Why Are We So Bad at Killing Games?
Here’s a scene you’ve probably witnessed: a game, months into soft launch, is missing its targets.
The team presents a plan:
The plan is full of hope. It talks about a future feature that will fix retention, a new offer that will solve monetization, or a new user acquisition strategy that will find the “right” players.
The meeting ends with a decision to extend the soft launch for “just one more month.”
Three months later, the game is finally killed, having burned hundreds of thousands of dollars in team salaries and UA spend for no additional insight.
This is the slow, expensive death that plagues our industry. The problem is rarely the game or the team; it’s the process. We are emotionally invested, so we optimize for hope. We ask, “What could this game become if everything goes right?” instead of, “What does the data tell us right now?”
This handbook offers a different, more disciplined approach: Red Light by Default. It’s a system built on a simple premise: assume every project is a failure until proven otherwise. The goal is not to find reasons to continue; it is to demand that the data provide overwhelming evidence not to kill.
Part I: The Mindset
Chapter 1: Understand Player Motivation
“What is the most exciting, aspirational part of our game, and how can we show it to the player in the first ten minutes?”
The Problem:
“Our playtesters love the game and retention looks okay, but our early conversion numbers are terrible. We think players need more time to see the value before they buy.”
The Fix: Stop Thinking Rationally. Conversion is Aspirational.
You believe players spend money as a rational transaction. This is incorrect. For your most valuable players, monetization is an emotional, aspirational act. They aren’t buying a solution to a past problem; they are buying into a future fantasy.
What the Data Says:
The timing of the first purchase is the single strongest predictor of a player’s lifetime value. The vast majority of first purchases happen almost immediately.
| Time to First Purchase | % of All Paying Users |
|---|---|
| First Hour | 15% |
| First Day (Day 0) | 25% |
| Day 1 | 10% |
| Day 2 | 5% |
| Cumulative by Day 2 | 40% |
Checklist:
- Does your First Time User Experience (FTUE) immediately showcase what is cool and aspirational?
- Is your first IAP offer framed around a compelling future state (“Become a Dragon Rider”)?
- Are you treating early spenders as curious explorers, not loyal customers who need convincing?
Chapter 2: Understand Your Business Model
“What systems are we building specifically to provide deep, long-term value to our top 5% of players?”
The Problem:
“We have a decent number of payers and our conversion rate is average, but our revenue is still flat. We’ve made our starter packs cheaper to attract more buyers, but it’s not working.”
The Fix: Stop Designing for Everyone. Your Business is Your Fans.
Your business model is not a volume play; it is a quality play. Designing for the “average player” is a recipe for mediocrity because the average player does not meaningfully contribute to revenue.
What the Data Says:
Revenue in Free-to-Play games is hyper-concentrated. The top percentile of your paying users will account for a hugely disproportionate amount of your revenue.
| Payer Percentile | % of Total Revenue |
|---|---|
| Bottom 50% | 3% |
| 50-75% | 6% |
| 75-90% | 12% |
| 90-95% | 13% |
| 95-99% | 30% |
| Top 1% (99+) | 36% |
Checklist:
- Have you stopped averaging your player data and started segmenting by spend depth?
- Is your monetization strategy designed to serve your most engaged fans?
- Are you prepared to build deep, expensive systems (like VIP tiers)?
- Do your price points reflect premium positioning, not mass-market accessibility?
Chapter 3: Design for Monetization and Engagement
“How does our monetization system make our core loop more engaging and rewarding for the players who choose to spend?”
The Problem:
“We’ve built a great core loop with fantastic early retention, and now we’re bolting on monetization. Players are calling it a ‘cash grab’.”
The Fix: Monetization is a Core Feature, Not a Layer.
If spending money feels like a tax you have to pay to keep playing, you’ve already failed. It should feel like an enhancement that makes the game more fun.
What the Data Says:
Successful games monetize from day one. Waiting to test monetization until after you’ve “fixed” retention is a fatal error.
| Day After Install | Cumulative % of D360 LTV Realized |
|---|---|
| Day 1 | 10% |
| Day 7 | 20% |
| Day 30 | 35% |
| Day 90 | 60% |
Checklist:
- Did your initial concept document include a detailed, written monetization strategy?
- Does spending money in your game feel like a strategic choice, not a punishment?
- Are you testing both engagement and monetization KPIs from the very first build?
Chapter 4: Set Ambitious Targets
“What does ‘great’ look like in today’s market, and have we set that as our minimum standard for success?”
The Problem:
“Our gate meetings are endless debates. The data is always ‘almost there,’ and we’re not sure if we should kill the project or pivot again.”
The Fix: Make the Decision Obvious. Set Brutally High Targets.
Your KPI targets are not goals; they are kill criteria. The gray area is where studios go to die. Eliminate it.
Simulated Scenario: The Proximity Trap
Imagine two studios. Studio A sets a “realistic” D7 retention target of 18%. Studio B sets an “ambitious” target of 25%. Both games hit 16% D7 retention.
- Studio A’s Decision: “We’re only 2 points off! Let’s do a two-month art polish sprint.”
- Studio B’s Decision: “We are 9 points off. This core isn’t working. Kill it.”
Outcome: Studio A spends an extra $400k+ to confirm what Studio B learned in a month.
Checklist:
- Are your KPI targets based on top-quartile performance for the current market?
- Has the entire team agreed that anything less is a failure?
- Is your go/no-go decision based on hitting targets, not a subjective “feeling”?
Chapter 5: Assume Failure by Default
“Does the data show, with high confidence, that this game will be a hit, or are we operating on hope and sunk costs?”
The Problem:
“We’ve been in soft launch for six months. We’ve pivoted twice. We all feel like it could work, but we’re burning cash… Should we kill it?”
The Fix: Flip the Question. Prove a Game is Worthy of Investment.
The question is never “Why should we kill this game?” It is “Has this game given us sufficient data to prove we shouldn’t kill it?”
Simulated Scenario: The Risk Matrix
You have two choices.
- Choice A: Extend Investment. Spend another $200k. You have a 90% chance of lighting your $200k on fire.
- Choice B: Kill the Project. You save the $200k. There is an extremely low probability (<1%) that you just killed a future hit.
The Rational Decision: The financial risk of continuing is mathematically far greater than the opportunity cost risk of killing a potential winner.
Checklist:
- At every gate meeting, is the default decision “kill”?
- If the data is ambiguous, is the decision to kill final?
- Is killing a project treated as a successful outcome of a rigorous process, not a team failure?
Part II: The Process
Chapter 6: Ditch Lagging Indicators for Universal Truths
“Is our game’s design fighting against universal truths or capitalizing on them?”
The Problem:
“We keep trying to use D7 retention to predict D30 LTV, but our models are always wrong… by the time we have a clear signal, we’ve wasted weeks.”
The Fix: Ground Your Strategy in Reality, Not Flawed Correlations.
Player activity is not random; it follows predictable life patterns. Your design must accommodate these realities instead of fighting them.
What the Data Says (The Universal Truths):
| Cycle | Key Pattern | Implication |
|---|---|---|
| Daily | Peaks at commute times, lunch, and evening couch time. | Design for short, interruptible sessions (3-5 minutes). |
| Weekly | Activity and spending peaks on Saturday and Sunday. | Schedule your most compelling events & sales for weekends. |
| Platform | Mobile is a secondary activity. | Game must be simple to pick up and put down. UI must work with one hand. |
Checklist:
- Does your core loop deliver a satisfying experience in 3-5 minutes?
- Are your appointment mechanics timed to match daily peaks?
- Is your event and sales strategy built to maximize the weekend surge?
Chapter 7: Find Your North Star
“What is the single player action that proves our game’s core fantasy is working and that players are deeply engaged?”
The Problem:
“Our team is chasing too many metrics… We don’t know what’s most important, and we’re making shallow optimizations everywhere.”
The Fix: Find the One Metric That Matters. Align Everyone.
A true North Star is the single, game-specific player action that best quantifies the value your players get from your core loop. It’s the “magic moment” of your game, turned into a number.
What the Data Says (Finding the Right Action):
| Game Genre | North Star Metric Example | Rationale |
|---|---|---|
| Match-3 | Levels Completed Per Session | Measures core loop mastery and correlates to booster spend. |
| RPG | Heroes Upgraded to Max Level Per Week | Measures long-term investment and predicts high LTV. |
| Strategy | Battles Initiated Per Day | Measures competitive engagement and resource consumption. |
Checklist:
- Have you identified the single most important action in your core loop?
- Does improving this action demonstrably correlate with long-term retention?
- Is your entire team aligned on improving this single metric above all others?
Chapter 8: Decide in Time-Bounded Phases
“What is the absolute deadline for our final kill/scale decision, and what is the minimum data we need to make it?”
The Problem:
“Our soft launch is dragging on with no clear end date… We say we’ll make a decision ‘when we have enough data,’ but that day never seems to come.”
The Fix: Use Rigid, Time-Bounded Gates to Force a Decision.
A soft launch is not open-ended research. Waiting for D90 or D180 data is a colossal waste of time and money. Hit games perform well from the start.
What the Data Says:
By Day 30, the fundamental engagement and spending patterns of your cohort are set in stone. If your core monetization isn’t working by Day 30, it never will.
| Player Action | % of Lifetime Actions by Day 30 |
|---|---|
| Players Converting (1st purchase) | 75% |
| Players Redepositing (2nd+ purchase) | 65% |
Checklist:
- Extend ONCE Rule: If D7 targets are missed, do you have a single, clear, data-backed hypothesis to fix it? You get one shot.
- Final Call: Are you making the final kill/scale decision by Day 30, with absolutely no exceptions?
Chapter 9: Insist on Statistical Confidence
“What is the exact number of users we need to confidently make this decision, and what is our plan to acquire them?”
The Problem:
“We have data, but we don’t trust it. Two whales skewed our LTV… We’re arguing about whether the data is even real.”
The Fix: Don’t Make Million-Dollar Decisions on Hundred-Dollar Data.
Making decisions on noisy data is worse than using no data at all. You must ensure you have a large enough cohort to provide a statistically significant reading.
What the Data Says:
| Metric to Measure | Desired Margin of Error | Minimum Installs per Cohort |
|---|---|---|
| 20% D7 Retention | +/-2% | ~1,540 |
| 10% D30 Retention | +/-2% | ~865 |
| 2.5% D7 Conversion | +/-0.5% | ~3,680 |
Checklist:
- Have you defined your required confidence level and margin of error for each core KPI?
- Have you used a sample size calculator to determine the minimum installs needed?
- Are you refusing to hold any gate meetings until your cohort reaches the required size?
Chapter 10: Optimize for Speed
“What is the total cost to get a confident decision, and how can we spend money on UA to reduce that time?”
The Problem:
“We’re trying to save money by limiting our UA spend… but the project is dragging on forever and costing a fortune in salaries.”
The Fix: Your Most Expensive Resource is Time. Buy a Faster Decision.
Soft launch UA is not a marketing cost. It is an investment in producing the data required to make a decision. Being “conservative” on UA spend is often the most expensive choice you can make.
What the Data Says (The Math of Speed):
A faster decision dramatically reduces total project cost by minimizing OPEX burn.
| Scenario | UA Spend / Month | Time to Decision | Total Cost to Kill |
|---|---|---|---|
| “Slow & ‘Safe’” | $30,000 | 3 months | $690,000 |
| “Fast & ‘Smart’” | $100,000 | 1.5 months | $450,000 |
Spending more on UA upfront saved $240,000.
Checklist:
- Have you calculated your true total weekly/monthly burn rate?
- Are you choosing the plan that minimizes total cost to decision, not just the initial UA budget?
- Is your CFO aligned on UA as a data production investment?
Conclusion: Ask Better Questions, Build Better Games
The Red Light by Default system is not a magic formula for creating a hit game. It is a disciplined framework for killing games that won’t be hits, faster and cheaper than your competition.
It forces you to ask harder, better questions at every stage of the process:
Are we building something players aspire to?
Are we designing for our most valuable fans?
Are our targets ambitious enough to force a clear decision?
Does the data provide overwhelming proof this game is a winner?
Are we optimizing for the fastest possible path to a confident decision?
By instilling this discipline, you protect your most valuable resources—your team’s time, money, and morale. You will still fail. But you will fail better. And that makes all the difference.
