Running win Loss Analysis

• Source: Report research library
• Frameworks applied: Unkover 7-step win/loss operating process, Clozd question-design framework (5 principles, 4 core topics), Five Whys probing, win/loss quadrant (Leverage Zone / Hidden Strengths), quote-level thematic coding, Corporate Visions status-quo-bias and Unconsidered Needs

Hard sample floors with a stated response rate, quote-level flat coding scheme, and buyer-stated versus rep-stated reasons kept in separate columns.

Known weakness, published as found: The header says 'The seven steps' but the checklist has eight items and the body numbers eight; align the count. Move the twelve-question core set and interview mechanics into a gated 'interview guide' subsection, since they apply only to the fieldwork phase and not to transcript-coding requests.

1. 1.Copy the SKILL.md text below, or download the raw file.
2. 2.Create a folder named exactly running-win-loss-analysis in your agent's skills directory.
3. 3.Save the file inside that folder as SKILL.md.
4. 4.Ask the agent one of the trigger requests below.
5. 5.Check the output against what you already know before it leaves your desk.

• Set up a win/loss program for us — we have no idea why we lose to Competitor X
• Here are 40 closed-lost deals from Q3. Design the interview guide and the sampling plan
• Analyze these lost-deal interview transcripts and tell me what to change in our GTM

• Score this discovery call transcript and give the rep one coaching point
• Write me a battlecard for Competitor X with trap-setting questions

---
name: running-win-loss-analysis
description: >-
  Runs a structured win/loss program end to end — sampling closed deals,
  writing the buyer interview guide, coding transcripts into themes, and
  synthesizing findings into GTM decisions with owners. Use when the user
  says "win loss analysis", "win/loss program", "why are we losing deals",
  "loss reasons", "buyer interviews", "lost deal interviews", "churn exit
  interview", "competitive loss review", "win rate by competitor", "post
  mortem on closed deals", or hands over closed-won and closed-lost deal
  lists, CRM loss-reason exports, or interview transcripts to analyze. Use
  this skill whenever the task involves learning from a set of already-closed
  deals through buyer feedback, even if the user does not say "win/loss". Do
  NOT use for forward-looking pipeline or forecast reviews, for coaching a
  single recorded call (see scoring-sales-calls), for authoring battlecards
  from the findings (see building-competitive-battlecards), or for NPS and
  general customer satisfaction surveys.
metadata:
  version: "1.0"
---

# Running win/loss analysis

Design and run a buyer-interview-based review of closed deals, then turn the pattern into decisions with named owners. Pipeline reviews, single-deal post-mortems, and CRM data mining are explicitly not this: a pipeline review is forward-looking, a post-mortem covers one deal, and CRM captures what reps think happened rather than what buyers experienced ([Unkover](https://unkover.com/blog/win-loss-analysis/)).

## The seven steps

Define goals and scope → build the interview pipeline → craft interview questions → collect supplemental data → analyze patterns not individual deals → report findings and drive action → close the loop ([Unkover](https://unkover.com/blog/win-loss-analysis/)). Copy this checklist and tick as you go:

```
- [ ] 1. Scope: decision the program must inform
- [ ] 2. Sample: pipeline sized to hit the interview floor
- [ ] 3. Guide: open questions, core topics, drill-downs
- [ ] 4. Supplemental data: CRM, call recordings, pricing
- [ ] 5. Code transcripts into themes; count, do not anecdote
- [ ] 6. Validate the sample and coding, fix, re-validate
- [ ] 7. Report by audience; assign owners and deadlines
- [ ] 8. Close the loop: measure whether the change moved win rate
```

**1. Scope.** Start from the decision, not the data: name the GTM decision this round must inform (which competitor to build positioning against, whether to change packaging, which segment to defund). Judgment step — you choose the scope. Decision criterion: if no one would act differently based on any possible finding, the scope is wrong and you should narrow it to one competitor or one segment.

**2. Sample.** These floors are non-negotiable because below them you are reporting anecdotes as patterns ([Unkover](https://unkover.com/blog/win-loss-analysis/)):

| Requirement | Value |
|---|---|
| Minimum interviews per quarter | 15–20 |
| Minimum split | at least 10 wins and 10 losses |
| Segment-level analysis (by competitor, deal size, persona) | 30–40 per quarter |
| Deals to recruit from, at a 10–30% response rate | 50–60 |
| Interview window after close | 2–4 weeks |
| Interview length | 20–30 minutes |

Interview within 2–4 weeks of close: after 30 days buyers rationalize the decision and after 60 days they can barely remember who they evaluated (same source). Interview the decision-maker — for losses the person who chose the competitor, for wins the person who signed off. Frame the outreach as "help us improve", not "tell us what went wrong" (same source). Exclude no-decision and stalled deals from win-rate math: win rate = wins ÷ total decisions × 100; win/loss ratio = wins ÷ losses (same source). Track no-decisions separately, because roughly 40% of pipeline deals end in no decision ([Corporate Visions](https://corporatevisions.com/blog/unconsidered-needs/)) and folding them into losses hides a status-quo problem behind a competitive one.

**3. Craft the guide.** Open broad and open-ended so the buyer tells their story, then let that story reorder your later questions ([Clozd](https://www.clozd.com/blog/a-framework-for-creating-the-right-win-loss-analysis-questions)). Openers, verbatim from that framework: "What prompted you to evaluate ________ solutions at this time?" / "How did you structure and run your evaluation?" / "What was your final decision?" / "Why was that the right decision for your business?"

Cover four core topics in every interview — solution requirements, pricing, competitive landscape, sales experience — with a few open questions each (same source). Anchor the core set on these twelve questions ([Unkover](https://unkover.com/blog/win-loss-analysis/)): how the need was identified; how the shortlist was built; who was involved and how their priorities differed; top three requirements; how price was weighed against features, support, and ease of use; whether any capability was a dealbreaker; which other vendors were evaluated and how they compared; anything a competitor offered that they wished you had; the single biggest factor in the decision; the moment that tipped the scales; how their impression of you changed through the evaluation; and one thing they would change about working with your team.

Interview mechanics, stated as rules because each one has a specific failure mode:
- Avoid closed-ended and scale questions during the interview; push "on a scale of 1-5" items to a short post-interview survey, because a rating ends the story you were trying to elicit ([Clozd](https://www.clozd.com/blog/a-framework-for-creating-the-right-win-loss-analysis-questions)).
- Never ask "Are integrations important to you?"; ask "What are the most important integrations to you and why?" (same source).
- Never ask a leading question. Replace "Why didn't you choose us?" with "Walk me through your evaluation process" ([Unkover](https://unkover.com/blog/win-loss-analysis/)).
- Let silences breathe and run the Five Whys: "we went with them on price" → "what about the pricing was different?" (same source).
- When a buyer names a product gap, pivot immediately into why that capability was necessary and what the ideal solution would have offered ([Clozd](https://www.clozd.com/blog/a-framework-for-creating-the-right-win-loss-analysis-questions)).
- Run a couple of practice sessions internally before the first real buyer interview (same source).

**4. Collect supplemental data.** Pull CRM stage history, call recordings, pricing and discount data, and the rep's stated loss reason for each interviewed deal. Keep the rep-stated reason in a separate column from the buyer-stated reason; never merge them, because the gap between the two is one of the program's main outputs.

**5. Code the transcripts.** Analyze patterns, not individual deals ([Unkover](https://unkover.com/blog/win-loss-analysis/)). Build a flat coding scheme with one code per decision factor (price level, pricing model fit, specific capability gap, integration, security or compliance, sales experience, implementation risk, incumbent inertia, champion turnover). Apply codes at the quote level, allow multiple codes per interview, and record for each code: frequency, deal outcome, competitor, segment, and ACV band. Keep the scheme flat rather than hierarchical, because nested schemes make cross-quarter comparison impossible once a subcategory is renamed.

Cut the coded data on: win rate by competitor, by deal size/ACV band, by lead source, and by buyer persona; and top cited decision factors, wins versus losses (same source). Plot factors on the win/loss quadrant — X-axis is win rate when the criterion matters, Y-axis is how often buyers mention it. High win rate plus high frequency is the **Leverage Zone**: proven strengths to double down on in positioning. High win rate plus low frequency is **Hidden Strengths**: surface them proactively in sales conversations (same source). The two low-win-rate quadrants are unnamed in that source, so describe them by their coordinates rather than inventing labels.

**6. Validate, fix, re-validate.** Run all checks. If any fails, fix it and re-run the full list before writing a single finding, because a finding published from a nine-interview sample cannot be withdrawn from the sales floor once it circulates.

```
- [ ] At least 15 completed interviews, with at least 10 wins and 10 losses
- [ ] Segment-level claims are backed by 30-40 interviews, or labeled directional
- [ ] Every interview happened within 2-4 weeks of close, or the delay is noted
- [ ] No-decision and stalled deals excluded from win-rate math and tracked separately
- [ ] Every theme reports a count and denominator ("8 of 12 competitive losses")
- [ ] Buyer-stated and rep-stated loss reasons are compared, not merged
- [ ] Every recommendation has a named owner and a deadline
```

**7. Report by audience.** Same data, four cuts ([Unkover](https://unkover.com/blog/win-loss-analysis/)): sales leadership gets win-rate trends, the top three loss reasons with examples, competitor gaps to push into battlecards, and recommended training; product gets feature gaps in lost deals ranked by revenue impact, unmet must-haves, and a competitive parity matrix; marketing gets messaging gaps, perception versus intended positioning, content that influenced or was missing, and comparison-page accuracy; executives get a one-page summary plus three recommendations with owners and deadlines, and trend lines rather than snapshots. Write executive findings in the revenue-attached form: "We lost $1.2M in pipeline to Competitor X last quarter because buyers cited [specific gap] in 8 out of 12 competitive losses" (same source). Cadence: quarterly reports, monthly metrics dashboard (same source).

**8. Close the loop.** Update battlecards, then measure whether the change moved outcomes — if "lack of enterprise SSO" was the top loss reason and product shipped it, measure win rate against that competitor next quarter. Expand scope from the top two competitors outward, revise questions as the product evolves, and keep a running log of insights, decisions, and results (same source).

## Output format

Use this section order for the main report. Executives and product read different sections, and reordering breaks the one-page executive extract.

```markdown
# Win/Loss Findings — <period>, <scope>
Sample: N interviews (W wins / L losses) | Response rate: X% from D deals recruited
Interview window: <median days from close> | Excluded: <no-decisions, stalled>

## Headline
<One sentence with revenue attached and a count/denominator>

## Win rate cuts
| Cut | Win rate | n |  (by competitor, ACV band, lead source, persona)

## Themes
| Theme | Frequency (n/N) | Outcome skew | Competitor | Representative quote |

## Quadrant
Leverage Zone: | Hidden Strengths: | Low win rate / high frequency: | Low win rate / low frequency:

## Buyer-stated vs rep-stated loss reasons
| Deal | Rep reason (CRM) | Buyer reason (interview) | Match? |

## Recommendations
| # | Recommendation | Audience | Owner | Deadline | Metric that will move |

## Next round
<Scope, sample target, questions to revise>
```

## Gotchas

- Only 15% of CRM loss reasons match what buyers say in first-hand interviews, meaning 85% are inaccurate or incomplete — Clozd's comparison of 1,000 closed-lost deals against buyer interviews, reported in [Unkover](https://unkover.com/blog/win-loss-analysis/). Never seed the coding scheme from CRM loss-reason picklists; you will reproduce the rep narrative with a research veneer.
- "We lost on price" is almost never the finished answer. Run the Five Whys on it in the interview ([Unkover](https://unkover.com/blog/win-loss-analysis/)), and when coding, split price level from pricing-model fit — the fixes are opposite (discounting versus repackaging).
- Recruiting is the binding constraint, not analysis. At a 10–30% response rate you need 50–60 recruited deals to complete 15–20 interviews ([Unkover](https://unkover.com/blog/win-loss-analysis/)), so build the recruitment list at the start of the quarter rather than after close.
- Losses to "no decision" are a status-quo-bias problem, not a competitive one. Around 40% of pipeline deals end in no decision ([Corporate Visions](https://corporatevisions.com/blog/unconsidered-needs/)), and the corrective is Unconsidered Needs messaging that disrupts the status quo, not a sharper competitive comparison.
- Buyers cannot tell you about needs they never considered — Unconsidered Needs are by definition not discoverable through voice-of-the-customer research ([Corporate Visions](https://corporatevisions.com/blog/unconsidered-needs/)). Treat interview output as authoritative on the evaluation as run, and never as a product roadmap generator.
- Spreadsheets stop scaling past about two quarters of data, and above 20 interviews per quarter you need interview-management or call-analysis tooling; third-party interview studies run $2K–$5K per study and take 4–6 weeks ([Unkover](https://unkover.com/blog/win-loss-analysis/)). Say this before a leader commits to a program cadence the team cannot staff.
- Expected impact is real but slow: systematic win/loss is associated with a 15–25% win-rate improvement over two years (McKinsey) and comprehensive programs with up to 50% win-rate improvement and 15–30% revenue increase (Gartner), both cited in [Unkover](https://unkover.com/blog/win-loss-analysis/). Do not promise movement inside one quarter.
- Common win-rate "benchmarks" circulating for B2B SaaS (20–30% competitive win rate, 40–50% for top performers, SMB versus enterprise splits) carry no named source on that page (same source). Exclude them from any report; use the org's own baseline instead.

What does the Running win Loss Analysis skill do?

Runs a buyer-interview win/loss program from sampling through coding to audience-specific findings with named owners

Where does the Running win Loss Analysis skill come from?

Report research library. It was written by The Revenue AI Report against a 12 criterion quality rubric and graded in an independent scoring pass.

Why was the Running win Loss Analysis skill chosen for this library?

Hard sample floors with a stated response rate, quote-level flat coding scheme, and buyer-stated versus rep-stated reasons kept in separate columns.

When should the Running win Loss Analysis skill not be used?

Do not use it for: Score this discovery call transcript and give the rep one coaching point Or: Write me a battlecard for Competitor X with trap-setting questions

How do I install the Running win Loss Analysis SKILL.md file?

Download the file, create a folder named exactly running-win-loss-analysis inside your agent's skills directory, and save the file inside it as SKILL.md. The agent loads it when a request matches the description.