The Strategy
Push quality relative to price point using first and second-party research. Launch Early Access to fund R&D while honing 1.0. Go big at 1.0 with optimized platform/geo mix.
The goal: Find innovative differentiators and quality levers through internal playtesting and competitive analysis of past and upcoming games targeting the same audience. Price point anchors consumer expectations for scope and polish—research identifies where to exceed those expectations.
The path: Ship to Early Access with a great core that convinces consumers the game can become an excellent 1.0. EA revenue offsets R&D and validates direction. The better the EA release, the higher your steady-state revenue during the honing period. 1.0 is the main event—platform mix (PC-only, cross-platform, geo rollout) determined by EA learnings.
Methodology: This playbook derives strategy from mechanisms (cause → effect), not historical success patterns. Examples inform pattern identification; strategy derives from understanding WHY something works. "Helldivers 2 did X" is not a strategy.
1
Game Type Classification
Critical first step before applying any framework
Before applying ANY framework, classify the game. Kill criteria, quality gates, and strategic options differ fundamentally between types.
Systemic-Infinite
Replayable loops; emergent gameplay; theoretically endless
Roguelikes, Sims, Strategy, Survival, Sandbox
Key Metrics: Session length, D7/D30 retention, replay rate
Content-Finite
Linear progression; narrative-driven; consumed and finished
Narrative RPG, Horror, Adventure, Walking Sim
Key Metrics: Completion rate, churn points, time-to-finish
Hybrid
Core loop + narrative wrapper; replayable but with endpoint
Action RPG, Metroidvania, Immersive Sim
Key Metrics: Both metric sets; weight by design emphasis
Why This Matters
Applying Systemic metrics to Content-Finite games kills profitable projects. A linear narrative game with 15% D30 retention might be perfectly healthy if 80% of players completed it.
2
Foundational Patterns
12 mechanism-based patterns for AA publishing
1
Feedback Quality Degrades with Audience Mismatch
When the feedback source differs from the target buyer, iteration optimizes for wrong variables. Preference heterogeneity exists WITHIN demographics—two players with identical profiles can have wildly different tolerances.
- Define target buyer FIRST (not "everyone")
- Screen testers for ownership/playtime in 3-5 reference games your target buyer loves
- Post-test survey: "Compared to [reference game], this game is..." on 5 dimensions
- Alignment score = % of testers whose evaluation matches your positioning
- Benchmarks: >70% = proceed | 50-70% = investigate | <50% = redefine target
2
Reputation Damage Compounds Through Discovery Algorithms
Steam's algorithm responds to revenue velocity (dollar spikes in short windows), not review scores directly. Reviews affect conversion rate, which affects revenue velocity. Recovery cost exceeds typical AA capitalization.
- No public exposure until quality gates pass
- If launching EA, monitor daily revenue and review sentiment
- Intervention trigger: <70% positive reviews after 2 weeks → intensive patch/communication
- Escalation trigger: <60% positive after 30 days → evaluate pull vs. continued investment
- Recovery arcs require 12-24 months + $500K-2M budget
3
Revenue Timing Creates Discipline Problems
EA revenue during development creates optionality but also pressure. If studio NEEDS EA revenue to survive (not just extend), the framework fails regardless of market.
- Studios must be capitalized to ship 1.0 WITHOUT EA revenue
- EA revenue = upside for polish, not survival
- Model EA revenue conservatively: Regional-focused EA = 40-50% of Western per-unit
- Warning sign: EA revenue projections required to close budget gap = undercapitalized
EA Is The Launch (Dec 2025 Data, n=225)
80% of EA games generate MORE revenue during EA than at 1.0. Median 1.0 revenue = 40% of EA first-month. Even massive hits decline at 1.0: Hades II (-75%, still $11M), Backpack Battles (-87%), Supermarket Simulator (-95%, still 3.2M units). These aren't failures—they're proof that EA has become the primary revenue event. Model accordingly: EA is your launch window, 1.0 is closer to a major update or DLC drop.
Review Score Debt Is Fatal
Games graduating with "Mixed" reviews from early EA have near-zero 1.0 recovery rate. The 1.0 launch cannot fix a damaged "Recent Reviews" score fast enough to hit discovery algorithms. If reviews go Mixed in EA, intervention must happen immediately—not at 1.0.
4
Platform Leverage is Portfolio Fit, Not Quality
Sony/Xbox negotiate based on portfolio gaps and timing—not quality alone. A mid-tier exclusive filling a calendar gap is worth more than a better game competing with first-party.
2024-2025 Market Reality
Game Pass deals have contracted significantly. MG covering 50%+ of dev cost is now optimistic, not baseline. Have a viable PC-only plan if platform deals don't close.
5
Market Windows Exist But Duration is Unknowable
Category maturity follows S-curves. Early entrants face less competition but unproven demand. Late entrants face more competition but validated demand.
- Test demand signals before committing to window-dependent strategy
- Build validation checkpoints, not assumptions
- Define explicit "window closed" triggers and pivot protocols
- Monitor competitor announcements monthly
6
Price Sets Genre-Contextualized Expectations
Players evaluate games against what they paid AND genre norms. An 82 MC at $40 in tactics = triumph. An 82 MC at $40 in open-world RPG = "what's wrong with it?"
| Genre | Typical Range | Notes |
|---|---|---|
| Tactics/Strategy | $25-50 | Higher end for 4X/grand strategy |
| Roguelike/lite | $10-30 | Premium roguelikes pushing $25-30 |
| Action RPG | $40-60 | Scope-dependent |
| Narrative/Adventure | $20-40 | Length-dependent |
| Survival/Crafting | $20-40 | EA often lower |
| Horror | $20-40 | Length and production value dependent |
7
Virality is Enabled by Design, Not Marketing
Games go viral when they have shareable moments + low friction to group adoption + social proof loops. These are design properties. Features that make games watchable can inversely correlate with retention.
- Design for virality enablement (co-op, emergent systems, streamable moments)
- Base case P&L must work WITHOUT viral lift
- Audit for anti-viral properties at greenlight
- Don't sacrifice retention mechanics for streamability
- Short-form audit: Does the game produce 15-60 second clips that work on TikTok/Reels?
8
Discovery is Algorithmic, Triggered by Revenue Velocity
Steam algorithms respond to revenue spikes (dollars earned in short windows). Reviews, wishlists, and CCU matter because they drive revenue velocity. First 48 hours determine discovery trajectory.
- Concentrate marketing around algorithmic trigger points
- Sustained low-level spend < surgical spikes
- Key widgets: New & Trending, Popular Upcoming, Discovery Queue
- Missing launch window = relying entirely on long-tail discounting
9
Wishlist Velocity > Wishlist Volume
Recent wishlists convert better than old wishlists. Games with shorter pre-release visibility (~214 days vs ~411 days) tend to perform better. Correlation ≠ causation—better games may simply move faster.
- Median wishlist-to-first-week-sales: ~0.15-0.20x (for 10K+ wishlists)
- Games priced >$10: conversion drops to ~0.10x median
- Variance is extreme: 10-20x range between similar games
- Wishlists >6 months old may need re-activation
10
Development Velocity Determines Market Timing
Delays compound non-linearly. A 6-month delay shifts competitive window, may invalidate platform timing, and allows sentiment to shift. Games that ship well iterate in days/weeks; games that ship poorly iterate in months.
11
Post-Launch Economics Determine LTV
First 30 days determine base sales. Months 2-24 determine lifetime value through discount cycles and DLC. AA games with "AAA content expectations" face margin compression.
- Decide at greenlight: Ship-and-move-on OR live service
- If live service: budget content roadmap from day 1
- If ship-and-move-on: plan discount cadence (20% at 4-6wk, 33% at 3mo, 50% at 6mo)
- First major DLC 3-6 months post-launch captures engaged players
12
Steam Next Fest is Primary Wishlist Generation Mechanism
Steam Next Fest (Feb, June, Oct) is now make-or-break visibility for games without existing brand. Demo exposure drives wishlist velocity. Post-July 2024, demos appear permanently in player libraries.
Cold Start Problem
Games with <2,000 wishlists entering Next Fest struggle to gain traction. The system rewards existing momentum. Bad demo is worse than no demo.
3
Strategic Options
Four paths with failure modes and intervention triggers
A
Closed Alpha → Global Simultaneous Launch
Best for: Content-Finite games, narrative-heavy titles, games where "fresh" launch matters
Execution Sequence
- Closed alpha (NDA): 1-5K players matching target demo
- Iterate until quality gates pass
- Demo at Steam Next Fest (2-4 months before launch)
- 1.0 launch: PC + platforms simultaneously
- Marketing: Surgical spikes at launch + major updates
⚠️ Failure Modes
- Cold Start: No EA user base to drive Day 1 CCU. Marketing alone may not trigger algorithms.
- Capital Runway Wrong: Development extends; no EA revenue to bridge.
- Market Window Closes: Competitor ships during closed development.
- Demo Underperforms: Next Fest reveals market isn't there.
Intervention Triggers
D30 retention <15% (Systemic) or Completion <60% (Content-Finite) after 2 cycles
Deep design review; consider pivot or additional cycle
Runway <12 months before quality gates pass
Evaluate: scope cut, bridge funding, or pivot to Option B
Demo wishlist conversion <3% at Next Fest
Market validation failed; reassess positioning or viability
B
Early Access as Refinement Engine
Best for: Systemic-Infinite games only. NOT appropriate for Content-Finite games.
Execution Sequence
- Closed alpha: Validate core loop with target demo
- Quality gates must pass before public EA
- Demo at Steam Next Fest (optional, can align with EA launch)
- Public EA launch
- Iterate based on live feedback + revenue
- 1.0 launch after quality threshold hit
- Platform deals negotiated during late EA using live data
Not Appropriate For
Content-Finite games (linear narrative, horror, adventure), games where story spoilers matter, games where "fresh" launch perception is critical.
⚠️ Failure Modes
- Review Spiral: Ship below threshold; negative reviews compound.
- EA Revenue Disappoints: Regional pricing + patient buyers = 40-50% of model.
- Feedback Misalignment: Optimize for EA audience, degrade for 1.0 audience.
- EA Fatigue: Game stays in EA too long; audience moves on.
📊 EA Market Reality (Dec 2025, n=225 graduates)
| Metric | Value | Planning Implication |
|---|---|---|
| Median EA duration | 14 months | Plan 12-18 month runway from EA to 1.0 |
| Average EA duration | 21 months | Outliers pulling average up; >24 months = fatigue risk |
| 1.0 vs EA revenue | 40% | Model 1.0 as tail revenue, not second peak |
| 1.0 outperformers | 20% | Only 45 of 225 did better at 1.0—don't plan for this |
📈 The 20% That Beat EA (What They Did)
| Game | 1.0 vs EA | Mechanism |
|---|---|---|
| News Tower | +138% | Novel concept re-viralized; 3x CCU at 1.0 |
| Mars First Logistics | +181% | Massive localization expansion at 1.0 |
| Escape The Backrooms | +189% | Evergreen streamer/community engine; 6M copies LTD |
Pattern: All three expanded their addressable market at 1.0 (new geographies, new audiences, new platforms)—they didn't just "finish" the game.
🎯 Actionable Rules
- Discount strategy determines 1.0 potential: Heavy EA discounting trains audience to wait; minimal discounting preserves 1.0 demand
- Console parity multiplies 1.0: Simultaneous console launch creates unified marketing "beat" beyond Steam algorithms
- Localization is market expansion: Adding languages at 1.0 = new TAM, not just translation
- Price increase signaling: Announcing 1.0 price hike drives "buy now" surge in final EA week
Intervention Triggers
<70% positive reviews after 2 weeks
Emergency patch cycle; increase communication; identify top complaints
<60% positive after 30 days despite intervention
Evaluate: major pivot, pull from sale, or abandon 1.0 plans
D30 retention <10% (Systemic games)
Core loop broken; major design intervention required
C
Platform-First (Game Pass / PS Plus)
Best for: Games where guaranteed floor matters more than upside ceiling
Execution Sequence
- Closed alpha: Validate core loop
- Pitch platforms during development with prototype + team pedigree
- Secure platform deal (if achievable in current market)
- Continue development with platform involvement
- 1.0 launch on platform + PC (manage exclusivity terms)
2024-2025 Market Reality
Game Pass deals have contracted significantly. Many developers report deals "not worthwhile" outside Game Pass. Don't plan around platform deal that may not materialize. Have PC-only viable path as backup.
Intervention Triggers
Platform deal negotiation stalls >3 months
Activate PC-only backup plan; don't hold development
MG offer <40% of dev cost
Evaluate trade-offs; may be better to decline and go PC-first
D
Genre Community Incubation
Best for: Niche genres with passionate, organized communities (50K+ combined)
Execution Sequence
- Identify underserved genre with active community
- Engage community during pre-production (genuine design input, not marketing)
- Closed alpha with community superfans
- Community becomes organic marketing engine
- Demo at Steam Next Fest with community support
- EA or 1.0 launch supported by community momentum
⚠️ Failure Modes
- Community is Ceiling: 50K subreddit doesn't equal 50K buyers.
- Tyranny of Superfans: Hardcore demands make game impenetrable to broader audience.
- Community Turns Hostile: Transparent development backfires on unpopular decisions.
4
Portfolio Management
Risk distribution and intervention discipline
Risk Distribution (Limits, Not Targets)
These are risk exposure limits, not target allocations. Actual mix depends on deal flow.
| Risk Profile | Max Exposure | Rationale |
|---|---|---|
| High-risk (Option A) | 30% of capital | Highest upside, highest variance |
| Medium-risk (Option B) | 60% of capital | Bread and butter; data-driven |
| Low-risk (Option C) | 40% of capital | Cash flow stability; capped upside |
Key Principle
Don't reject exceptional opportunities to maintain spreadsheet ratios. Limits prevent overconcentration, not optimal allocation.
Project Intervention Discipline
The goal is not to kill projects but to intervene before they become unrecoverable.
| Flag | Condition | Response |
|---|---|---|
| 🟡 Yellow | Metric below benchmark | Increased monitoring, identify root cause |
| 🟠 Orange | Below benchmark after intervention | Leadership review, consider pivot |
| 🔴 Red | Multiple failures, interventions unsuccessful | Kill review with default to kill |
Expected: 30-40% of greenlit projects should be killed or pivoted pre-launch. This is healthy, not failure.
5
Quality Gates
Pre-greenlight, pre-public, and demo gates
Pre-Greenlight Gates (Market Validation)
| Gate | Method | Benchmark |
|---|---|---|
| Concept Appeal | Steam page test (coming soon) | Wishlist/visit >7% |
| Competitive Clarity | Landscape analysis | <3 direct competitors announced |
| Team Velocity | Prototype iteration cycles | <2 weeks per meaningful change |
| Budget Reality | Estimate vs. capital | Capital = 140%+ of estimate |
| Cold Start Path | UA test | <$3/click OR organic traction |
Pre-Public Gates: Systemic-Infinite Games
| Gate | Metric | Benchmark |
|---|---|---|
| Core Loop | Avg session length | Genre-appropriate |
| Retention | D7 retention | >25% (adjust by genre) |
| Replay Signal | % starting second session | >60% |
| Technical | Crash rate | <1% |
| 2-Hour Hook | Refund proxy | >80% |
Pre-Public Gates: Content-Finite Games
| Gate | Metric | Benchmark |
|---|---|---|
| Completion | % finishing game | >60% |
| Pacing | Churn by chapter | No section >20% drop |
| Satisfaction | End-of-game NPS | >50 |
| Technical | Crash rate | <1% |
| 2-Hour Hook | Refund proxy | >85% |
Demo Quality Gates (Pre-Next Fest)
| Gate | Requirement |
|---|---|
| Represents core loop | Accurately shows what the game IS |
| Polish level | Near-release quality; not "work in progress" |
| Length | 15-30 minutes; enough to hook, not exhaust |
| Completion rate | >70% of demo starters finish |
| Technical | <0.5% crash rate |
6
Steam Next Fest Playbook
Timing, execution, and benchmarks
Timing Strategy
| Launch Window | Next Fest Timing | Rationale |
|---|---|---|
| Q1 launch | October (prior year) | 3-4 month wishlist build |
| Q2 launch | February | 2-4 month wishlist build |
| Q3 launch | June | 2-3 month wishlist build |
| Q4 launch | October | 1-3 month wishlist build |
Ideal gap: 2-4 months between Next Fest and launch.
Pre-Fest Checklist (4-6 weeks before)
Demo-to-Wishlist Benchmarks
| Performance | Conversion | Implication |
|---|---|---|
| Strong | >15% | Market validated; maintain momentum |
| Good | 10-15% | Viable; identify optimization opportunities |
| Marginal | 5-10% | Concerns; analyze feedback carefully |
| Weak | <5% | Market validation failed; reassess |
7
Commercial Terms Framework
Publisher-developer economics
Traditional recoup (60/40 until recouped, then 40/60) creates misaligned incentives. Alternative structures:
Revenue Share from Dollar One
30/70 split from first dollar; no recoup
✓ High-confidence projects; attracts top talent
Partial Recoup
50/50 until 50% of advance recouped; then 30/70
✓ Balanced risk sharing
Traditional Recoup
60/40 until full recoup; then 30/70
✓ Higher-risk projects; more publisher control
Milestone-Based Equity
Publisher takes equity stake; no recoup
✓ Franchise bets; long-term alignment
Minimum Terms Checklist
8
Localization Framework
Language prioritization by tier
TIER 1
Launch Languages
Covers ~75% of Steam revenue
English
Simplified Chinese
German
Russian
TIER 2
Within 30 Days
Covers ~90% of Steam revenue
French
Spanish (LATAM)
Japanese
Brazilian Portuguese
TIER 3
Based on Performance
Korean
Polish
Turkish
Italian
Genre Adjustments
| Genre | Prioritize |
|---|---|
| Strategy/Tactics | German, Russian |
| JRPG-style | Japanese, Korean |
| Horror | Japanese, Korean |
| Survival | Chinese, Russian |
Budget guidance: Full localization per language = $15-50K depending on text volume and voice
9
Decision Checklists
Greenlight, pre-public, launch, and post-launch
At Greenlight
Before Public Exposure
At Launch Decision
Post-Launch
Version 3.1 Key Changes (Dec 2025)
- EA Is The Launch: Reframed EA strategy based on n=225 graduates data—80% generate more revenue in EA than at 1.0
- Review Score Debt: Added warning that Mixed reviews in EA = near-zero 1.0 recovery (Gemini/Grok validated)
- EA Market Reality Table: 14-month median duration, 40% revenue ratio, planning implications
- The 20% That Beat EA: Case studies with specific mechanisms (News Tower, Mars First Logistics, Escape The Backrooms)
- Actionable Rules: Console parity, price increase signaling, localization as market expansion
Sources: GameDiscoverCo (Dec 2025), cross-validated via SteamDB, VG Insights, Gamalytic, Game World Observer
v3.0 Changes
- Added Game Type Classification (Systemic vs. Content-Finite)
- Added Steam Next Fest playbook (Pattern 12)
- Softened kill criteria to intervention triggers
- Added Commercial Terms Framework
- Updated Platform Deal economics for 2024-25 market
- Changed price anchors from data to methodology
- Added Localization Framework
- Converted portfolio allocation from targets to limits