Appified Games Playbook  ·  Tapugo Internal
Tapugo · Internal Strategy Document
The Appified Games Playbook
A complete product framework covering category definition, the 5 core pillars, the ARM funnel, live ops strategy, performance benchmarks, and the four user personas that define this new category.
38 chapters 6 parts Footballer Life Simulator · PMF confirmed

Definition: An appified game is an interactive product that uses game mechanics — progression, simulation, consequence — as the engine for delivering an aspirational identity experience, designed around high-frequency habitual engagement rather than entertainment sessions. The user does not play a character. They inhabit one.

Part I · Framework
Category definition, competitive positioning, the 5 non-negotiable pillars, and the distinction table that separates appified games from mobile games and lifestyle apps.
Part II · ARM Funnel
Acquisition creative strategy, retention ritual architecture, monetization tiers, live ops calendar, and the benchmark health dashboard with category-specific KPIs.
Part III · Personas
Four psychographic archetypes — The Dreamer, Obsessive, Escapist, and Insider — with full ARM implications, churn triggers, and monetization opportunities per persona.
Part IV · Narrative Design
The content architecture system — beats, voice specs, state variables, decision trees, consequence chains, career arcs, calendar integration, AI production pipeline, and 20 writing rules.
Part V · Onboarding
The first 10 minutes — five phases from Drop-In to Open Thread. Identity commitment design, first beat architecture, D1 notification strategy, sport variants, and onboarding success metrics.
Part VI · Content Scaling
How 2 people scale narrative without it feeling thin — the multiplier system, calendar as content engine, community intelligence, weekly pipeline, AI prompt templates, authenticity audit, and depth runway math.
Cover · Start here
Part I · Framework
01
Part I · Framework
Category Overview
How appified games differ from mobile games and lifestyle apps — and why that distinction changes every product decision downstream.
DimensionMobile GameLifestyle AppAppified Game
Core pullChallenge / entertainmentUtility / informationFantasy / identity
Session designLong, intentionalQuick, functionalShort, emotional, repeated
Retention driverProgression loopsHabit / dependencyAspiration maintenance
Failure stateGame over / losingN/ANarrative stagnation
Monetization modelIAP / adsSubscriptionSubscription + belonging IAP
Success metricSession length / DAUMAU / churn rateDaily ritual completion
Community roleOptional featurePeripheralCore product layer
Identity simulation as a product category
People don't want to play a footballer. They want to feel like one. The game mechanics — decisions, upgrades, stats — are the delivery mechanism for that emotional experience. This is why the ASMR comparison holds: ASMR isn't entertainment in the traditional sense, it's a sensation delivery product. Appified games are aspiration delivery products.
Why the retention numbers look like an app, not a game
D1 retention of 55–62% and 10× daily opens aren't game metrics. They're habit metrics — the behavioral signature of something users feel they need to check, not something they choose to play. The product has successfully crossed the line from entertainment into lifestyle infrastructure.
01 of 23
Part I · Framework
02
Part I · Framework
The 5 Core Pillars
Structural requirements that must exist before the ARM funnel can function. An appified game without all five collapses into either a mediocre idle game or a forgettable lifestyle app.
1
Pillar One
Aspirational Identity Specificity
The fantasy must be precise, not generic. "Be a footballer" is too broad. "Be a 22-year-old attacking midfielder at a French Ligue 1 club trying to earn a national team call-up" creates the conditions for genuine immersion. Specificity is what makes the user feel seen.
Failure signal: Users describe it as "a game" not "my footballer"
2
Pillar Two
No-Lose Decision Architecture
Every decision must carry consequence without failure. Choosing the wrong training costs form. Declining a transfer costs money but gains loyalty. There is no game over — only narrative direction. Anxiety is the enemy of habit.
Failure signal: D1 retention drops below 45%
3
Pillar Three
Real-World Calendar Sync
The product's narrative must pulse in sync with actual world events: match days, transfer windows, cup draws, international breaks. The world is the content engine. A footballer app that doesn't register the Champions League final as a meaningful moment has failed this pillar.
Failure signal: DAU flat during major real-world sport events
4
Pillar Four
Witnessed Social Status
The user's progress must be visible to an audience. Discord communities, shareable career summaries, narrative stat leaderboards — these transform personal achievement into social identity, dramatically deepening retention past D30.
Failure signal: Discord below 500 active members at 10K DAU
5
Pillar Five
Authentic Narrative Voice
The writing, scenarios, and consequences must feel true to the actual lived experience of that identity. This is Tapugo's deepest moat: the CEO's footballer background means he knows what an agent conversation actually sounds like. This authenticity cannot be faked or outsourced.
Failure signal: Reviews mention "feels fake" or "doesn't feel real"

The pillar test: Before greenlighting a new title, score each pillar 1–5. Any pillar below 3 is a product risk. Any pillar below 2 is a launch blocker. The pillars are a coherent system — weakness in one undermines all others.

02 of 23
Part II · ARM Funnel
03
Part II · ARM Funnel · A
Acquisition — Selling the Fantasy
You are not acquiring mobile game players. You are recruiting people who have always wanted to live a specific life. Every creative, channel, and targeting decision must reflect that distinction completely.
"The creative doesn't sell the game. It sells the feeling of being the player."
POV-first creatives
No gameplay footage in the first 2 seconds. Open on a first-person moment: reading a transfer offer, walking into a stadium. The hook must be emotional before it is mechanical.
Identity statement ads
"You've always wanted to know what it's like." Copy that validates the fantasy without making it feel like a game. Target audiences who follow athletes, not people who play sports games.
Authenticity signals
Real terminology, real scenarios. A transfer offer that uses actual contract language. A physio report with real injury terminology. Credibility converts.
UGC amplification
Discord users sharing career milestones are your best ads. Build shareable career cards from day one. User-generated "I just signed for Real Madrid" posts outperform any studio creative on ROAS.
TikTok / Reels — primary
Fan culture lives here. Target by interest graph: football fans, athlete vlogs, transfer news. POV format is native. CPI efficiency highest in France.
Meta — scale engine
Lookalike audiences built from D7 retained users, not installers. The retained user profile is the acquisition target — not the curious clicker.
YouTube pre-roll
Target sports documentary and athlete vlog viewers. 15-second unskippable POV creative. The audience is already in aspirational consumption mode.
App Store optimization
Category: Sports, not Games. Screenshots show narrative moments, not UI. "Live the life" as the primary value proposition. Reviews mentioning immersion are the social proof to surface.
Audience: fan, not gamer
Target people who follow athletes, not people who play sports games. The fan has the fantasy already formed. The gamer may not have the emotional investment in the identity.
Geo: depth over breadth
France is the proof-of-concept market. UK, CA, AUS follow a similar cultural-intensity profile. Scale within these before expanding. CPI efficiency degrades in low fan-intensity markets.
Event-driven spend spikes
Double UA spend 72 hours before major real-world events: transfer deadline day, tournament draws, World Cup opening. Intent is at its highest. CPI lower, LTV higher.
Target CPI: $0.40–$0.80 T1
Primary: TikTok + Meta
ASO: Sports category
Creative: POV first-person
Spike: ±72h major events
03 of 23
Part II · ARM Funnel
04
Part II · ARM Funnel · R
Retention — The Ritual Architecture
Retention in appified games is not about keeping users inside the app. It is about making the app feel like a necessary part of the user's day — like checking the news, tracking a portfolio, or maintaining a relationship.
"A retained user doesn't think 'I should play today.' They think 'I wonder what happened overnight.'"
☀ Morning trigger 06:00–09:00
Push notification as news, not reminder. Agent, press, physio. Land on narrative inbox — not a menu. Story first.
⚡ Decision moment
30–90 second choice. No fail state. Consequence felt in narrative direction, not stats lost.
📰 Midday world update
Real-world event surfaces in narrative. Transfer news becomes your transfer rumor. The world feeds the product.
🏟 Evening match / event
Simulated match result tied to real fixture. Longer engagement window. Discord most active here.
💬 Night — social witness
Discord, shareable career card. Status visible to community. Loop seeds tomorrow's morning trigger.
The gap problem
D7 at 20–25% dropping to D14 at 15% is 30–40% churn in one week. The "novelty cliff" — initial fantasy explored, deeper ritual habit not yet formed.
The bridge mechanic
At D7, introduce the first major career inflection: a transfer offer, contract renewal, or squad selection. A "second first impression" that resets emotional investment.
The community handoff
D7 is when Discord integration deepens. Surface career stats publicly, invite to community events. Social belonging becomes the retention layer that pure narrative cannot sustain alone.
D1 target: ≥ 55%
D7 target: ≥ 20%
D14 target: ≥ 18%
Daily opens: ≥ 6×
Below 3× opens: not appified
04 of 23
Part II · ARM Funnel
05
Part II · ARM Funnel · M
Monetization — Belonging Over Transactions
Appified games monetize identity, not entertainment. Every mechanic must feel like an extension of the fantasy — something a real person in that position would pay for.
"If the user ever thinks 'I'm being asked to pay for a game feature,' the monetization has failed."
Tier 1 — Subscription (core)
Monthly access framed as "your career, uninterrupted." Anchored to streaming service pricing ($4.99–$7.99/month). Target: 25–35% of active users convert within 30 days.
Tier 2 — Belonging IAP
Identity-expressive purchases: boot customization, agent upgrades, premium career stats. Must feel like lifestyle choices, not game boosts. Never make them feel like a paying player.
Tier 3 — Season Pass
Event-driven premium track: World Cup Edition, Transfer Window Pass. Drives LTV spikes around cultural relevance. Time-bound creates urgency without being predatory.
Tier 4 — Portfolio Pass
When 3+ titles exist: cross-sport bundle ($9.99–$14.99/month). "All sports. One life." The LTV multiplier that shifts the VC conversation from studio to platform.
Never gate narrative progression
If a user can't experience the next story beat because they haven't paid, they disengage permanently. The narrative must always progress. Monetization adds depth, never access.
Never offer stat boosts or pay-to-win
The user's career must feel earned. A paid stat boost breaks narrative authenticity — the player knows they cheated the fantasy. This destroys the very thing the product is selling.
Never show ads inside the experience
A banner ad inside the agent's message inbox shatters immersion instantly. Ad revenue is incompatible with identity immersion. Choose subscription and accept the constraint.
Never make IAP feel like a game store
"Your agent has arranged a meeting with Adidas" is a monetization moment. "Visit the Store" button is not. The fantasy must hold through the transaction.
Conversion target: 25–35%
Sub price: $4.99–$7.99/mo
D30 ROAS: ≥ 100%
Ads: never
05 of 23
Part II · ARM Funnel
06
Part II · ARM Funnel
Live Ops — A Living World
Live ops in appified games is not a content calendar. It is a world management system. The product must feel like it exists in real time and responds to the world exactly as real sporting careers do.
"You're not shipping content updates. You're running a season."
Always on
Daily narrative triggers
Weekly form updates
Discord events
Match week
Pre-match build-up
Match day simulation
Post-match press
Fan reaction feed
Jan / June
Transfer window season
Transfer window pass IAP
Deadline day event
Intl break
National team selection
International arc
Country leaderboards
World Cup
World Cup Edition pass
Tournament arc
Daily elimination events
10× organic UA spike
🌍

World integration

Real events feed the narrative engine

Real fixture calendar mapped to match simulations. User's team plays when their real counterpart plays.
Transfer window mirrors real window. Rumors, deadline day tension, post-window arc.
Real world news injected as narrative. A high-profile injury surfaces as a career context moment.
📖

Narrative seasons

Story arcs that last 6–10 weeks

Each season has a career theme: Breakthrough, Injury comeback, Contract year, National team push.
Season arc creates anticipation. Users know something significant is building. Retention holds.
Season finale creates a shareable moment. Career summary card generated. Discord erupts.
💰

Economy events

Monetization woven into the world

Sponsor deals as IAP narrative moments. "Adidas wants to upgrade your deal" — a purchase framed as a milestone.
Limited edition season passes. World Cup pass. Only available for the duration — no evergreen equivalent.
Gifting mechanics between Discord users. Belonging IAP becomes social currency.
👥

Community ops

Discord as a product feature

Weekly "Career of the Week" spotlight. Community-voted. Status is the prize — not money.
Community-driven narrative votes. Users shape the world. Ownership deepens retention.
Community co-designed events. Discord runs itself during these — ops cost near zero.
06 of 23
Part II · ARM Funnel
07
Part II · ARM Funnel
Benchmarks & Health Signals
Category-specific KPIs. These differ significantly from standard mobile game benchmarks because the product category is different. Use these to diagnose whether an appified game is functioning as designed.
Metric
Appified game target
Casual game baseline
D1 retention
≥ 55%
Below 45% = pillar failure
25–35%
D7 retention
≥ 20%
Ritual forming threshold
8–12%
Daily sessions / user
6–12×
Below 3× = not appified
1–2×
CPI (Tier 1)
$0.40–$0.80
$1.50–$3.00
D30 ROAS
≥ 100%
Footballer at 120%
40–70%
Sub conversion
25–35%
Of D14+ retained users
2–5%
App Store rating
≥ 4.5★
3.8–4.2★
Discord / DAU ratio
≥ 5%
< 0.5%
Signal 1 — Session count dropping
If average daily sessions per active user drops below 4, the product is transitioning from appified game to regular game. Audit push notification performance and morning trigger copy first.
Signal 2 — Discord going quiet
Discord activity is a leading indicator for D30 retention. When community drops, D30 churn follows within 2 weeks. Run a community event immediately.
Signal 3 — Reviews mentioning "repetitive"
The narrative layer has been exhausted. The response is a narrative season launch, not a feature update. The problem is content depth, not product mechanics.
07 of 23
Part III · User Personas
III
Part III · User Personas
Four Personas. One Category.
Segmented not by demographics, but by what the fantasy is doing for them psychologically — the variable that actually explains behavioral differences in acquisition, retention, and monetization.
1
Persona 01
The Dreamer
"I always wondered what my life would have been like."
42%
User base
High
Sub rate
Daily opens
"The guy who played football as a kid and still thinks about what might have been."
2
Persona 02
The Obsessive
"I know more about this sport than most professionals."
28%
User base
V.High
Sub rate
14×
Daily opens
"The super-fan. Already consuming sports content 3 hours a day."
3
Persona 03
The Escapist
"I just want to feel something different for a few minutes."
20%
User base
Med
Sub rate
Daily opens
"Uses the product as emotional regulation. The ASMR connection is strongest here."
4
Persona 04
The Insider
"I want to understand how the sport actually works from the inside."
10%
User base
V.High
Sub rate
11×
Daily opens
"Small but high-value. Loves realism, hates anything that feels gamified."
You are designing for The Dreamer but you will be discovered by The Obsessive
The Dreamer (42%, highest emotional investment) is your core product target. But The Obsessive (28%) is your organic growth engine — they share, post, and bring their friends. The product must be deep enough to satisfy the Obsessive without being complex enough to alienate the Dreamer. That tension is the central design challenge of the category.
08 of 23
1
Persona 01 · 42% of user base
The Dreamer
"I always wanted to know what it would have been like. I had the talent but not the path. Or maybe I just needed to believe that."
Daily opens
High
Sub rate
Low
Churn risk
Med
IAP spend
"I played until I was 19. Then real life happened. This isn't a game to me. It's the life I didn't get to live."
Discord message · Footballer Life Simulator community
Who they are
Former youth or amateur player who left the sport before it became professional
Age 25–40, predominantly male, heavy football media consumer
The fantasy is autobiographical — imagining a version of themselves, not a celebrity
What they need
Specificity of detail — the physio session, the dressing room, the agent call. Authenticity validates the fantasy
No failure states — a bad game is not game over. It's a career setback with a path forward
Quiet mornings — the morning ritual is almost meditative. Sit with the fantasy, not race through it
Churn triggers
Narrative starts to feel repetitive or templated
A decision feels arbitrary rather than consequential
The product starts to feel like a game rather than a life
A real-world event happens that the product ignores
Monetization opportunities
Subscription converts quickly — the fantasy is worth paying for
Career milestone cosmetics: custom boot deal, custom agent
World Cup / tournament season passes resonate strongly
Memoir feature: exportable career summary to keep and share
09 of 23
2
Persona 02 · 28% of user base
The Obsessive
"I consume everything about this sport. Podcasts, documentaries, transfer trackers, player stats. This product just... fits."
14×
Daily opens
V.High
Sub rate
Low
Churn risk
High
IAP spend
"I already spend 3 hours a day on football content. This just gave me a version where I'm in the story."
Composite from Discord research · top-engagement users
Who they are
Hardcore fan — encyclopedic about the sport, active on Reddit, Twitter/X, Discord
Already spending significant time on sports media daily — this product is additive
Tests the product's authenticity aggressively. Will call out anything that feels wrong.
What they need
Depth and realism — transfer mechanics that mirror real processes, accurate agent dynamics
Real-world sync — they are already tracking transfer deadline day. The product must reflect it
Community status — Discord is not a side feature for them, it's a core destination
Churn triggers
Content depth runs out — they've seen all the story branches
Real-world event not reflected within 24 hours
A factual inaccuracy — they will notice and they will post about it
Community feels inactive or unmoderated
Monetization opportunities
Highest IAP spenders — buy everything that deepens the world
Season passes convert immediately and at full price
Community status IAP: exclusive Discord roles, career badges
Your best organic UA channel — post career milestones unprompted
The Obsessive is your growth engine, not your design target
Do not design the product around them. Their depth requirements push toward complexity that alienates the Dreamer. Instead build a content depth layer accessible to them — extra narrative branches, advanced career stats, deep Discord integration — that sits beneath the surface. Make them leaders, not power users: Discord roles, community event hosting, early access to narrative arcs.
10 of 23
3
Persona 03 · 20% of user base
The Escapist
"My actual life is stressful. This is somewhere I can go where things are simpler and I'm always okay."
Daily opens
Med
Sub rate
Med
Churn risk
Low
IAP spend
"It's like ASMR but I'm living the life. I don't want to win. I just want to feel what it's like to exist in that world for a few minutes."
Composite from App Store reviews · 4–5 star ratings mentioning "relaxing"
Who they are
Fan of the sport who uses the product for emotional regulation, not fan engagement
Broader age and gender range — most likely to include women and non-core demographics
This is the ASMR connection — comfort content consumed habitually
What they need
Emotional safety — no sudden crisis arcs, no punishing consequences, no high-stakes pressure
Sensory texture — rich descriptive writing, atmospheric onboarding into each session
Predictable ritual rhythm — same comforting structure each morning. Disruption causes churn
Churn triggers
A narrative arc introduces high-stakes crisis without resolution pathway
The product introduces competitive or social pressure elements
Session pacing feels rushed or demanding
Real-world stressful events bleed too literally into the narrative
Monetization opportunities
Subscription converts on emotional value — "my calm daily ritual"
Atmosphere IAP: custom home city, apartment upgrades, lifestyle cosmetics
Off-season narrative packs — quiet moments between matches resonate most
Gifting mechanic: likely to send gifts to friends who share the sport
The Escapist is the underserved opportunity in the category
No existing product is designed for them. Sports games are competitive. Football apps are informational. ASMR is passive. This product is the only thing that gives them an active emotional experience with zero pressure. This likely explains a portion of the strong female-skewing reviews Footballer is receiving. Consider a "calm career" narrative track — slower pace, less transfer drama, more lifestyle texture.
11 of 23
4
Persona 04 · 10% of user base
The Insider
"I want to understand how the sport actually works — the business of it, the psychology, the decisions behind the decisions."
11×
Daily opens
V.High
Sub rate
Low
Churn risk
High
IAP spend
"I work in sports management. This product gets the dynamics right in a way no game ever has. The agent negotiation feels real. The contract language feels real."
Composite from Discord · professionals, coaches, sports industry workers
Who they are
May work in or adjacent to the sport — coaching, management, media, sports business
Small percentage but highest credibility signal — their presence validates authenticity to everyone else
Most likely to become Discord community leaders and vocal advocates
What they need
Factual accuracy — they will test every system against their knowledge of how the sport actually works
Zero gamification signals — XP bars, level-up sounds, achievement badges all break their immersion immediately
Community authority — they want their domain expertise recognized and useful within the community
Churn triggers
A factual inaccuracy that reveals shallow research — they tell everyone
Gamification elements appearing — any visible game mechanics break the spell
Community goes unmoderated — they disengage from low-signal environments
Product simplifies a mechanic they know to be complex
Monetization opportunities
Highest LTV per user — pay for depth, pay for access, pay for status
Expert narrative tracks: agent career mode, manager perspective mode
Advisory Discord role with product team access — priceless for them, free for you
Consultant pipeline: some will become actual product advisors
The Insider is your credibility infrastructure
10% of users, disproportionate influence. When an Insider validates the product's authenticity publicly — on Twitter, in a Reddit thread — it converts Dreamers and Obsessives at a rate no paid creative can match. Before launching any new title, run a 30-person closed beta with Insider-profile users from that sport's community. Their authenticity verdict determines whether Pillar 5 holds.
12 of 23
Part III · User Personas
13
Part III · User Personas
Persona Spectrum
How the four personas compare across the dimensions that matter most for product and business decisions.
Escapist 6×
Dreamer 9×
Insider 11×
Obsessive 14×
Escapist — Med
Dreamer — High
Obsessive — V.High
Insider — V.High
Escapist — Low
Dreamer — Med
Obsessive — High
Insider — Critical
Decision 1 — Narrative complexity
Design for the Dreamer, layer depth for the Obsessive
The surface narrative should be emotionally rich but mechanically simple — the Dreamer is your 42%. Obsessives and Insiders will find the depth if it's there. Like a great novel: a reader can finish it and feel moved without understanding every technique. Another reader can spend weeks unpacking the structure. Same product. Both satisfied.
Decision 2 — Community design
Let the Obsessive and Insider run the community. Don't run it for them.
Tapugo's community ops cost is near zero because passionate Obsessives and Insiders are doing the work. Give them tools and roles, not managed content. A community event designed by an Obsessive resonates more than one designed by the studio. The studio's job is to create the conditions — real-world moments, career shareables — that give the community something to talk about.
Decision 3 — New title validation
The Insider beta test is your launch insurance
Before any new title ships, run a 30-person closed beta with Insider-profile users from that sport's community. Their authenticity verdict determines whether Pillar 5 holds. A single "this doesn't feel real" verdict in the first week of App Store reviews is disproportionately damaging in a category where the entire value proposition is immersion.
13 of 13 · End of playbook
Part IV · Narrative Design
Tapugo · Internal Document
Narrative Design System
The content architecture behind appified games — how to write scenarios, build decision trees, and sustain authentic narrative at scale with a 2-person team.
10 chapters Beat → Voice → State → Arc AI-native production pipeline

The core insight: Narrative at scale in appified games is not a writing problem — it is an architecture problem. Studios that treat it as a writing room hire writers, produce scenarios, and churn at D30 when users exhaust the content. Studios that treat it as a software system build modular, composable, consistently voiced narrative that compounds. This document is the architecture.

The four layers of narrative architecture
Layer 1 — The Beat
The atomic unit. Stimulus → Decision → Consequence. Every scenario is a chain of beats. Master the beat and everything else is composition.
Layer 2 — The Voice System
Who speaks and how. Each character has a documented psychology, agenda, and register. Any writer — human or AI — can produce authentic copy for any voice.
Layer 3 — State Variables
The career is not a story, it's a set of states: form, reputation, fitness, relationships. States shift incrementally and unlock different beat types based on their values.
Layer 4 — Calendar Integration
The real world is the content engine. Transfer windows, match days, cup draws — predictable triggers that fire the right arc types automatically.
NDS Overview
Part IV · Narrative Design
01
Foundation · Layer 1
The Beat — The Atomic Unit
The smallest meaningful narrative piece. Everything in the product — morning triggers, match day arcs, transfer negotiations — is a chain of beats. Build the beat correctly and the entire system scales from it.
"A beat is not a scene. It is a moment where the user's world changes based on their choice."
Beat anatomy
Stimulus
Something arrives
A message, news, event, or change in the world that demands a response. Delivered via the character who owns the relationship — agent, physio, press, manager.
Decision
The user chooses
2–3 options. Never more. Each option must feel like a real human choice — not "good option" vs "bad option" but two legitimately compelling paths with real tradeoffs.
Consequence
The world shifts
One or more state variables change. A relationship warms or cools. A future beat is unlocked or closed. The consequence is felt in narrative — not a stat number — first.
Example beat — transfer rumor
Stimulus: "Marco (agent) · 07:14 — A Premier League club made an inquiry overnight. They want an answer before the weekend. I need to know where your head is."

Options: A) "Tell them I'm interested — let's see the numbers."  |  B) "Not now. I'm focused on the title race."  |  C) "Ask them to wait until the window. Don't close the door."

Consequence A: Agent relationship +2. Club loyalty −3. Transfer probability 40%. Press starts writing speculation articles. Manager calls a meeting.
The three beat types
Type 1 — Tension beat
Creates a micro-narrative that needs resolving. Transfer rumors, contract standoffs, form dips, injury concerns. The user feels mild anxiety that only closes when they respond. Highest open urgency. Use 40% of the time.
Type 2 — Reward beat
Validates the career. Press recognition, sponsor interest, national team consideration, fan appreciation. Opens a longer session — users sit in the good feeling. Use 30% of the time.
Type 3 — World beat
Something real happens that enters the career narrative. A rival gets injured. A league decision affects your club. The World Cup squad is announced. Highest reopen rate. Triggered by the real calendar. Use 30% of the time.
What makes a beat fail
False choices — one option is obviously better
If the user never deliberates, you've written a menu not a decision. Both options must have something genuinely worth losing. "Accept transfer to Champions League club" vs "Stay for the title push" — that's a real choice. "Be kind to your physio" vs "Be rude to your physio" is not.
Consequence felt only in numbers
If the only consequence is "Form −2," the beat is a stat update, not a narrative. The consequence must land in the story first: the manager drops you from the starting XI, the press asks about your fitness, your agent calls with concern. Numbers follow story.
Stimulus without stakes
"How did training feel today?" is a stimulus without stakes. "The physio pulled you aside after training. Your knee is reacting. He's recommending a rest day before Saturday's derby." has stakes. Every stimulus must carry something the user cares about losing or gaining.
More than 3 choices
Decision paralysis breaks the habit loop. Appified game sessions are 30–90 seconds. Three options is the ceiling. Two is often better. One is sometimes correct (acknowledge a consequence, accept a reality).
14 of 23
Part IV · Narrative Design
02
Foundation · Layer 2
The Voice System
Every character in the user's world has a documented psychology, distinct register, and specific agenda. This is what allows narrative to scale — any writer or AI prompt can produce authentic copy for any character because the character's inner logic is fully specified.
"The user doesn't read a message from 'the game.' They read a message from Marco, their agent — a man with his own ambitions."
Core voice roster — football
Marco — The Agent
Career broker · primary relationship
Agenda
Maximise client value. Commission-driven but genuinely cares about long-term career.
Register
Direct, conspiratorial. Short sentences. Never uses the word "game." Texts like a trusted advisor who's also always working an angle.
Tension
His interests occasionally diverge from the user's — a bigger club means a bigger commission.
"Heard something this morning. Don't go public. Call me when you're done with training."
Coach / Manager
Authority figure · consequence giver
Agenda
Win matches. The user is an asset to be managed, not a person to be pleased.
Register
Formal, measured. Speaks in team terms. Never praises without purpose. When he's warm, it means something.
Tension
Tactical decisions can override loyalty. Being dropped is always possible regardless of form.
"I want to see more pressing in the final third. We'll talk before the team meeting on Friday."
Press Digest
Reputation mirror · public witness
Agenda
The story. Neutral but amplifying — will run with whatever narrative has momentum.
Register
Third-person. Newspaper headline style. Short. Uses ratings and rankings. References real publications by name for authenticity.
Tension
Creates the Obsessive user's dream: their career being covered like a real player's.
"L'Équipe 8.2 — 'The most complete midfielder in Ligue 1 right now. But questions remain about the knee.'"
Physio / Dr. Roux
Body steward · friction creator
Agenda
Protect the body. Will always recommend rest when the user wants to push. Honest to a fault.
Register
Clinical but warm. Never panics. Uses actual medical terminology lightly — enough to feel real, not so much it alienates.
Tension
The body vs the ambition. Playing through an injury to not miss the derby.
"Grade 1 strain. You can play Saturday. But if it goes to Grade 2, we're looking at three weeks minimum."
Voice spec template — for new titles
Required documentation for every voice character
Agenda: What does this character actually want? How does it overlap and conflict with the user's goals?
Register: Sentence length, formality level, vocabulary range, signature phrases. Two or three example messages that define the ceiling and floor of their voice.
Relationship arc: How does this voice change as the user's career progresses? A loyal agent at level 1 becomes a conflicted one when a rival club calls.
Trigger ownership: Which beat types does this character deliver? The physio owns injury beats. The agent owns transfer beats. Press owns reputation beats. Never cross streams — it breaks authenticity.
Never says: A short list of phrases that would break the character. The manager never says "amazing job." The agent never says "it's just a game." The physio never says "you'll be fine" without qualification.
Voice consistency at scale
The voice bible
A living document (Notion or similar) with full spec for every character. Updated when new scenarios expose edge cases. Every piece of content — human or AI-generated — is checked against it before shipping.
Voice-first content review
When reviewing AI-generated copy, the first question is always: "Does this sound like Marco?" Not "Is the story good?" Voice breaks first. Story problems are usually voice problems in disguise.
Character ownership
Each character has a single internal owner on the team who approves all copy for that voice. Consistency requires a single ear, not committee review. Two people = two character owners each.
15 of 23
Part IV · Narrative Design
03
Foundation · Layer 3
State Variables
The career is not a story — it is a set of states that shift incrementally with every beat. State variables determine which beats are available, which characters are active, and what the consequence of each decision feels like. This is the invisible scaffolding that makes the narrative feel alive rather than linear.
"A player with Form 80 and Club Loyalty 20 lives in a completely different narrative than one with Form 30 and Club Loyalty 90. Same beats, different world."
Core state variables — football
Form
0 – 100
Affects press coverage, manager relationship, transfer interest, and national team consideration. Below 40: manager beat becomes available. Above 85: sponsorship beats unlock.
Fitness
0 – 100
Low fitness locks out match beats and triggers physio narrative. Pushing through injury risks Grade 2+ and 3-week arc. Full fitness unlocks intensity training beats.
Reputation
0 – 100
Drives press coverage volume, sponsor deal tier, fan interaction tone, and national team manager interest. Reputation >80 unlocks celebrity / crossover beats.
Club loyalty
0 – 100
Determines transfer beat tone. Low loyalty + high form = maximum transfer drama. High loyalty unlocks contract extension arc and captain consideration narrative.
Agent trust
0 – 100
Low trust means agent starts withholding information in beats. High trust unlocks exclusive opportunity beats. Agent can be fired — triggering a full replacement arc.
Career stage
Emerging → Peak → Late
Unlocks different arc types. Emerging: breakthrough arcs. Peak: legacy and transfer arcs. Late career: retirement consideration, coaching future, mentorship beats.
How states gate content
State-gating rules
Beat availability: Every beat type has a minimum state threshold. Transfer beats require Form > 60. Injury recovery beats require Fitness < 50. The state system ensures users never receive contextually inappropriate content.
Consequence tuning: The same decision produces different narrative consequences depending on current state. Declining a transfer when Club Loyalty is 80 produces warmth. Declining when it's 20 produces skepticism from the press about your motives.
Arc unlocking: Career arcs are not on a timer — they unlock when the right combination of state variables aligns. The "Contract Year" arc unlocks when Club Loyalty < 50 AND Form > 70 AND a rival club has been in contact. This makes arcs feel earned, not scheduled.
Never expose numbers to the user: The user never sees "Form: 72." They see the world reacting. The manager picks them for the big match. The press writes favorably. The agent gets calls. The numbers are purely for the system — the narrative is the output.
State drift — keeping the narrative alive
Passive drift
States drift slowly without input. Form decreases if no training beats are completed. Fitness erodes during match runs. This creates natural tension that demands the user's attention — the career needs maintenance, like a real one.
Volatile variables
Some variables can spike sharply: a single press incident can drop Reputation 15 points. A dominant performance in a derby can push Form 20 points. Volatility creates drama — the career is never on cruise control.
Long-burn variables
Career Stage moves only once every 6 months of real time. The user knows a transition is coming, they just don't know exactly when. The slow burn of approaching a career phase change is its own narrative engine.
16 of 23
Part IV · Narrative Design
04
Decision Architecture
Decision Trees
How individual beats connect into multi-step scenarios. A transfer negotiation is not one beat — it is a tree of 4–6 beats, where each decision narrows or expands the available branches. The tree design determines how many meaningful paths exist without requiring an infinite content library.
A worked example — transfer negotiation tree
Scenario: Premier League approach · mid-season
Beat 1 · Stimulus from Marco (Agent)
"Arsenal have made an approach. 40% increase on your current wage. They want an answer in 48 hours."
Option A — Engage
"Tell them I'm listening. Get me the details."
Opens Beat 2A: Contract details arrive. Manager senses something. Press starts speculating.
Club Loyalty −5 · Agent Trust +3
Option B — Deflect
"Not now. We're in the title race. Tell them to wait."
Opens Beat 2B: Marco pushes back. Arsenal's interest stays warm but the window is tightening.
Club Loyalty +4 · Agent Trust −2
Option C — Refuse
"I'm not leaving. Close it down."
Closes transfer arc for this window. Unlocks loyalty arc — captain conversation opens within 2 weeks.
Club Loyalty +12 · Reputation +5
If Option A → Beat 2A unlocks
"The contract sheet came through. The numbers are real. And this morning, the gaffer pulled you aside after training. He knows something."
Tree design principles
Width over depth
A tree with 3 branches at Beat 1 and 2 branches at Beat 2 gives 6 paths from 5 beats. A linear story of 6 beats gives 1 path. Same content budget, 6× the replayability. Keep trees wide and shallow — not narrow and deep.
Converging paths
Most trees should converge to a shared outcome beat — the transfer happens or it doesn't — with different narrative texture depending on how you got there. Full divergence is expensive. Convergence with earned variation is efficient.
Dead ends are not failures
A path that closes an arc (Option C above) is not a failure state — it's a different success. Make every dead end feel like a deliberate career choice that opens something else. The user chose loyalty. That should feel consequential and respected, not punished.
The maximum tree
No scenario tree should exceed 6 beats. Beyond that, maintenance becomes impossible and the user loses the thread. A 6-beat tree with 2–3 options per beat is 12–18 pieces of content that generates months of distinct paths.
Standard tree templates
Tree typeBeat countTriggerConverge point
Transfer approach4–6 beatsForm > 65 + rival club activeWindow closes with/without move
Injury arc3–4 beatsFitness < 45Return to training beat
Contract standoff5 beatsClub Loyalty < 40 + contract <6moSign, walk, or extend short-term
Manager conflict3 beatsForm < 50 + 2 consecutive poor matchesDropped / restored to starting XI
National team call-up4 beatsReputation > 70 + Form > 75Squad named — in or out
Sponsor deal2–3 beatsReputation > 60Deal signed or declined
Rival emergence3 beatsWorld beat triggers + Reputation > 50Response to comparison coverage
17 of 23
Part IV · Narrative Design
05
Decision Architecture
Consequence Chains
How the effects of a decision ripple forward in time. A good consequence chain means a choice made in January is still affecting the narrative in March — without the user needing to remember the original decision. The world remembers for them.
"Consequences are not punishments. They are the world's response to the person you're becoming."
The ripple structure
Three consequence horizons
Immediate (same session): The character who sent the stimulus responds. A relationship warms or cools. The next beat in the tree becomes available. The user feels the world react within minutes.
Short-term (1–7 days real time): A new character reacts to the consequence. Press runs a story. Manager references it in the next team meeting beat. The agent circles back. The decision lives in the world for a week.
Long-term (weeks to months): State variable shifts from the original decision compound into arc unlocks or closures. The user who chose loyalty in January finds the captain arc opening in March. The chain is long enough that it feels like real career consequence, not scripted cause-and-effect.
Consequence chain example — injury decision
Day 1 · The decision
Physio says rest. You choose to play through the discomfort in Saturday's derby. Fitness −8. Immediate beat: You perform well. Form +5. Press praises the commitment.
Day 4 · Short-term ripple
Physio beat: "The scan came back. Grade 1 strain confirmed. I need you to be honest with me about what you're feeling." Agent beat: "The manager asked me directly. I told him you're managing it. Don't make me a liar." Club Loyalty +3 (played through it for the club).
Day 14 · Long-term consequence
Two weeks later: a second knock in training. Physio beat: "I told you. Grade 2 now. Three weeks minimum. I'm sorry." The injury arc that was avoidable at Day 1 has now arrived — with a different emotional texture because the user caused it through their own choice.
Making consequences feel earned, not punitive
Consequences must honor the original choice's internal logic
If you chose to play through injury for the club, the consequence chain should acknowledge the sacrifice even while delivering the downside. The manager's gratitude is real. The injury is also real. The user made a hard choice and the world reflects its complexity.
Positive consequences compound too — not just negative ones
Consistently choosing loyalty unlocks a captain arc. Consistently managing fitness unlocks a "physical peak" beat that gives you one season of elevated performance. The user who plays well gets reward chains, not just the user who makes mistakes getting consequence chains.
Never punish without a clear link to the original decision
If a consequence arrives without the user being able to trace it to their own choice, it feels arbitrary and unfair. The narrative must always signal the connection — even subtly. The physio's "I told you" is a narrative link, not just a reminder. It validates the user's memory and their agency.
18 of 23
Part IV · Narrative Design
06
Decision Architecture
Career Arcs
Arcs are multi-week narrative structures that give users something to be in the middle of at all times. They are the chapter-level structure above beats and trees. A user who has just entered the "Contract Year" arc behaves differently from one in the "Breakthrough" arc — even if they're seeing similar beats.
"The user's daily ritual has depth because they know they're somewhere in a story, not just receiving notifications."
Arc library — football
Arc nameDurationUnlock conditionNarrative themeEnds with
The Breakthrough6–8 weeksCareer Stage = Emerging + First big performanceProving you belong. Proving it to the manager, the press, and yourself.First team regular or loan move decision
Injury comeback4–6 weeksFitness < 30 for 7+ daysThe fear of not returning the same. Physio as confidant. Small victories.Return match beat — full performance or setback
The Contract Year8–10 weeksContract < 6 months + Club Loyalty variable activeLoyalty vs ambition. Every match now has a subtext.Sign, leave, or enter free agency
National team push6 weeksReputation > 70 + Form > 75 + international window approachingThe dream within the dream. Does your club form translate internationally?Squad selection beat — called up or not
The rival4–5 weeksWorld beat introduces a direct competitor at same positionComparison, competition, respect. Not hatred — complexity.One of you moves club, retires, or earns clear seniority
Late career legacy10–12 weeksCareer Stage = Late + Reputation > 60What do you want to be remembered for? Mentorship, one last title, a dignified exit.Retirement decision beat or one-year extension
The World Cup6 weeks (real calendar)Triggered by real World Cup calendar + national team activeOnce-in-a-career moment. Pressure unlike any club beat.Tournament elimination or final — shareable career card generated
Arc design rules
One active arc at a time
Never run two major arcs simultaneously. Daily beats can reference both but the dominant emotional thread is singular. The user is either in the Contract Year or in the National Team Push — not both at full intensity.
The arc must change the user's relationship to daily beats
If the Injury Comeback arc doesn't change how a training beat feels — if it's just the same content with an arc label — it has failed. Arcs are an emotional lens, not a content category.
Arc endings are the most important beats
The arc ending is the shareable moment, the Discord spike, the "second first impression" that re-hooks D14 users. Spend disproportionate effort on arc endings. The rest of the arc earns the right to a great ending beat.
Arcs seed the next arc
The final beat of every arc should plant a seed for the next one. Signing the contract extension in the Contract Year arc should immediately activate a low-level "club leadership" narrative thread that will eventually become the captain arc.
19 of 23
Part IV · Narrative Design
07
Scale & Production · Layer 4
Calendar Integration
The real world is the content engine. The football calendar generates high-stakes narrative triggers on a predictable schedule — for free. Calendar integration is the system that maps real-world events to in-product beat types automatically, ensuring the product always feels alive without proportional content effort.
"Transfer deadline day is not a live ops event you design. It's an event that already exists. You just have to be ready for it."
The football calendar as a content schedule
Always on
Matchday −1: physio + prep beats Matchday: real result → narrative beat Matchday +1: press reaction
Weekly
Form update beat Training selection beat Agent check-in
Jan / June
Window opens: transfer beat queue activates Deadline day: maximum tension arc
Intl break
National team squad announcement beat Country leaderboard on Discord
World Cup
Full tournament arc activates Daily elimination beats Career summary card generated
The 48-hour real-world response rule

Rule: Any major real-world football event must surface as a narrative beat within 48 hours of occurring. A high-profile injury, a shock result, a managerial sacking — these enter the career world as Press Digest beats, agent commentary, or manager reactions. The Obsessive user is already talking about it on Discord. If the product hasn't acknowledged it, the product feels fake. 48 hours is the maximum acceptable lag.

Pre-written beat libraries for calendar events
Event beat library
For every predictable calendar event — transfer deadline, World Cup draw, international break — maintain a library of 8–12 pre-written beats that can be deployed immediately when the event fires. No reactive writing required. The content exists; the calendar triggers it.
Rapid beat production
For unpredictable events (shock signing, player scandal, managerial drama), a single writer should be able to produce a reactive beat in under 2 hours using the voice spec and beat template. The system makes fast production possible because the structure is pre-defined.
Calendar as content roadmap
Map the entire football calendar 6 months forward. Every significant event has a beat type assigned, a voice character nominated, and a state variable interaction planned. Content production follows the calendar, not the reverse.
20 of 23
Part IV · Narrative Design
08
Scale & Production
The AI Content Engine
Tapugo runs on 2 people. The narrative design system only scales if AI can generate production-ready beat content reliably. This chapter defines the prompt architecture, quality requirements, and human review layer that makes AI-native narrative production possible without losing voice consistency.
"AI doesn't replace the narrative designer. It makes one narrative designer as productive as ten."
The production pipeline
Step 1 — Beat brief
A human writes a 3-line brief: beat type, triggering state conditions, character delivering it, and the emotional register. This takes 5 minutes. The brief is the creative decision. Everything else is production.
Step 2 — AI generation
A structured prompt containing the voice spec, beat template, and brief generates 3 variants of the beat. Variants give the human reviewer options without requiring full rewrites. The best variant is usually 80% done.
Step 3 — Voice review
The character owner does a 5-minute review: Does this sound like Marco? Does the consequence feel earned? Is the decision genuinely difficult? One pass. Ship or return with a single revision note.
Step 4 — State wiring
The developer connects the approved beat copy to the correct state variable triggers and consequences in the content database. This is a 10-minute technical task, not a creative one.
Master prompt architecture
Beat generation prompt — master template
You are writing narrative content for Footballer Life Simulator, a mobile product where users live the daily life of a professional footballer. Your output will be delivered directly to users as messages, press reports, and character communications.

Character voice: {CHARACTER_NAME} — {CHARACTER_REGISTER_SUMMARY}
Beat type: {BEAT_TYPE} (tension / reward / world)
Current player state: Form {FORM}, Fitness {FITNESS}, Club Loyalty {LOYALTY}, Career Stage {STAGE}
Brief: {3_LINE_BRIEF}

Write the stimulus message (max 3 sentences, in character), then 2–3 decision options (max 12 words each), then the consequence narrative for each option (1–2 sentences, felt as story not numbers).

Rules: Never use the word "game." Never show stat numbers to the user. Every decision must have a genuine cost and benefit. Write in the character's voice — not a narrator's voice.
{CHARACTER_NAME} {CHARACTER_REGISTER_SUMMARY} {BEAT_TYPE} {FORM} {FITNESS} {LOYALTY} {STAGE} {3_LINE_BRIEF}
Prompt variants for specific beat types
Transfer rumor beat — agent voice
Add to master prompt: "The user's club is in the title race. The offer is real and substantial. The agent's conflict of interest (commission) should be subtly present without being stated. The message should feel like a confidence being shared, not a transaction being proposed."
Injury beat — physio voice
Add to master prompt: "Use one real medical term for the injury type. The physio is never alarmist but never falsely reassuring. The decision should force the user to choose between the body and the moment. Make the trade-off feel real, not melodramatic."
What AI cannot do — the human layer
AI cannot assess authentic sports texture
An AI will write plausible-sounding football scenarios that feel slightly off to anyone who has actually played or worked in the sport. The CEO's domain knowledge is the irreplaceable quality gate. Every beat that involves real tactical, medical, or contractual detail must pass through a domain expert.
AI cannot make arc-level decisions
Which arc to activate when, how to sequence the long-term consequence chains, when a career needs a new emotional direction — these are design decisions that require understanding of the whole narrative system. AI produces beats within arcs; humans design the arcs.
AI cannot catch voice drift over time
After 200 AI-generated Marco beats, the character will have subtly drifted. Small inconsistencies compound. Monthly voice audits — reading 20 recent beats for each character — catch drift before users do. This is a human task that takes 30 minutes per character.
21 of 23
Part IV · Narrative Design
09
Scale & Production
The Writing Rules
20 rules that govern every piece of narrative content in the system — whether written by a human or generated by AI. These are not style preferences. They are structural requirements. Content that violates them breaks immersion and triggers churn.
The 20 rules
01
Never use the word "game"
Not in any character's voice, not in any consequence, not anywhere. The word breaks the fourth wall instantly.
02
Never show numbers to the user
No "Form +5." No "Reputation: 72." Numbers are backstage. Narrative is the stage.
03
Every stimulus must have a named sender
"Marco (Agent) · 07:14" not "Notification." Anonymous messages break immersion.
04
Stimuli are 1–3 sentences maximum
Users open this product 10 times a day in 30-second windows. The stimulus must be scannable at a glance.
05
Decision options are 12 words maximum
The choice must be instantly clear. "Tell them I'm interested — let's see the numbers." No more.
06
Both options must have something worth losing
If one option is obviously correct, it's not a decision. The user must feel the cost of each path.
07
Consequences land in story before state
The manager drops you from the XI. The agent goes quiet. The press asks questions. Then the numbers shift.
08
Use real sports terminology precisely
Grade 1 strain, not "minor injury." Pressing trigger, not "the thing you do when you don't have the ball." Specificity = credibility.
09
Morning beats are quieter than evening beats
The morning trigger sets the emotional tone for the day. It should feel like news, not crisis. Evening match beats can carry more intensity.
10
No villain characters
Every character has understandable motivations. The rival isn't evil — they're ambitious. The agent isn't corrupt — they're conflicted. Complexity retains better than antagonism.
11
Reward beats must earn the warmth
A reward beat that arrives without context feels hollow. The press calling you "the best midfielder in Ligue 1" after three good matches feels earned. After one means nothing.
12
Never resolve a tension beat in the same session
If the agent sends a concerning message in the morning, the consequence arrives in the evening beat or the next day. The unresolved tension is the pull that brings users back.
13
The career always moves forward
No matter what the user chooses, the narrative progresses. There is no stalling, no "you failed," no reset. Different paths — never a stopped clock.
14
World beats reference real events, not fictional ones
The Champions League draw is real. The rival's injury is based on real news. The press article cites an actual publication. Fictional world events break the product's core value proposition.
15
Physical details ground the fantasy
"The warm-down session ran 20 minutes over." "Your ankle was taped before the captain's walk." Sensory specificity is what makes the Escapist user feel like they're inside the world.
16
Each character speaks once per session maximum
Receiving three messages from Marco in one session feels like a conversation — it should feel like a world. Distribute voices across sessions.
17
Arc beats are emotionally escalating
Beat 1 of the Contract Year arc has lower stakes than Beat 5. The arc must build. Flat arcs feel like repetition, not story.
18
The player is always the most capable person in the room
The fantasy requires the user to feel like a professional of genuine ability. Write them as intelligent, skilled, and considered — not as someone being managed by everyone around them.
19
Match results follow real scores where applicable
If the user plays for PSG and PSG won 3–0 in real life yesterday, the match beat reflects that result. Calendar sync is a narrative commitment, not just a feature.
20
When in doubt, ask: would a real footballer's agent actually say this?
This is the ultimate authenticity test. If the answer is no — if it sounds like a game designer wrote it rather than a footballer lived it — rewrite it.
22 of 23
Part IV · Narrative Design
10
Scale & Production
Quality & Testing
A narrative system is only as good as the weakest beat a user encounters. Quality control in appified games is not about grammar — it is about immersion. A single beat that breaks voice, presents a false choice, or delivers a consequence that feels arbitrary can end a habit loop that took two weeks to form.
The beat quality checklist
Run every beat through these 7 questions before shipping
Voice test: Read the stimulus aloud. Does it sound like a real message from this character? Would the character's owner approve it without edits?
Decision test: Are both options genuinely tempting? Could a reasonable person in this career situation make either choice? If one is obviously correct — rewrite.
Stakes test: What does the user stand to lose or gain? If the answer is "not much," the beat has no tension. Increase the stakes or cut the beat.
Consequence test: Does the consequence feel proportionate and earned? Does it land in story before state? Is there a clear narrative link back to the decision?
Authenticity test: Does any word, phrase, or scenario feel like it was written by someone who hasn't played or worked in professional football? If yes — it probably was. Fix it.
Length test: Is the stimulus 3 sentences or fewer? Are the options 12 words or fewer? If not — cut.
Context test: Is this beat reaching the right user at the right state? Would a user with Form 20 and Fitness 30 receiving a sponsorship beat feel authentic? Check the state conditions.
Regular quality practices
Monthly voice audit
Read 20 recent beats per character in sequence. Listen for voice drift, register inconsistencies, and any phrases that have crept in from the "wrong" character. Takes 30 minutes per character. Non-negotiable.
Insider user review
Before any new arc or beat library ships, send it to 3–5 Insider-profile users (sports professionals, domain experts) for a 24-hour authenticity review. Their notes surface the inaccuracies no internal review catches.
D7–D14 churn signal analysis
When D14 churn spikes, pull the beats users saw in their D8–D14 window. Look for beat type distribution, voice character concentration, and decision quality. Churn spikes almost always have a content signature.
Session exit timing
Track which beats precede session exits most frequently. A beat with high exit rate is either too long, a false choice, or has an unsatisfying consequence. The data tells you which beats need rebuilding before users tell you in reviews.
The content depth runway

Target: At 30 days of engagement, a user should have seen no more than 40% of available beats. At 60 days, no more than 65%. The system should feel inexhaustible within the first 3 months. With 7 arc types × 4–6 beats per arc × 3 variants per beat × 3 tree branches = approximately 500 distinct narrative moments. That is the minimum viable content depth for a product aiming at 90-day retention.

Min beats at launch: 200+
Arc types: 5 minimum
Voice characters: 4–6
Beat variants per scenario: 3
Monthly new beats: 30–50
Voice audit: monthly
Insider review: every arc
31 of 31 · End
Part V · Onboarding Architecture
10
Tapugo · Onboarding Architecture
The First 10 Minutes
How to drop a user inside an identity they've never been offered before — and make them believe in it before the session ends.
5 phases D1 retention architecture No tutorials. No overlays.

The fundamental problem: You can't tutorial someone into a fantasy. The moment onboarding feels like onboarding — overlays, feature tours, "tap here to continue" — the illusion collapses. Users have a model for games. They have a model for apps. They have no model for what this is. You have to drop them inside it before they realize what's happening.

The five phases at a glance
0:00 – 0:45
Phase 1 — The Drop-In
No splash screen. No feature list. The world begins mid-scene.
45 seconds
0:45 – 3:00
Phase 2 — Identity Commitment
4 decisions. No forms. Every choice expands the world.
2 minutes
3:00 – 5:30
Phase 3 — The First Beat
The best content in the product. Front-loaded, not saved.
2.5 minutes
5:30 – 8:00
Phase 4 — The World Opens
Second and third voices arrive. The career has texture.
2.5 minutes
8:00 – 10:00
Phase 5 — The Open Thread
Something stays unresolved. The pull to return is set.
2 minutes
Overview
Part V · Onboarding Architecture
T
Design Challenges
Three Core Tensions
The specific design problems that make appified game onboarding different from any other category — and why conventional onboarding playbooks actively harm it.
"Every onboarding decision either reinforces the fantasy or breaks it. There is no neutral."
Tension 1 — Identity commitment before investment
The problem
The user needs to make enough identity choices to feel like the career is genuinely theirs. But too many choices feel like character creation in an RPG — which signals "game," breaks the fantasy, and raises the cognitive cost of entry.
The line: 4 decisions is the maximum. Position, club region, playing style, age. These four choices generate infinite narrative specificity without feeling like a form. Five or more choices feel like registration.
The framing: Decisions must be presented as narrative moments, not settings screens. "Where do you play?" not "Select position." "Which league calls to you?" not "Choose league." The language of the world, not the language of the app.
Tension 2 — Best content must be first, not saved
The problem
Most products save their best content for when users are retained. Appified games cannot. The user's entire mental model of what this product is — and whether it's worth returning to — forms in the first 90 seconds of narrative.
The implication: The very first message the user receives must be the most emotionally resonant, most authentic, most specifically crafted beat in the entire product. Not a tutorial. Not a welcome message. The best possible version of the fantasy, delivered immediately.
The test: Read the first beat aloud. Does it make you feel something? Does it sound like a real person in a real situation? If the answer to either is no — it's not ready to ship.
Tension 3 — Habit signal before session end
The problem
Before the user closes the app for the first time, they must feel that something is unresolved in their world. Not a cheap cliffhanger — a genuine open thread that creates the pull to return. This converts a curious first session into a D1 return.
The mechanism: A stimulus is delivered — the agent sends a message, a press report references you, a manager meeting is scheduled — and then the session ends before resolution. The unresolved thread lives in the user's mind until they return to close it.
The D1 push: The next morning's push notification should reference the open thread directly. "Marco left another message. He needs your answer." The notification lands with weight because the user already knows what it means.
Tensions
Part V · Onboarding Architecture
01
Phase 1 · 0:00 – 0:45
The Drop-In
No splash screen. No loading spinner with a tagline. No feature carousel. The world begins mid-scene — and the user's job is to find themselves inside it.
"The first thing they see is not an app. It is a moment already in progress."
What the user sees first
Pre-dawn · 05:47
FOOTBALLER LIFE
The stadium lights are still on from last night. You can see them from your apartment window. You've been awake for an hour.
MARCO — AGENT · 05:52
Call me when you're up. Something came in overnight. Don't read the press before we talk.
Why this works

The user has not been welcomed. They have not been shown features. They have been placed inside a moment that has already happened — a message that arrived while they were asleep, from someone who clearly expects them to already know the context.

The atmospheric opener (stadium lights, pre-dawn) establishes sensory texture before anything mechanical happens. The user's brain is building a world before they've made a single decision.

The agent's message creates immediate mild tension — "Don't read the press" — without explaining why. This is the first pull. The user wants to know what happened.

Drop-In design rules
Begin in the middle of a moment, not at the beginning of a story
The career has history before the user arrived. The agent's message implies prior conversations. The stadium lights imply last night's match. The user joins a world that was already running.
Atmospheric text before interactive elements
8–12 words of sensory scene-setting before the first message. Grounds the user physically in the world before they're asked to do anything. The Escapist persona in particular needs this landing pad.
Never open with "Welcome to Footballer Life Simulator"
The word "Welcome" signals an app. The product name in the first sentence signals a game. Both break the fantasy before it begins. The product name appears nowhere in the first 90 seconds of narrative.
Never open with an explanation of how the product works
No "You'll receive daily messages from your team..." No "Make decisions to shape your career..." The product explains itself by doing, not by describing. If it requires explanation before the first beat, the first beat is wrong.
Timing and pacing
0:00 – 0:15 · Scene
Atmospheric text appears. No interaction required. Slow fade-in. The user is given permission to read and feel before acting.
0:15 – 0:35 · Message arrives
The agent's message appears with a notification animation — as if it just arrived. The familiar notification format makes the world feel real before the user has processed that it isn't.
0:35 – 0:45 · Invitation
A single quiet prompt: "Who are you?" No button. No call to action. Just the question, in the agent's voice, that begins Phase 2.
Phase 1 of 5
Part V · Onboarding Architecture
02
Phase 2 · 0:45 – 3:00
Identity Commitment
Four decisions. Each one delivered as a narrative moment rather than a settings screen. By the end, the user has a career that feels specifically theirs — not a template.
"The user isn't creating a character. They're remembering who they already are."
The four decisions
MARCO — AGENT
"Before we get into the call — remind me. Where do you play best?"
Your answer
Attacking midfielder — I create and score
Centre-forward — I live in the box
Deep-lying playmaker — I see the game early
Right winger — pace is my language
Decision 1 — Position
Framed as a question from the agent, not a settings screen. The choices are written in the first person and carry personality — "pace is my language" tells you something about the player you're becoming before you've chosen. The choice persists as a state variable that shapes future narratives.
MARCO — AGENT
"And the club situation — which league are we thinking? I've had interest from a few directions."
Your preference
Ligue 1 — France is where I want to prove myself
Premier League — the biggest stage
La Liga — technical football, my style
Bundesliga — hungry, ambitious, rising
Decision 2 — League
Each option has a personality beyond the factual. "France is where I want to prove myself" implies an underdog arc. "The biggest stage" implies a peak ambition arc. The user isn't just selecting a game setting — they're selecting a narrative identity. The league choice gates which real-world calendar events become relevant to the career.
Decision 3 — Playing style
Marco: "Last thing — how would you describe your game? I need to know what to sell."
Options: Clinical / Technical / Physical / Intelligent. Each maps to a different press voice and manager relationship style. "Intelligent" players get more tactical manager beats. "Clinical" players get more goal-centric press coverage.
Decision 4 — Career stage
Marco: "How long have you been at this level? I want to pitch you right."
Options: Breaking through / Establishing myself / In my prime / Building a legacy. Sets the Career Stage state variable and unlocks the appropriate arc type from day one. This is the most consequential of the four choices.
Identity confirmation screen
MARCO — AGENT
"Alright. I've got what I need. Give me a few minutes."
PositionAttacking midfielder
LeagueLigue 1
StyleTechnical
StageEstablishing myself
This is your career. It begins now.
The confirmation moment

The identity card appears in the style of the world — not a settings summary, but a dossier. No "confirm" button. The statement "This is your career. It begins now." does the work of a button while staying inside the narrative register.

This screen is the product's first emotional beat before the first gameplay beat. For The Dreamer persona especially, seeing their identity documented triggers the "this is real" response that everything downstream depends on.

The 4-decision constraint — why not more
Cognitive load ceiling
Research on mobile engagement shows decision fatigue sets in after 4 choices in the first session. The 5th decision is where users start to feel like they're doing work rather than living a life. Four is the ceiling regardless of how much identity richness the system could support.
Derivable attributes
Dozens of narrative attributes can be derived from 4 choices. A "Technical / Ligue 1 / Establishing" player generates a completely different press voice, agent dynamic, and first beat than a "Clinical / Premier League / Prime" player. Depth is in the system, not the input.
Editable within the first week
Allow one revision of any identity attribute within the first 7 days, framed narratively: "Your agent suggests repositioning how he's pitching you." This removes the anxiety of permanent commitment without cheapening the initial choice.
Phase 2 of 5
Part V · Onboarding Architecture
03
Phase 3 · 3:00 – 5:30
The First Beat
The best content in the product, delivered first. The first beat must be so emotionally resonant that the user's mental model of this product upgrades from "app I downloaded" to "world I'm in" — permanently.
"Most products save their best for later. Appified games have no later if the first beat fails."
The first beat template
Design requirements for the first beat
It must carry genuine stakes. Not simulated stakes. The agent's call, referenced in the Drop-In, must now arrive with real content — a transfer inquiry, a contract decision, a press situation. Something the user actually cares about resolving.
It must use the identity choices. If the user chose "Attacking midfielder / Ligue 1 / Establishing myself," the first beat must reference those specifics. A generic first beat after a personalized identity phase breaks the promise of the first two minutes.
The decision must be genuinely hard. Two options, both compelling. The user's first real choice in the product must feel like a choice they'll remember making — not a tutorial decision with an obvious answer.
The consequence lands immediately. In the first beat, don't defer the consequence to the next day. Show the immediate ripple — the world shifts in real time — so the user understands the product's core loop from a single interaction.
A worked first beat — Ligue 1 establishing player
MARCO — AGENT · 06:14
It's Lyon. They've been watching you for six weeks. They want you in January. It's not confirmed — but it's serious. Your current club doesn't know yet. How do you want to handle this?
Your response
"Find out what they're offering. Keep it quiet for now."
"Tell my club. I want to be honest about where I am."
Why this first beat works

The stakes are real: Lyon vs. current club is a genuine career inflection. It references a real French club, which immediately signals authenticity to Ligue 1 fans.

Both choices have weight: Keeping it quiet preserves optionality but risks loyalty. Being honest builds trust but closes the negotiation before it begins. A Dreamer persona will agonize. An Obsessive will immediately understand the implications.

The consequence arrives in 2 seconds: After the choice, the agent responds immediately — planting the next thread before the user has processed the first one.

Immediate consequence delivery
MARCO — AGENT · 06:16
Smart. I'll set up a call with their technical director for next week. In the meantime — there's a press request waiting. Le Monde wants 10 minutes before training. They've heard a rumor. What do you want to say?
The consequence plants the next beat
The agent's response resolves the immediate choice (smart — we're moving forward) and immediately plants the next decision (press request) before the user has left the screen. This is the "one more" mechanic — the session continues not because the product demands it, but because the world demands attention. The press question is Phase 4's entry point.
Phase 3 of 5
Part V · Onboarding Architecture
04
Phase 4 · 5:30 – 8:00
The World Opens
A second and third voice arrive. The career now has texture beyond the agent. The user understands that multiple relationships constitute this world — and that each one has its own agenda.
"One voice is a product. Three voices are a world."
Introducing the second voice — the press
PRESS DIGEST · 07:02
Le Monde Sport — 'One of Ligue 1's most technically gifted midfielders is being linked with an ambitious winter move. Sources say the player's form this autumn has attracted attention from several top clubs.'
The story is out. It wasn't you. Marco is calling.
Why the press arrives here

The press voice serves two functions in Phase 4. First, it provides external validation — the user sees themselves being written about, which deepens the identity investment. Second, it creates a new layer of consequence: the agent call that follows is now about damage control, not opportunity.

Using Le Monde specifically — a real, named, prestigious publication — signals authenticity immediately to the French market. The specificity is the credibility. "A sports outlet reported" would feel like a game. "Le Monde Sport" feels like the world.

Introducing the third voice — the physio or manager
Branch A — Physio arrives
If the user's first beat involved any mention of physical demands (intensive training, pushing through), the physio appears with a gentle check-in. Introduces the body as a narrative variable. Establishes that the career has physical limits — not just social and strategic ones.
Branch B — Manager arrives
If the press story is now public, the manager appears. His tone depends on club loyalty state — warm if the user has been transparent with the agent, cooler if they haven't. The manager's first appearance establishes his authority without being antagonistic.
The inbox view — first full look at the product
Your inbox · 07:14
MARCO — AGENT · unread
Call me. The press got ahead of us. We need to coordinate.
PRESS DIGEST · read
Le Monde Sport · transfer link story live
COACH BERTRAND · unread
Team meeting tomorrow at 9. I'll want to talk to you beforehand.
The inbox as a world map

The first full inbox view — three unread threads — gives the user a spatial understanding of their world. They see the agent, the press, and the manager in one view. They understand instinctively that each relationship is a separate narrative thread that they manage.

This view should arrive at approximately the 7-minute mark — late enough that the user is invested, early enough that the session doesn't end before they've seen it.

The coach's message ("I'll want to talk to you beforehand") is intentionally unresolved. It is Phase 5's entry point.

Phase 4 of 5
Part V · Onboarding Architecture
05
Phase 5 · 8:00 – 10:00
The Open Thread
Before the session ends, something must be left unresolved — a genuine thread in the user's world that lives in their mind until they return. This is the mechanism that converts a first session into a D1 return.
"The session doesn't end when the user closes the app. It ends when the open thread closes — which it never fully does."
The open thread mechanism
How it works
A new stimulus arrives that cannot be immediately resolved. The coach's meeting is tomorrow. The agent needs 48 hours to hear back from Lyon. The press request window closes at noon — and the user sees it with 20 minutes remaining. One of these threads is planted deliberately to require time to pass before it can close.
The app suggests — doesn't demand — that the user return. No "Come back tomorrow!" push prompt. Instead, the agent's last message in the session ends with something like: "I'll have the numbers from Lyon by morning. Check in then." The pull is in the story, not in the product's hand.
The D1 notification references the specific open thread. Not "New content available." Not "Your career needs attention." Specifically: "Lyon's offer just came in. Marco needs your answer before training." The user knows what this means because they left the thread open.
The push notification — D1 morning
FOOTBALLER LIFE · NOW
Lyon's offer just came in.
Marco needs your answer before training. The club is waiting too.
Why this notification converts
The user already knows who Marco is. They already know about Lyon. They already made the choice to keep it quiet. This notification lands with narrative weight because the user has context — it is not an anonymous prompt from a product they downloaded yesterday.
"The club is waiting too" adds a second layer of tension — the coach's meeting is now clearly connected to the transfer situation. Two threads are converging, and only the user can resolve them.
This notification is not a retention mechanic. It is the next beat in the story. The distinction matters — users who feel manipulated by notifications churn. Users who feel that something genuinely happened in their world return.
The session-end screen
Your morning begins here.
Lyon's offer arrives in the morning.
The coach is waiting.
The press is already writing.
🔔
NOTIFICATION · TOMORROW 07:00
Marco has news from Lyon.
Set for tomorrow morning.
Your career continues tomorrow.
The session-end screen design

No "Rate us" prompt. No "Share with friends" CTA. No feature highlights. The session ends in narrative — a summary of the three open threads, and the time the next beat will arrive.

The notification permission request is embedded here, framed as "Set for tomorrow morning" — not as a product asking for access but as a career that has a scheduled development the user would want to know about. Permission conversion should be above 70% in this framing.

The phrase "Your career continues tomorrow" is the only product-level language in the entire onboarding sequence. It earns this moment because everything before it was lived, not described.

Phase 5 of 5
Part V · Onboarding Architecture
Design Rules
What Never to Do
The onboarding failures that look like reasonable product decisions but actively destroy the fantasy. Each one has been seen in adjacent products and each one produces measurable D1 churn.
Tutorial overlays or "tap here" prompts
The moment a hand appears pointing at a UI element, the product is a game. The user's brain recategorizes it and the fantasy collapses permanently. Every mechanic must be learned through participation, not instruction. If a mechanic requires instruction, redesign the mechanic.
A "how to play" or "how this works" screen
If this product needs to explain itself before the first beat, the first beat is wrong. The product explains itself by being experienced. A first-time user should understand what this is by the end of Phase 3 — not by reading a modal at the start of Phase 1.
Generic first beat content
A first beat that could be sent to any user — regardless of their identity choices — is a first beat that will fail. "Your agent has a message for you" is generic. "Lyon have been watching you for six weeks" is specific. The specificity must be generated from the identity choices made in Phase 2.
Asking for notification permission before Phase 5
Notification permission requested before the user cares about their career converts at under 20%. Requested at the session-end screen after 10 minutes of narrative investment, framed as "your next career beat is scheduled," converts at 60–75%. The timing is the permission strategy.
A paywall or subscription prompt in the first session
The first session must be entirely free. No "unlock full career" prompt. No premium badge on the best choices. The user must believe unreservedly in the world before they're asked to pay for it. The subscription conversation begins in Session 2, framed as "continue your career uninterrupted."
Resolving all tension before session end
A session that ends with every thread closed gives the user no reason to return. The most common onboarding failure is a satisfying first session — because a satisfying session is a complete session, and a complete session has nothing pulling the user back. Leave one thread genuinely open.
Showing stats, meters, or progression bars in the first session
Form: 72. Reputation: 45. Club Loyalty: 60. These numbers — or any visible progression system — signal "game" immediately to the Insider persona and "homework" to the Escapist. All state variables operate invisibly in the background. The first session has zero visible gamification. None.
Design Rules
Part V · Onboarding Architecture
V
Sport Variants
Adapting the Architecture
The five-phase structure is universal. What changes across sports is the specific voice roster, the identity commitment decisions, the drop-in atmospheric texture, and the first beat's emotional register.
What stays constant across all sports
Five-phase structure
4-decision identity limit
Agent-first voice
Open thread at session end
No tutorial overlays
No paywall in session 1
D1 push references specific thread
Sport-specific adaptations
Sport Drop-In texture Identity decisions First beat type Open thread
Football ✓ Pre-dawn, stadium lights, agent message Position, league, style, career stage Transfer approach Club's response pending
F1 Driver Paddock morning, engineers on radios, manager call Team tier, driving style, career ambition, contract status Seat negotiation / team politics Team principal meeting tomorrow
MMA Fighter Gym morning, wrapping hands, coach texts Fighting style, weight class, camp type, ranking tier Fight offer — opponent and timeline Opponent's camp response pending
Basketball Shoot-around, arena empty, front office call Position, conference, role (starter/sixth), contract year Trade rumor / contract extension offer GM meeting scheduled
The Insider beta test — sport-specific validation

Before any new sport onboards its first users: Run the complete 10-minute onboarding sequence past 5–10 Insider-profile users from that sport's community. Their single most important question: "Does this feel like something that could actually happen in my sport?" If more than 2 of 10 say no — the identity decisions or first beat have authenticity gaps that will destroy D1 retention before UA spend matters.

Sport Variants
Part V · Onboarding Architecture
%
Success Metrics
Measuring the First 10 Minutes
The specific signals that tell you whether the onboarding architecture is working — and the diagnostic questions that identify which phase is failing when D1 retention drops.
Primary KPIs — onboarding health
≥ 85%
Phase 2 completion
Users completing all 4 identity decisions
≥ 70%
First beat completion
Users making a decision in Phase 3
≥ 65%
Notification opt-in
At Phase 5 session-end screen
≥ 55%
D1 retention
Return within 24 hours of first session
≤ 15%
Phase 1 drop-off
Users who leave before identity phase
≥ 8 min
Median session length
For users who reach Phase 3
Diagnostic — which phase is failing
P1
Drop-off > 25% in Phase 1
The atmospheric opener is too slow or too vague. Shorten the scene-setting to 1 sentence. Make the agent's first message more specific and more urgent. The user must feel the pull within 20 seconds.
P2
Identity completion below 80%
The decisions are too complex, too many, or feel like a form. Reduce to 3 decisions. Rewrite option copy in first-person personality language. Ensure each decision arrives as a question from the agent, not a settings panel.
P3
First beat decision rate below 60%
The first beat has a false choice — one option is obviously correct. Rewrite both options so each has something worth losing. If the beat is getting read but not decided, the stakes feel abstract. Make them more personal and immediate.
P4
Session exit before inbox view
The world is opening too slowly. Phase 4 content is arriving too late or the second voice is arriving before the first beat's consequence is resolved. Tighten the timing — the press voice should arrive no later than 6 minutes in.
P5
D1 return rate below 45%
The open thread is either too weak (the user doesn't care enough about what's unresolved) or fully resolved (nothing was left open). Audit the session-end state: is there a genuinely compelling thread with a specific arrival time? If the open thread is vague, the D1 notification will feel like spam.
23 of 23 · End
2
Part VI · Content Scaling
Why Volume Isn't Enough
The answer to "how do 2 people scale narrative?" is not "write more content." It's build a system that turns one human creative decision into 15–20 authentic narrative moments — then let the world and the community do the rest.
7 chapters 3 levers · 1 production system

The real problem: AI without architecture produces volume without coherence. 500 beats written by unconstrained AI feel like they were written by 5 different people who never read each other's work. The product starts feeling thin not because there's less content, but because the content loses the specific texture of authenticity that makes users believe in the world. The solution is architectural first, generative second.

The three-lever model
Lever 1 — The multiplier
One human scenario brief generates 15–20 distinct narrative moments through structured branching, state variations, and consequence chains. Leverage ratio is in the architecture.
Lever 2 — The living world
The real football calendar generates free content triggers every 48–72 hours. Wire the narrative system to the calendar and the world writes the drama for you.
Lever 3 — Community as co-author
The Discord community surfaces content gaps before they become churn signals — and sometimes fills them. A free content intelligence system that runs itself.
Overview
M
Part VI · Content Scaling
The problem quantified
The Content Math
What 90-day retention actually demands in terms of narrative depth — and why it's achievable for 2 people with the right system.
"The goal is not to write more. The goal is to make the user feel like there is always more."
What 90-day retention requires
The depth target
A user retained to D90 opens the app ~600 times across 90 days (roughly 6–7 opens per day). Each open delivers 1 beat. That's 600 beat exposures before the product needs to feel truly exhausted.
With 3 variants per beat and 40% branch divergence, a library of 300 base beats generates approximately 900 unique experiences. Users hit true repetition at around exposure 600 — right at D90.
At 30 new beats per month (the sustainable 2-person output using the system in this document), the library grows by 90 per quarter — staying 50% ahead of the D90 exposure ceiling permanently.
The 2-person content capacity
Weekly time allocation (2 people, ~20h content time/week)
3h
4h
5h
4h
2h
2h
Scenario briefs (3h)
AI generation + review (4h)
Calendar event beats (5h)
Arc design + wiring (4h)
Quality audit (2h)
Community reading (2h)
Output per week at this allocation
ActivityTimeRaw outputMultiplied output
Scenario briefs3h4 briefs48–60 beats
Calendar event beats5hDirect write12–18 beats
Arc design4h1 arc designedStructure for 20+ beats
Community gap reads2h5–10 signals3–5 brief ideas
Total weekly beat output~75–90 beats

The monthly compound: 75–90 beats per week = 300–360 beats per month. The D90 depth target (300 base beats) is achievable in a single month of sustained production. After month 1, every week of output is extending the depth runway further.

The Math
01
Lever 1
The content multiplier
One Brief → Many Beats
A single 5-minute scenario brief, processed through the multiplier system, generates 15–20 distinct narrative moments that feel specifically written — not templated. The leverage is in the structure, not the writing speed.
"A brief is a creative decision. A beat is the execution of that decision. Only humans make the creative decisions. AI handles all the execution."
What a scenario brief contains
Brief template — 5 minutes to complete
Situation: What has just happened in the career world? (1 sentence. Real or derived from calendar.)
Voice: Which character delivers this? What is their emotional state right now?
Stakes: What does the user stand to gain and lose? Name both sides explicitly.
State condition: Which state variables trigger this brief? (e.g. "Form > 70 AND transfer window open")
Authentic detail: One real-world specific — a club name, a medical term, a contract clause — that makes it feel lived-in rather than invented.
How one brief multiplies
Multiplication layerHow it worksBeats generated
Base beatAI generates 3 variants from the brief3
Decision branches2–3 options per beat × consequence beats+6
State variationsSame brief, different Form/Loyalty values → different character tone+4
Consequence chainShort-term and long-term ripple beats auto-generated+4
Voice recastsSame event, different character delivers it (press vs agent)+2
Total from 1 brief (5 min human time)~19 beats
State variations — the biggest multiplier most studios miss
Same brief, Form 85, Loyalty 70
Agent tone: Confident, pressing. "Three clubs. The number you want is there. I think we move."
Press voice: Speculative, excited. "One of Ligue 1's most in-form midfielders linked with ambitious winter exit."
Same brief, Form 48, Loyalty 85
Agent tone: Cautious, protective. "I've heard something but the timing is terrible. Your form right now won't get you the deal you deserve."
Press voice: Uncertain, probing. "Surprise link for struggling midfielder — insider questions whether a move makes sense."
1 brief = ~19 beats
4 briefs/week = ~75 beats
Human time per brief: 5 min
AI generation time: ~2 min
Review time per beat: ~3 min
01 of 07
02
Lever 2
The living world
The Calendar Writes the Drama
The real football calendar generates high-stakes, emotionally resonant content triggers on a predictable schedule — for free. A 2-person team that treats the calendar as a content system, not a live ops checklist, gets more narrative richness than a 20-person team inventing fiction.
"Transfer deadline day doesn't need a writer. It needs someone to wire the system to the clock."
The calendar as a content system
How calendar integration works in practice
Pre-written beat libraries per event type. For each predictable calendar event — transfer deadline, international break, cup draw, World Cup — maintain a library of 8–12 pre-written beats that can deploy immediately when the event fires. Zero reactive writing required.
The 48-hour rapid response protocol. For unpredictable events (shock signing, injury to a star player, managerial sacking), a single brief + AI generation produces a reactive beat within 2 hours. The voice spec and beat template mean structure is never invented under pressure.
Event intensity tiers. Not all calendar events are equal. Tier 1 (World Cup, transfer deadline, national team call-up) get full arc treatment. Tier 2 (match weeks, international breaks) get pre-written beat battery activation. Tier 3 (league cup rounds, mid-table matches) get passive state variable updates.
Annual calendar mapped to content intensity
Always on
Match day beats (×38) Weekly form beats Training beats
Tier 2
Intl breaks × 4/yr Cup rounds × 6/yr Champions League groups
Tier 1
Jan window opens · 3 weeks June window opens · 3 months Deadline day × 2/yr
Peak
World Cup · 6 week arc Euros · 4 week arc 10× organic UA window
The 5-minute rapid response brief
Template — unpredicted world event
Event: {REAL_WORLD_EVENT} just happened.
How it enters the career world: Through {CHARACTER} — because they own this beat type.
Emotional register: {TONE} — the character's reaction, not a neutral report.
User implication: What does this mean for the user's career specifically? Connect to their current {STATE_VARIABLE}.
Authentic detail: One real name, real number, or real terminology from the actual event.
{REAL_WORLD_EVENT} {CHARACTER} {TONE} {STATE_VARIABLE}
02 of 07
03
Lever 3
Community as co-author
The Discord Intelligence Layer
The community is not a feedback channel. It is a content intelligence system that runs itself — surfacing gaps before they become churn, validating authenticity in real time, and occasionally generating the product's best content ideas unprompted.
"A user who posts 'I can't believe my agent just did this' has written the brief for the next beat without knowing it."
What the community tells you for free
Gap signals
When users start asking "what happens if I..." — that's a content gap. The specific question is the brief. Track these in a simple running list. Every week, the top 3 questions become brief candidates.
Authenticity flags
When an Insider-profile user says "that's not how contracts work" or "no agent would say that" — that's a voice quality failure caught before it becomes an App Store review. Community catches what internal review misses.
Organic brief ideas
Users describing their career moments — "I declined Lyon and now my agent won't talk to me" — are narrating consequence chains the system hasn't built yet. These are some of the best scenario briefs available because they're generated by people living the product.
Depth horizon warnings
When top-engaged users (D30+) start saying "I feel like the same things keep happening" — that's the D30 cliff approaching. It arrives in community before it appears in analytics. Acting on it with 2 weeks of buffer prevents the churn spike entirely.
The 2-hour weekly community read
Monday morning · 2 hours · one person
30 min — scan Discord for question patterns. What are users asking about that the product hasn't addressed? Log 5–10 questions. Circle the 2 most common.
30 min — read career sharing posts. What moments are users describing that surprised them, delighted them, or frustrated them? Each surprise is a brief candidate. Each frustration is a quality flag.
30 min — note Insider corrections. Any factual pushback from domain-expert users gets logged immediately and actioned in the next content sprint. Don't wait for a review cycle.
30 min — write 3 briefs from community signals. Convert the strongest observations directly into scenario briefs before the week starts. These go into Wednesday's AI generation session.
03 of 07
04
The production system
The weekly engine
The Weekly Pipeline
A precise weekly production system that two people can run without collision, without creative fatigue, and without letting AI erode the authentic voice that makes the product work.
The week, day by day
Monday · Person A · 2h
Community intelligence read
Scan Discord. Log gap signals, authenticity flags, career moment descriptions. Write 3 briefs directly from community observations. This is the creative anchor for the week — all other content follows from what users are experiencing.
Output: 3 community-derived briefs · 5–10 gap signals logged
Tuesday · Person B · 3h
Scenario brief writing
Write 4 scenario briefs from the week's calendar events and state-variable gaps. Each brief takes 5 minutes. The remaining time is spent on arc design — mapping the next 4–6 week narrative structure so AI generation has a coherent container to work within.
Output: 4 scenario briefs · 1 arc structure mapped
Wednesday · Person A · 4h
AI generation session
Run all 7 briefs (3 community + 4 scenario) through the AI generation pipeline. Each brief produces 3 variants. Review each variant for voice consistency — 3 minutes per beat. Flag any that drift from character spec for quick rewrite. Approve and tag all passing beats for state wiring.
Output: ~40–50 approved beats · flagged list for rewrites
Thursday · Person B · 3h
Calendar beats + state wiring
Write calendar event beats for the coming week's fixtures and world events. These are direct writes — no AI generation — because they reference real outcomes and require specific voice. In parallel, Person B wires Wednesday's approved beats into the content database with correct state-variable triggers.
Output: 10–15 calendar beats · Wednesday beats wired and live
Friday · Both · 2h
Quality audit + voice check
Both team members read 20 beats each from the week's production — out of order, without context. Does each beat sound like the character? Does the decision feel genuinely difficult? Does the consequence land in story before numbers? Anything that fails either question goes back for a single revision note, not a full rewrite.
Output: Weekly batch QA'd · voice drift caught before live
The two-person split — who owns what
Person A — The World Watcher
Community intelligence reads every Monday
AI generation sessions — produces volume
Voice character owners: Agent + Press
Real-world event monitoring + rapid response
Person B — The Architect
Scenario brief writing — the creative decisions
Arc design — the long-form narrative structure
Voice character owners: Manager + Physio
State variable wiring + content database
04 of 07
05
The production system
AI generation
Prompt Templates
The exact prompt structures that produce production-ready beats reliably. Each template is designed so the AI cannot stray outside the voice spec, even across hundreds of consecutive generations.
"The prompt is a fence, not a script. It defines the boundary of the character's world. The AI fills the space inside it."
Master generation prompt
Beat generation — master template
You are writing narrative content for Footballer Life Simulator — a product where users live the daily life of a professional footballer. Your output is delivered directly to users as messages, reports, and character communications. It must feel like it was written by a person with lived experience of professional football, not by someone who has read about it.

Character: {CHARACTER_NAME} — {CHARACTER_REGISTER} — never breaks this voice, not even slightly.
Situation brief: {BRIEF}
Player state: Form {FORM} · Fitness {FITNESS} · Club Loyalty {LOYALTY} · Career Stage {STAGE}
Beat type: {TYPE} (tension / reward / world)

Produce 3 variants. Each variant contains: (1) stimulus message — max 3 sentences in character voice, (2) 2–3 decision options — max 12 words each, both genuinely difficult, (3) consequence for each option — 1–2 sentences felt as story, not statistics.

Hard rules: Never use the word "game." Never show stat numbers. Every decision must have something worth losing on each side. Write as {CHARACTER_NAME} — not as a narrator describing {CHARACTER_NAME}.
{CHARACTER_NAME} {CHARACTER_REGISTER} {BRIEF} {FORM} {FITNESS} {LOYALTY} {STAGE} {TYPE}
Character-specific add-ons
Agent (Marco) add-on
Marco is direct and conspiratorial. Short sentences. He always has an angle — his commission interests are present but never stated. He uses "we" for career decisions that benefit him and "you" for risks. He never comforts without a reason. His conflict of interest should be subtly present — never explicit.
Press Digest add-on
Third person. Newspaper headline register. Always references one real, named publication. Uses actual ratings (L'Équipe scores, ratings out of 10). Never invents a quote — reports what "sources say" or what "observers noted." Should feel like reading a real match report.
Physio (Dr. Roux) add-on
Clinical but warm. Uses one real medical term per message — Grade 1/2 strain, adductor, metatarsal. Never panics and never falsely reassures. The decision should always force a choice between the body and the moment. He has been right before. The user knows it.
Manager (Coach Bertrand) add-on
Formal and measured. Speaks in team terms — "we need," "the squad requires." Praises rarely and only when it means something. When warm, it should feel significant. He is authority — not antagonist. The user respects him even when in conflict.
The variant selection rule
How to choose between the 3 AI variants
Read variant 1 aloud. Does it sound like a real message from this character? If yes — approve. If not — move to variant 2.
Check the decision balance. If both options clearly favour one path, the beat has a false choice. Reject regardless of how well the voice works.
If none of the 3 variants pass: Add one line of context to the brief ("the agent is specifically under time pressure here") and regenerate once. If the second set also fails — the brief itself is the problem, not the generation. Rewrite the brief.
Never ship a beat that required more than one revision note. If you're editing heavily, the prompt architecture isn't doing its job. Fix the system, not the output.
05 of 07
06
The production system
The authenticity problem
Staying Authentic at Scale
The hardest thing about scaling narrative with AI is that authenticity degrades slowly — not in a single bad beat, but through dozens of small compromises that compound over weeks until the product sounds like it was written by someone who studied football rather than lived it.
"Authenticity doesn't break in one beat. It erodes in fifty."
The three authenticity failure modes
Voice drift — the character slowly stops sounding like themselves
Happens gradually over 50–100 AI generations. The agent starts using slightly more formal language. The physio starts over-explaining. The press voice starts editorialising. Each individual beat passes the voice check — but read 20 in sequence and the drift is audible. Detection: monthly read of 20 consecutive beats per character out of context. Fix: reanchor the prompt with 3 "canonical examples" drawn from the best-performing early beats.
Scenario repetition — the same emotional shape recycled with different nouns
The transfer rumor beat that felt fresh in week 2 starts feeling mechanical in week 8 because the structure is identical — just with a different club name. The solution is brief diversity, not variant diversity. Each new batch of briefs must introduce a new emotional situation, not re-skin a previous one. Monthly brief audit: flag any brief that shares the same stakes structure as an existing beat.
Domain accuracy slip — details that real football professionals would notice
AI will occasionally invent a plausible-sounding but inaccurate piece of football detail. A contract clause that doesn't exist. An injury timeline that doesn't match real recovery data. A transfer process that skips a step. These are invisible to the casual user but immediately obvious to the Insider persona — who will post about it publicly. Fix: every beat that contains a specific factual claim (medical, legal, tactical) gets one Insider review before shipping.
The monthly authenticity audit
First Friday of every month · 2 hours · both people
Voice audit: Read 20 random beats per character (4 characters = 80 beats total, 40 per person). Flag any where you notice register drift. Update the character spec if the drift exposed a gap in the original documentation.
Structure audit: List the emotional shapes of the last 30 briefs. Group by similarity. Any cluster of 4+ similar structures gets a creative block — no more briefs in that shape for 4 weeks.
Accuracy spot-check: Pull 10 beats containing specific factual claims and verify each one against real football knowledge. Log any errors found. If 2+ errors found in a month — add Insider review to the generation pipeline for beats with factual claims.
CEO domain review: The CEO's lived footballer experience is the irreplaceable quality gate. 30 minutes of his time once per month reading 15 beats is worth more than any other quality mechanism. Block it. Protect it. Never deprioritise it for shipping velocity.
06 of 07
07
The production system
The long game
The Depth Runway
How to think about content depth as a compound asset — one that grows faster than users can consume it when the system is running correctly, and that creates a permanent competitive moat when it's been running for 6+ months.
Depth as a moat
Why narrative depth compounds
A copycat studio launching 6 months after Tapugo doesn't just need to match the current beat count — they need to match the current beat count plus 6 months of compound production. At 300+ beats per month, the gap widens by 300 beats every month they wait.
More importantly, they cannot replicate the voice. The CEO's domain expertise, the community's co-authorship, and the authenticity earned through Insider validation — these are not purchasable at any production speed. Volume can be copied. Authenticity cannot.
At month 6 of running the system: ~1,800 base beats. With 3 variants each = 5,400 unique narrative moments. Users would need to engage at 10× daily opens for 180 days before hitting meaningful repetition. The depth runway is essentially permanent.
Depth runway projection
MonthCumulative beatsUnique experiencesD90 exposure ceilingRunway status
Launch30090060050% buffer
Month 26001,8006003× buffer
Month 39002,7006004.5× buffer
Month 61,8005,4006009× buffer
Month 123,60010,80060018× buffer — moat
The one metric that tells you the system is working

Track this weekly: "Percentage of D30 users who have seen more than 60% of available beats." If this number is rising above 40%, the system is not producing fast enough and users are approaching the repetition ceiling. If it stays below 30%, the system is healthy. Below 20% means users are discovering new content faster than they're exhausting it — that's the ideal state. Run this report every Monday morning alongside the community read.

Target: <30% beat exposure at D30
Monthly output: 300+ beats
Month 6 runway: 9× buffer
Voice audit: monthly
CEO review: monthly
Community read: weekly
07 of 07 · End