PLAYBOOKS BYLIFECYCLE STAGE.

1. Open the Game Database and search for comparable titles in your genre
2. Add 2–5 competitor or reference games to your catalogue as "tracked" titles
3. Import reviews for each game and run analysis
4. Compare across games — look at polarity and emotion breakdowns, top topics, dominant named emotions, and player demographics
5. Identify gaps and opportunities — where do competitors fall short? What do players wish for that nobody delivers?

• Game-level polarity and emotion distributions across your catalogue
• Topic comparisons — which themes dominate feedback for each game
• Player demographic differences — how the audiences compare in playtime, engagement, and experience
• Emotional profiles — how each game makes its players feel

Analysing competitor games costs the same as analysing your own: 1 credit per snippet during analysis. Use filters to focus on the most relevant reviews and keep costs manageable.

💡 Tip: You don't need to analyse all 50,000 reviews for a major title. Import the most recent 200–500, filter by your area of interest (e.g. negative polarity or frustration tagged to combat), and analyse that focused set. You'll get the patterns you need without the cost.

• You're in pre-production and want to understand audience expectations in your genre
• You're positioning your game against specific competitors
• You want to validate that your unique selling points actually resonate differently from the competition

Game Database → Inbox → Snippets → Clusters (repeated per game)

1. Identify 3–5 reference games that share your genre, audience, or design pillars
2. Add them to your catalogue and import reviews for each
3. Run analysis — focus on the topics relevant to your design (e.g. filter by topic or keyword)
4. Build clusters around your core design questions: "How do players feel about crafting systems in survival games?" or "What frustrates players about roguelike progression?"
5. Generate insights to synthesise findings across games — look for patterns that hold across multiple titles, not just one

• Cross-game patterns — themes that recur across multiple comparable titles
• Audience expectations — what players in your genre take for granted vs. what delights them
• Risk signals — features or design choices that consistently generate frustration
• Opportunity gaps — things players wish for that no game in the space delivers well

Same as any analysis: 1 credit per snippet when you run analysis, 2 credits per accepted cluster, 5 credits per cluster when accepting an insight. Start small — 200 reviews per game is enough to surface the dominant themes.

💡 Tip: Frame your clusters as design questions, not just topics. Instead of a cluster called "Combat", build one called "What frustrates players about melee combat in action RPGs?" — filter for negative polarity or frustration + combat topic + high playtime. The evidence you get back directly answers your design question.

• You're in concept or pre-production and want to de-risk design decisions
• You're choosing between two creative directions and want evidence to inform the choice
• You're building a pitch and want to show you understand the audience landscape

Game Database → Inbox → Snippets → Clusters → Insights (across multiple reference games)

1. Analyse your own game's reviews through the full pipeline (Inbox → Snippets → Clusters → Insights)
2. Analyse 2–3 comparable titles to show how your game compares
3. Build clusters and insights around the themes that matter to your pitch — audience engagement, emotional resonance, competitive differentiation
4. Pull key data points: snippet counts, sentiment percentages, topic and emotion/taxonomy breakdowns, player demographics, trend direction
5. Use evidence in your materials — pitch decks, one-pagers, and live conversations backed by structured data

• Quantified sentiment — not just "positive", but "72% positive, with 'world-building' and 'atmosphere' as the top drivers" — and we go beyond that baseline with rich emotional taxonomies: named emotions (e.g. frustration vs. boredom vs. disappointment), not a single polarity bucket, so stakeholders see what kind of player experience you're delivering
• Competitive positioning — how your game's feedback profile compares to reference titles
• Audience depth — player demographics showing who your audience is (experience, engagement, playtime commitment)
• Momentum indicators — polarity and emotion trends, review velocity, and emerging themes over time
• Evidence chains — every claim traces back through insights → clusters → snippets → original reviews

Standard pipeline costs apply. For a pitch-ready analysis of your own game plus 2–3 comparables, budget approximately 2,000–5,000 credits total (depending on review volume and analysis depth).

💡 Tip: Frame your evidence around the questions publishers actually ask: "Who is your audience?" "How does player experience (emotion, topic, polarity) compare to competitors?" "What's your retention story?" "What are the biggest risks?" Build a cluster and insight for each question, and you'll walk into the meeting with answers, not hopes.

• You're preparing a pitch deck for a publisher or investor meeting
• You need to justify continued investment in a live title
• You want to prove product-market fit with structured data instead of anecdotes

Full pipeline: Game Database → Inbox → Snippets → Clusters → Insights → CoreDecisions + CoreProfile for audience depth

1. Find comparable EA titles in the Game Database — games in your genre that launched into Early Access
2. Import their reviews and filter for the EA period (use date filters to isolate early feedback)
3. Run analysis and focus on negative signal — polarity, named emotions, and topics — what did players complain about most in the first weeks?
4. Build clusters around EA-specific themes: missing features, performance expectations, content depth, communication cadence
5. Compare across titles to identify patterns — what do EA players in your genre consistently expect?

• EA expectation patterns — the features and polish levels players in your genre consider table stakes
• Forgiveness thresholds — what players tolerate in EA vs. what triggers refunds and negative reviews
• Communication themes — how dev responsiveness and roadmap transparency affect player emotion and narrative (not just average polarity)
• Early churn signals — what drives players to leave negative reviews in the first 2 hours

Same as any analysis pipeline. Focus your budget: import 200–300 reviews per reference game, filter for the first 30 days of EA, and analyse that focused set.

💡 Tip: Pay special attention to reviews from players with under 2 hours of playtime and strong negative emotion or polarity. These are your early churn signals — the things that make players bounce before giving your game a real chance. If you can address these before your EA launch, you'll have a much stronger first impression.

• You're 1–3 months from an Early Access launch
• You're deciding which features to include in your EA build vs. your post-EA roadmap
• You want to shape your store page messaging around what EA players actually care about

Game Database → Inbox → Snippets → Clusters → Insights (across EA reference titles)

1. Add a game to your catalogue from the Game Database
2. Open the Inbox and import your first batch of Steam reviews (up to 1,000 on first load)
3. Filter reviews by polarity (sentiment), emotion, playtime, language, or date to focus on the feedback that matters most
4. Run analysis — the system breaks your selected reviews into snippets, each tagged with topic, emotion/taxonomy, and polarity
5. Explore snippets — filter by topic (e.g. "Combat", "Performance"), named emotion (e.g. "Frustration"), and polarity to isolate specific themes

• Snippet cards showing the exact player quote, tagged with topic, emotion, and polarity (sentiment)
• Player context for each snippet — who said this, how long they played, how reliable their data is
• Filter controls to narrow by any dimension: topic, emotion, polarity, language, playtime, date range

Analysis costs 1 credit per snippet classified. A typical review produces 2–5 snippets, so analysing 100 reviews might cost 200–500 credits.

💡 Tip: Start with 50–200 focused reviews (e.g. negative polarity or strong negative emotion, high playtime) rather than analysing everything at once. You'll get meaningful results faster and spend fewer credits.

• You've just launched and want to understand first impressions
• You're preparing a patch and need to know what players care about most
• You want to move beyond "positive/negative" into the specific topics and emotions driving player experience

Inbox → Snippets → optionally Clusters and Insights for deeper analysis

1. Run analysis on your reviews in the Inbox (if you haven't already)
2. Open Clusters to see system-generated suggestions based on statistical patterns in your snippets
3. Review suggestions — each one shows the theme, snippet count, emotion and polarity distribution, and a confidence indicator
4. Accept or customise — accept a suggestion as-is, refine it by adding or removing snippets, or build your own from scratch
5. Explore the cluster detail view — volume, velocity, temporal charts, taxonomy breakdowns, emotional VAD shift, and polarity bars

• Suggested clusters with theme labels, snippet counts, and emotion/polarity distributions
• Cluster detail view with temporal charts (when did this feedback appear?), topic breakdowns, emotion analysis, and player signal aggregates
• Sub-clusters for drilling into a pattern without losing the parent theme's context

Accepting a cluster costs 2 credits. Exploring suggestions and previewing clusters is free — you only pay when you confirm.

💡 Tip: Start by reviewing the system's suggested clusters before building your own. They're generated from statistical patterns you might not spot manually — high-lift combinations of topic, emotion, polarity, and player segment that stand out from the baseline.

• You have 200+ snippets and want to see what themes dominate
• You suspect a recurring issue but want to quantify it before acting
• You want to compare how different player segments experience the same theme

Snippets → Clusters → optionally Insights for synthesis and CoreDecisions for action

1. Open Player Profiles for a game in your catalogue
2. Browse profiles to see experience bands, engagement levels, playtime commitment, and quality scores
3. Click a profile to see the full detail: summary, library, and signal tabs
4. Filter by segment — compare feedback from veterans vs newcomers, high-playtime vs casual, by region, platform, or commitment level
5. Cross-reference with snippets — see how the same theme plays differently across player segments

• Experience band — Steam account maturity based on library size, level, and playtime history
• Engagement band — how actively this player participates in the Steam community
• Playtime commitment — dedicated, invested, sampled, tried, or minimal
• Data quality — how publicly visible the profile is, affecting score reliability
• Confidence scores — per-dimension confidence (games, reviews, playtime, badges, achievements)
• Sample weight — a combined score used to weight this player's snippets in aggregate analysis

Exploring player profiles is free. Profiles are computed automatically when reviews are imported — no additional credit cost.

💡 Tip: Filter by "high playtime + negative polarity or strong negative emotion" to surface your most critical power-user feedback. These are the players who loved your game enough to invest serious time — and something drove them to leave a negative review. That signal is gold.

• You want to understand whether negative feedback comes from your target audience or drive-by reviewers
• You're comparing how different player segments experience the same feature
• You need to weight evidence in a cluster or insight by player credibility

Audience Intelligence takes the player-level data in CoreProfile and gives you a cohort view of the entire player population for a game. Rather than inspecting one profile at a time, you can filter the full reviewer population by experience band, playtime commitment, polarity, emotion, language, or any combination of signals and see aggregate breakdowns instantly.

This is especially useful when you want to know: "Of all the players who left negative reviews, what proportion are veteran players vs casual newcomers?" or "Do high-playtime players feel differently (named emotions, not just thumbs down) about this issue than samplers do?"

Open Audience Intelligence from the sidebar under CoreProfile → Audience Intelligence.

CoreProfile (cross-cutting) — feeds into Snippets, Clusters, and Insights

Audience Intelligence — cohort-level view of the full reviewer population

1. Run analysis on your reviews (or use existing snippets)
2. Filter snippets by emotion — isolate frustration, boredom, joy, or any other emotional signal
3. Build clusters around emotional themes — e.g. "Frustration with combat difficulty" vs. "Frustration with UI responsiveness"
4. Explore the cluster detail view — check the emotional VAD shift chart to see how emotion trends over time
5. Compare emotions across player segments — do veterans feel differently from newcomers about the same feature?

• Emotion tags on every snippet — precise emotional classification beyond positive/negative
• Emotion distribution in clusters — what's the dominant emotional response to a theme?
• VAD temporal charts — how valence (positive/negative), arousal (intensity), and dominance (feeling in control) shift over time
• Segment-level emotion differences — how different player types experience the same content emotionally

Emotion classification is included in the standard analysis cost: 1 credit per snippet. No additional charge for emotional signals.

💡 Tip: Combine emotion filters with playtime filters for the most actionable signals. "Frustration + 20 hours played" tells you your most dedicated players are hitting a wall — that's a retention risk. "Boredom + under 2 hours" tells you your opening isn't hooking players — that's an acquisition risk. Same emotion, very different problems.

• You want to understand the emotional experience of playing your game, not just whether players "like" it
• You're designing around feel — pacing, difficulty curves, narrative beats — and need evidence
• You want to detect emotional shifts after a patch or content update

Snippets (emotion filters) → Clusters (VAD analysis) → optionally Insights for synthesis

1. Review your clusters — identify the ones with the highest volume, strongest emotional signal, or fastest growth
2. Generate insights by selecting related clusters — the system synthesises them into a structured finding
3. Review the insight — check severity, linked evidence, emotion and polarity distribution, and player demographics
4. Create a decision from the insight — record what you're going to do about it and why
5. Set success metrics so you can measure whether the change worked

• Insight cards with severity (Critical/High/Medium/Low), type, and status
• Evidence trails linking back through clusters → snippets → original reviews
• Player demographics showing who is affected — veterans vs newcomers, high-playtime vs casual
• Decision records with linked evidence, status tracking, and revision alerts

• Generating candidate insights is free — you can preview before committing
• Accepting an insight costs 5 credits per attached cluster
• Monitoring a decision costs 1 credit per tracked metric per day

💡 Tip: Generate insights from multiple related clusters rather than one at a time. An insight built from 3 clusters that tell the same story is far more compelling than 3 separate single-cluster insights.

• You're planning a sprint and need to decide what to prioritise
• You need to justify a technical investment to your lead, publisher, or board
• You want to compare the severity of competing issues before committing resources

Clusters → Insights → CoreDecisions

1. Create a decision from an insight (or directly from Clusters)
2. Fill in the decision record — what you're doing, why, and what evidence supports it
3. Set success metrics (e.g. "negative polarity about matchmaking drops below 40%" or "frustration-tagged snippets in matchmaking fall by X%") and revision metrics (e.g. "if frustration increases post-patch, revisit")
4. Mark as Implemented when the change ships — this freezes the baseline snapshot
5. Monitor the dashboard — compare before/after polarity and emotion drivers, track emerging themes, and watch for revision alerts

• Baseline vs current comparison — polarity distribution, topic frequency, and named emotion intensity before and after your change
• Trend charts — daily tracking of key metrics over time
• Revision alerts — automatic notifications when reality diverges from expectations (e.g. polarity worsens or frustration spikes instead of improving)
• Evidence graph — trace the full chain from decision → insight → clusters → snippets → original reviews

Monitoring costs 1 credit per tracked metric per day. A typical decision tracks 2–3 metrics, costing 2–3 credits/day. Remove metrics you're no longer watching to save credits.

💡 Tip: Don't monitor everything. Pick 2–3 metrics that directly relate to the problem you were solving. If your decision was about matchmaking, track matchmaking-related emotion tags (e.g. frustration) and polarity — not overall game sentiment as one number. Focused monitoring gives clearer signals and costs less.

• You've shipped a patch or update and want to know if it worked
• You're running a live service and need to track emotion and polarity trends over time
• You want to build an evidence base for what works (and what doesn't) to inform future decisions

The Decision Surface gives you a real-time evidence pack for any game — showing the top drivers of player experience (topics, emotions, polarity), trend comparisons across time windows, and a narrative summary of decision pressure. Use it before creating a decision to understand the current state of play, and return to it after shipping to see how the driver rankings and emotional landscape have shifted.

For example: if "Matchmaking" was the top driver before your patch, check the Decision Surface after shipping to see whether it has dropped in the rankings and whether negative polarity and frustration in that area have decreased.

CoreDecisions (with evidence from Insights → Clusters → Snippets)

Decision Surface — for real-time driver analysis before and after implementation

1. Open Audience Intelligence from the sidebar under CoreProfile → Audience Intelligence, or from the game navigation bar
2. Select a game from your catalogue
3. Apply filters to define the cohort you want to examine — for example: Experience Band = Veteran, Sentiment = Negative
4. Review the aggregate breakdown — see how many players match, their polarity and emotion distribution, top topics, and engagement profile
5. Compare cohorts by adjusting the filters — swap Veteran for Newcomer and see how the same game lands differently across segments
6. Cross-reference with clusters and snippets — if a cohort shows unexpectedly high negativity on a specific topic, open Snippets filtered to that segment to read the actual feedback

• Cohort size — how many reviewers match your current filter combination
• Polarity breakdown — positive / negative / mixed distribution for the cohort (plus emotion tags where available)
• Top topics — the subjects this cohort mentions most
• Experience and engagement profile — the composition of the cohort by band and level
• Playtime distribution — how much time this cohort had invested before reviewing

Audience Intelligence is free to browse. Player profiles are computed automatically when reviews are imported — no additional credits are required to explore cohort views.

💡 Tip: Compare "High playtime + Negative" vs "Low playtime + Negative" for the same game. If veteran players are unhappy about something different from casual players, those are two separate problems that likely need two different solutions. Treating them as one issue is a common and expensive mistake.

• You want to know whether core players and casual players feel differently about a recent change
• You're preparing a decision and need to understand the demographic breakdown of the evidence
• You're monitoring a live service title and want to watch how cohort emotion and polarity shift over time
• You need to build a player-segmented argument for a stakeholder or publisher conversation

Audience Intelligence (CoreProfile) — cross-references with Snippets, Clusters, and CoreDecisions