, either to a work or to everything an AO3 writer publishes, mean readers get emailed every time you update. That makes a subscription the single most valuable metric to a growth marketer over time.

Why? Because the engagement action is a single click from your email inbox. As my subscriber count grew, so did the hit and kudos count week-to-week. I’m grateful to every reader who subscribed to the story or, eventually, to me.

My model only includes work subscriptions and not user subscriptions. When I started, I didn’t have any user subscriptions. You could adapt the model to track them separately or together as you see fit.

I can tell you that as Fishpond got more popular, I received double the subscriptions in the period 24 hours after a chapter published compared to the first 24 hours.

Why aren’t bookmarks tracked in Chapter Growth?

When a reader sets a bookmark, they still have to actively remember to check it to read the next piece of story, and some readers have dozens, if not hundreds or thousands, of bookmarks.

Readers can choose to make bookmarks private. That means the public-facing bookmark count on a story’s summary doesn’t give the whole picture.

I did love seeing my bookmark number rise; I tracked it in the General Growth tab.  I also really love it when someone annotates their bookmark on my story — I can sometimes tell from the description how far they’ve read or which kind of reader they think story would appeal to. 

But as growth stats go, it’s not as active or transparent an indicator of interest as a subscription.

Why aren’t comments tracked in Chapter Growth?

Comments are great. I love them.

From a growth perspective, their numbers are misleading, especially for works like mine. I responded to all my comments. That’s why I track them in the General Growth tab.

I respond to all of my comments because I enjoyed the interaction with the audience. They were a great source of qualitative feedback about whether the audience agreed with my instincts for the characters and plot — but I don’t count my comment responses in the model.

Given the number of people who never ever comment but will happily give a kudos, my colleague and I opted to give the kudos metric more weight.

For what it’s worth, I don’t think AO3 should count responses from creators on their own stories, either. Writers know we’re engaged in what’s happening with our work. I’d exclude them from the Dashboard counter.

Your mileage may vary and you can adapt your copy of the sheet how you like. Make it work for you!

If I could change one thing about AO3 stats  . . . 

I’d want to know how many users hit all the chapters in a work. Or, I’d ask for a ‘percentage read’ number or ‘completion rate’. Given how the Hit counter works, the number I see could easily be many, many people reading the first chapter and then stopping.

That’s why comments or annotated bookmarks are so nice when they crop up. It’s the only way to get a sense of how far people are reading. But I don’t believe they’re an accurate measure.

Growth marketers can be a bit obsessive

In general, growth marketers are data-driven people. We can be a bit (a lot?) obsessive about things we track.

Getting accurate numbers requires manual updates a couple times a week. I’m comfortable with that, but you may want to change it up. 

If the sheet is too much, feel free to cut back for your copy, adapt it, or ignore it. Track whatever feels right for you.

My goal is to give you a tool, not an obligation.

Data trends take time to develop on any platform

If you’re new to any Internet platform, it takes a while for people interested in your content to find your work.

I saw a big inflection point for Fishpond in Chapter 16 and again in Chapter 27. When I moved to the weekly model, I again saw a big jump with Chapter 39. 

My data and publishing frequency prior to Chapter 15 were erratic. If I’d decided then to pass judgement on whether the story was a success or failure, whatever answer I gave myself about what was happening would probably have been wrong.

It was simply too soon to know. 

Be patient with your work. It takes time to find an audience and for people to get invested.

That’s the lesson I’m taking forward with The Xenthian Cycle.

Other posts in this series:

• What AO3 Taught Me: Golden Route = Powerful Premise
• Overturning my Internet Bullshit: That time I wrote romantasy action on AO3

The post My AO3 Stats Model appeared first on E.M. Williams - Fantasy Author.

Live long enough, and you may overturn your past self’s Internet bullshit.

I’m a case in point.

Me circa 2015: “Fifty Shades of Grey is the hottest thing in publishing and I have formed negative opinions about fan fiction.”
Universe (grinning): “Hold my beer?”

Unpacking bullshit can be a long process

I’ve spent a lot of my middle adulthood digging out of elitism as a life choice. I’m also old enough that parts of that journey are visible in my Internet long-tail.

I grew up excelling in academics, and liked both school and learning. I won scholarships, awards, and medals. My family valued my achievements.

As I approached the end of my undergraduate degree, I had no idea what would make a viable career. What I wanted to do was write books. I didn’t believe that career would pay for my life, and neither did my family. At a loss, I leaned into earning a PhD in Science Fiction & Fantasy.

That way, I’d meet the writers I admired and study the books that I loved. Win-win, right?

Diligently, I earned a Master’s degree in Literary Theory. At some point on that journey, I realized I was more interested in writing books of my own than in talking about other people’s work for a living.

I started working, first in communications and then in marketing, and writing on the side.

Like many creatives, my ambitions and opinions about what makes a great book exceeded my ability to execute for a long, long time. It was frustrating. I despaired and felt envy for people I perceived as having what I wanted.

During this time, I wrote some unfortunate, judgemental things about fan fiction. The original essay isn’t worth your time.

My karmic turn in the river

In 2016, I read Liz Gilbert’s Big Magic. In it, she brings a non-elitist approach to creativity’s importance in our lives, while pointing out the elitism that exists in how we’re taught to think about how art and how it gets made.

During this period, I was also active on Twitter, which expanded my creative horizons. I listened to a lot of conversations about how widely people read, what participating in a fandom can mean, and why folks like what they like.

Reflecting on both experiences and my first essay after finishing Gilbert’s book, I realized I was doing nothing more profound than letting my fears about my writing prospects grab the mic while crapping on other people.

So I wrote “Death of a Fan Fiction Snob” to take myself to task.

For years, I thought that was it—lesson learned.

The muse had other plans.

The Switch game that changed everything

In 2019 and 2021, I played Fire Emblem: Three Houses (FE3H) for the Nintendo Switch.

It’s a strategy game set in a fantasy world with great characters, some heartbreaking choices, and a high degree of re-playability. I was quickly hooked. As I played the first of four possible routes, my passion for the game prompted me to read fan fiction for the first time.

Last February, I started rewriting the game as a romantasy action story on an Archive of Our Own, one of the biggest non-profit sites for fan fiction.

It’s called Into the Fishpond: A Three Houses Alliance. And it’s finished.

While Fishpond has a small yet kind following (~7,000+ hits), I’ve had some wonderful interactions with its readers (~170+ comments). Since it’s over 215,000 words, Fishpond is the longest story I’ve ever written. It taught me lots about writing faster, structuring scenes, and building sexual tension.

Like Lewis Carroll’s Through the Looking Glass, which loosely inspired the title, Fishpond is a portal story about someone from our world who ends up in the portal world.

It’s also got:

• A romance between two people dealing with grief (if you know FE3H well, Fishpond is loosely a Setleth ship)
• Found family
• Dragons
• Intense battle scenes (surprise!!)
• Magical twins (again!?) whose sibling bond low-key powers the whole thing (picturing your shocked face)
• My first non-binary character
• A lot of non-canon lore that I made up as I went along

Want to know more? Here’s the plot teaser from Fishpond’s intro on AO3:

Zara has just finished playing Fire Emblem: Three Houses for the fourth time when she’s struck by the injustice of watching her favourite characters die time and again. Why can’t she save all of them?

Fortunately, the Goddess Sothis agrees that this is a problem worth solving.

Into the Fishpond by E. M. Williams

Getting over your Internet Bullshit

My Internet bullshit almost stopped me from writing Fishpond. I was embarrassed to have once held negative opinions about an activity to which I was now devoting so much time. At first, I spent a lot of time minimizing this new novel and what it meant to me.

Some kind friends pointed out how deeply sad I got writing Chaos Armor, and how much joy came into my face when I talked about Fishpond. I’m glad I listened to them.

Besides great joy, this creative side quest proved I can build an audience entirely separate from my social circles and professional network.

I don’t expect the crossover audience for Fishpond to be high. Without game context, I’m not sure how much sense the story makes to Xenthian Cycle readers, although I did get comments from some people who’ve never played the game and enjoyed it.

Either way, I’m proud of the work and wanted to share that it exists.

Here’s to living long enough to set your Internet bullshit on fire.

Other posts in this series:

• What Writing on AO3 Taught Me: Golden Route = Powerful Premise
• My AO3 stats model

The post Overturning my Internet bullshit appeared first on E.M. Williams - Fantasy Author.