octavehq/win-loss-report

/octave:win-loss-report - Visual Win/Loss Report Builder

Generate beautiful, self-contained HTML win/loss analysis reports powered by your Octave deal intelligence. Unlike /octave:win-loss-report which outputs text-based analysis, this skill renders structured visual reports with CSS-based charts, progress indicators, comparison bars, and metric cards -- designed for leadership reviews, team retrospectives, and strategic planning.

Uses the same CSS variable / style preset system as /octave:deck.

On-brand styling — brand kit first, then generate

Resolve the brand before generating (do not skip this step). The report brand should be the workspace company's brand — that is, the Octave customer whose workspace you are operating in.

Step 1: Identify the workspace company. Call get_workspace_company to get the company name, domain/URL, and positioning. This is the company whose brand the report should use (whatever get_workspace_company returns is the brand, not the target account).

Step 2: Resolve the workspace company's brand kit. Slugify the workspace company name and check for a cached brand kit at ~/.octave/brands/<slug>/manifest.json. If a complete kit exists (has manifest.json and tokens.css), use it automatically:

• inline the kit's tokens.css (:root + the embedded @font-face) and get-brand-components/assets/kit_base.css into the output <style>;
• follow brand-kit.md → Signature moves, and reuse the kit's real logo for topbar and footer, images/, and icons.json. If no complete kit exists → build one. Run the get-brand-components skill (read ../../skills/get-brand-components/SKILL.md and follow it) for the workspace company's domain. If the first attempt returns incomplete results (no logo, no colors, partial data) → retry up to 3 times with different approaches (root domain, www. prefix, /about subpage). Only fall back to a generic preset after 3 genuine failures.

Step 3: Only use a generic preset as a last resort — after the workspace company's brand kit cannot be built.

Strong default: The report is always branded as the workspace company (the Octave customer). This is an internal report about deal outcomes — the workspace company's fonts, colors, topbar, and footer are used throughout.

Usage

/octave:win-loss-report [--period <timeframe>] [--segment <filter>] [--competitor <name>] [--style <preset>]

Examples

/octave:win-loss-report                                          # Last 90 days, all deals
/octave:win-loss-report --period "Q4 2025"                       # Specific quarter
/octave:win-loss-report --competitor "Gong"                      # Focused on deals vs Gong
/octave:win-loss-report --segment "enterprise"                   # Enterprise segment only
/octave:win-loss-report --period "last 6 months" --style paper-minimal
/octave:win-loss-report --period "2025" --competitor "Salesforce" --segment "mid-market"

Instructions

When the user runs /octave:win-loss-report:

Step 1: Define Scope

If not provided via flags, ask the user interactively:

Period -- "What time range?"

What time range should this report cover?

1. Last 30 days
2. Last 60 days
3. Last 90 days (default)
4. Last 6 months
5. Specific quarter (e.g., Q4 2025)
6. Custom date range

Your choice:

Filter -- "Any specific focus?"

Do you want to filter the analysis?

1. All deals -- full cross-deal analysis
2. Specific competitor -- focus on deals involving a competitor
3. Specific segment -- filter by market segment
4. Specific rep -- filter by sales rep
5. Custom filter -- describe what you want

Your choice:

If competitor or segment is selected, use list_entities to show available options:

# For competitor filter
list_entities({ entityType: "competitor" })

# For segment filter
list_entities({ entityType: "segment" })

Depth -- "How detailed?"

What level of detail?

1. Executive summary -- 1-page overview with key metrics and takeaways
2. Full report -- detailed analysis with all sections and drill-downs

Your choice:

| Depth | Sections Included | Best For | |-------|-------------------|----------| | Executive summary | Scorecard (headline metrics only), Win/Loss Analysis (top 3 each, no evidence), Data Sources | Board updates, weekly stand-ups | | Full report | All 4 sections with full detail | QBRs, strategy sessions, enablement |

Step 2: Octave Data Gathering

Based on scope, use Octave MCP tools to gather comprehensive deal intelligence. Always tell the user what you're researching and why.

Call as many tools as needed to build a complete picture. Win/loss reports are only as good as the data behind them. Layer multiple data sources -- deal outcomes + conversation findings + library context -- to produce analysis grounded in real evidence, not speculation.

See tool-reference.md for list-vs-search guidance and the tool reference tables for core deal data, conversation intelligence, library context, and competitor-focused data.

Output of this step: Present a report outline to the user for approval before generating. See outline-template.md for the outline template.

Wait for user approval before proceeding.

Step 3: Generate HTML

Load the shared rules before writing any HTML. Read each before producing output:

• Editorial rules — universal language rules, AI-ism kill list, banned vocabulary
• Presentation principles — universal visual rules, spacing, restraint
• Information principles — content structure, narrative arc, evidence quality
• HTML document format — format-specific visual rules for scrollable docs
• Octave value — prioritize grounded workspace data over generic AI content

Apply these rules during generation, not just during review. The review pipeline catches what slips through — but the generation step should internalize them from the start.

Build a single, self-contained HTML file. No external dependencies except Google Fonts. Everything else inlined.

After writing the file, proceed immediately to Step 4 (Review Pipeline). Do NOT open the file in the browser or present it to the user yet.

Output Directory

Reports go under .octave-reports/:

.octave-reports/
└── win-loss-<YYYY-MM-DD>/
    └── win-loss-report.html

Example: /octave:win-loss-report --period "Q4 2025" produces .octave-reports/win-loss-2026-02-11/win-loss-report.html

The .octave-reports/ directory is in .gitignore -- nothing here gets committed.

HTML Architecture

See html-architecture.md for the HTML architecture, structural requirements, and the CSS component system (metric cards, win rate ring, bar charts, use case cards, pattern cards, objection cards, head-to-head deal grids, tags, tab system).

Report Sections (Full Report)

See sections.md for the full 4-section report breakdown and the condensed executive summary variant.

Key structural principles:

• Every metric shows BOTH volume and rate with explicit axis labels
• Segments, personas, use cases, and competitors are attributed on every card — not floating as abstract statistics
• Section 2 (Win/Loss Analysis) is a tabbed pair: Win Analysis ←→ Loss Analysis
• Section 3 (Head-to-Head Deals) is tabbed by dimension: By Competitor | By Segment | By Persona
• Objections live inside the Loss tab (sorted by loss correlation, not frequency); where available, back each one with a verbatim quote from search_call_transcripts rather than a paraphrase
• Actionable changes live inside the Loss tab as "What to Stop Doing / Changes to Make," with an optional one-line "Winnable if: [X]" callout per pattern where the evidence points to a clean turning point
• "No Decision / Stalled" is an optional loss pattern category, included only if the workspace's loss data actually distinguishes no-decision from competitor losses, never fabricated

Step 4: Review Pipeline — MANDATORY GATE

Do NOT open the report in the browser, present the delivery summary, or tell the user the report is ready until the review pipeline has completed and you have a scorecard.

Load the review protocol and execute the review loop against the generated HTML file. The protocol specifies the full process, but here is the win-loss-report-specific wiring:

4a: Mechanical lint (before spawning reviewers):

bash <skill-dir>/scripts/lint.sh <path-to-report.html>

Fix every violation the lint surfaces. Deterministic checks: em-dashes, Tier 1 banned words, banned phrases, text density, leaked internals.

4b: Spawn two reviewers in parallel (both Task calls in a single message):

Editorial reviewer:

Task tool:
  subagent_type: "octave-editorial-reviewer"
  prompt: "Review the file at [FILE PATH].
           Read these principle docs and run each Review Checklist:
           1. [skill-dir]/../shared/editorial-rules.md (language quality)
           2. [skill-dir]/../shared/information-principles.md (information structure)
           Fix violations inline. Return scorecard."

Presentation reviewer:

Task tool:
  subagent_type: "octave-presentation-reviewer"
  prompt: "Review the file at [FILE PATH].
           Read these principle docs and run each Review Checklist:
           1. [skill-dir]/../shared/presentation-principles.md (universal visual rules)
           2. [skill-dir]/../shared/formats/html-document.md (format-specific visual rules)
           3. [skill-dir]/references/html-architecture.md (skill-specific CSS)
           4. [skill-dir]/references/sections.md (skill-specific structure)
           Fix violations inline. Return scorecard."

4c: Loop decision. Read both scorecards:

| Cycle | 0 fixes | 1-2 fixes | 3+ fixes | |---|---|---|---| | Cycle 1 | CLEAN → 4d | Apply, loop | Apply, loop | | Cycle 2 | CLEAN → 4d | Apply, STOP | Apply, loop | | Cycle 3 (cap) | CLEAN → 4d | Apply, STOP | Apply, STOP |

Max 3 cycles. Re-run both reviewers each loop (back to 4b).

4d: Output combined scorecard to the user. This is proof the pipeline ran. Step 5 cannot start without it.

REVIEW PIPELINE COMPLETE
=========================
Editorial:
  Mechanical:     [N fixes / PASS]
  Structural:     [N fixes / PASS]
  Quality:        [N fixes / PASS]
  Information:    [N fixes / PASS]

Presentation:
  Visual Rules:   [N fixes / PASS]
  Format Rules:   [N fixes / PASS]
  Design System:  [N fixes / PASS]
  Structure:      [N fixes / PASS]

Total fixes: [N]
Cycles: [1-3]
Status: [CLEAN / N remaining issues]

Step 5: Delivery

After the review pipeline scorecard has been output:

1. Open the report in the default browser
2. Present a summary:

REPORT READY
=============

File:    .octave-reports/win-loss-<date>/win-loss-report.html
Period:  [Date range]
Deals:   [N] won, [N] lost ([N]% win rate across [N] closed deals)
Style:   [Brand kit name]
Depth:   [Executive Summary / Full Report]

Key Findings:
- Win rate: [N]% ([N] of [N] closed deals)
- Top win use case: [Use case] ([N]% win rate across [N] deals)
- Top loss use case: [Use case] ([N]% loss rate across [N] deals)
- Highest-risk objection: [Objection] ([N]% loss rate when raised)
- Biggest competitor threat: [Competitor] ([N] deals, [N]% win rate)

Sections:
- Scorecard: metrics + segment/persona/competitor breakdowns
- Win/Loss Analysis: tabbed use cases, patterns, objections
- Head-to-Head Deals: tabbed by competitor/segment/persona
- Data Sources

Navigation:
- Scroll to navigate between sections
- Sidebar dots show your position
- Tabs switch between Win/Loss analysis and Head-to-Head dimensions
- PDF: Cmd+P / Ctrl+P -> Save as PDF (tabs flatten to labeled blocks)

---

Want me to:
1. Drill into a specific competitor
2. Drill into a specific segment
3. Add more detail to any section
4. Generate an executive summary version (or full version)
5. Export as PDF
6. Done

MCP Tools Used

Deal Intelligence

• list_events - Filter by DEAL_WON, DEAL_LOST, OPPORTUNITY_CREATED for deal outcomes and pipeline
• get_event_detail - Deep dive on notable wins and losses for deal stories

Conversation Intelligence

• list_findings - Objections, value prop presentations, competitor mentions, feature requests, proof points cited in calls
• search_call_transcripts - Verbatim objection/competitor quotes filtered by dealOutcome (WON/LOST/OPEN) — the evidence quotes behind the Win/Loss tabs
• get_entity_evidence - Best verbatim quotes evidencing one objection or competitor entity, for a specific card

Library Context

• list_entities - Quick inventory of competitors, segments, personas
• list_entities - Full entity data for proof points, references
• get_entity - Deep dive on specific competitors, personas
• list_motions - List Motions in the workspace
• list_motion_icps - List Motion ICP cells (persona × segment) under a Motion
• find_motion_icp - Motion ICP narrative + Learning Loop learnings
• search_knowledge_base - Semantic search for positioning content
• search_resources - Uploaded CRM exports, deal data files

Error Handling

No Deal Data in Period:

No won or lost deals found for [period].

This could mean:

1. CRM integration isn't syncing deal outcomes
2. The date range has no closed deals
3. Filters are too restrictive

Try:

1. Expand the date range
2. Remove filters (competitor, segment)
3. Check your Octave CRM integration settings

Insufficient Data (Fewer Than 5 Deals):

Only [N] deals found for [period]. Win/loss analysis is most useful with 5+ deals.

Options:

1. Proceed anyway -- I'll generate the report with available data (patterns may be unreliable)
2. Expand the time period to capture more deals
3. Remove filters to include all segments/competitors

No Findings Data:

Deal outcomes found, but no conversation findings available.

The report will include deal metrics and outcomes but won't have:

• Evidence quotes from calls
• Objection analysis
• Value prop effectiveness

For richer analysis, ensure calls are being recorded and findings extraction is enabled in Octave.

Octave Connection Failed:

Could not connect to your Octave workspace.

The win/loss report requires deal data from Octave. Check your Octave MCP configuration and reconnect.

Missing Competitor/Segment Data:

I couldn't find a competitor named "[name]" in your library.

Available competitors:

• [List from list_entities]

Pick one from the list, or proceed with "all deals" and I'll break down by competitor automatically.

Related Skills

• /octave:win-loss-report - Text-based win/loss analysis (this is the visual version)
• /octave:insights - Conversation intelligence analysis (conversational format)
• /octave:battlecard-doc - Competitive deep-dive (when patterns point to a specific competitor)
• /octave:pipeline - Current pipeline coaching and deal strategy
• /octave:deck - Present win/loss findings to leadership as a slide deck