Bigmind

Win-loss analysis

Bigmind’s agents analyze a set of closed deals and the calls behind them to explain why you win and lose, with win rates, sample sizes, real evidence and specific process and coaching actions.

By BigmindThe AI workspace for revenue teamsbigmind.ai
Built on Bigmind
Claude

Why did we lose deals last quarter?

Claude Opus 5.5

Why did we lose deals last quarter?

Your team

How it works

01

It defines the cohort

An agent sets the period, owners and what counts as won or lost in your CRM.

02

It joins the calls

The agent matches each sampled call to the right closed deal before reading it.

03

It counts the rates

The agent computes win rates and amounts and always shows the denominator.

04

You act on patterns

The agent names patterns with evidence and creates the tasks or playbooks you want.

The skill

Win Loss Review

Read the workflow contract before execution. For this workflow, also read pipeline and frameworks, knowledge and coaching.

Resolve period, owner scope and actual won/lost definitions. Retrieve the closed cohort with completeness/currency checks; use recorded outcomes rather than assuming a missing opportunity was won. Join each sampled conversation to the right deal.

Compute supported win rates and amounts with denominators. Inspect meaningful wins and losses for explicit reasons, contradictory evidence and unknown causes. Loss-stage distribution and time-in-stage need history.

Return patterns, sample sizes, representative evidence and specific process/coaching actions. Create requested tasks, playbooks, battlecards or coaching resources; do not claim causal effects or calibrated predictive weights from associations.

By Bigmind· Updated Oct 12, 2026· MIT· Source on GitHub

Run it

In your agentOne line installs it in Claude Code, Codex, Cursor and other agents, and it reaches your workspace through Bigmind’s MCP server.
npx skills add bigmindai/ext-plugins --skill win-loss-review

Install with:

Checks

Works with Bigmind
Built on Bigmind: it reads and writes your workspace through Bigmind’s tools.
Needs
A Bigmind workspace.
Tested by Bigmind
Bigmind builds and maintains it.

What it reads

Your CRMAccounts, contacts, deals and their stages, in HubSpot or Salesforce.
Your callsWhat was said on every recorded call, with the transcript and summary.

When to use it

  • The quarter just closed and you want to know why deals were lost.
  • Your win rate dropped and you need evidence before changing the process.
  • You’re updating a battlecard and want the real reasons from lost deals.

When not to

  • You want to know which competitors come up across calls. Use Battlecards.
  • The deal is still open. Use Deal review.
  • You want to predict which open deals will close. Use Sales forecast.

Questions

What is a win-loss analysis?

A win-loss analysis studies closed deals to learn why a company wins and loses. It combines outcomes from the CRM with what customers said on calls, then turns the patterns into changes to the sales process, messaging and coaching.

Does the win-loss analysis skill prove why deals were lost?

The win-loss analysis skill reports patterns, not proven causes. Bigmind’s agents show sample sizes and representative evidence, mark a cause as unknown when nobody stated one and never claim predictive weights from associations in a set of deals.

Does the win-loss analysis skill need Bigmind?

Yes. The win-loss analysis skill installs in Claude Code, Codex, Cursor and other agents, but it reads your closed deals and calls through Bigmind’s MCP server, so it needs a Bigmind workspace.

Skills to use with it