mcpField
C
customer-commitment-validation

customer-commitment-validation

github.com/lvtd-llcMarketing & growth161 installsUpdated 7w ago
Repository

About customer-commitment-validation

A meeting-outcome review skill for product, sales, partnership, investor, and customer conversations. It distinguishes concrete commitments from weak interest by assessing time, reputation, or money given up, then defines follow-up actions.

Key features

  1. 01Classify meeting outcomes
  2. 02Assess time, reputation, and money
  3. 03Identify concrete next steps
  4. 04Record owners, dates, and success criteria
  5. 05Treat rejection as evidence

Use cases

  1. 01Review a customer meeting
  2. 02Assess progress in a sales conversation
  3. 03Evaluate a partnership discussion
  4. 04Plan follow-up after a rejection

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/lvtd-llc/skills --skill customer-commitment-validation

After installing, type/customer-commitment-validationin Claude Code to confirm it loaded.

Bundled filesagents/openai.yamlworkflows/classify-meeting-outcome.md

Details

From the project README

Customer Commitment Validation

Use this skill after a product, sales, partnership, investor, or customer conversation where the team needs to decide whether the meeting created real progress. The core test is what the other person gave up: time, reputation, or money.

Source Traceability

Primary source: The Mom Test by Rob Fitzpatrick, especially chapters 5 and 8. Guidance is paraphrased for this MIT repo; authoring notes used converted EPUB lines 2365-2790 and 3613-4108.

Commitment Currencies

CurrencyExamplesSignal Strength
TimeScheduled next meeting with goals, trial setup work, real feedback sessionMedium
ReputationIntro to a decision maker, internal championing, public case studyMedium to high
MoneyDeposit, preorder, paid pilot, purchase order, signed letter of intentHigh

Compliments, vague "keep me posted" replies, and future-tense buying promises are not commitments.

An advancement must move the real-world funnel to a specific next decision or action. A next step without an owner, date, or customer-side effort is usually weak interest.

Review Workflow

  1. Restate the meeting context and what the team hoped would happen next.
  2. Classify the outcome as commitment, advancement, rejection, weak interest, or bad data.
  3. Identify what the other person gave up, if anything.
  4. If the outcome is fuzzy, propose a concrete next ask that creates a decision.
  5. If there was rejection, preserve the learning and decide whether it changes the product, segment, positioning, or sales path.
  6. If there was commitment, define the owner, date, success criteria, and follow up.

Outcome Classification

OutcomeMeaningAction
CommitmentThey gave up time, reputation, or moneyTrack and follow through
AdvancementThey moved the real-world funnel forwardConfirm next step details
RejectionThey declined a concrete askLearn and update assumptions
Weak interestThey were friendly but gave up nothingAsk for a concrete next step or stop counting it
Bad dataOutcome is only compliments or future hypotheticalsReturn to evidence or commitment

Output Format

# Commitment Review

## Meeting
- Person/company:
- Context:
- Stage:

## Outcome
- Classification, evidence, currency given up:

## Interpretation
- What this proves / does not prove:
- Risks, missing decision makers, approval path:

## Next Ask
[A concrete ask that creates advancement or rejection.]

## Follow-Up
- Owner, deadline, success criteria:

Workflow

Use workflows/classify-meeting-outcome.md when the user provides notes, transcripts, CRM entries, or a narrative of how a meeting ended.

Quality Bar

  • Do not call a meeting successful because it felt positive.
  • Do not count a lead as real until they had a chance to reject a concrete ask.
  • Do not overvalue free trials that cost the customer nothing.
  • Do not ignore missing decision makers or approval paths.
  • Treat rejection as useful evidence, not as failure.

Comments

Markdown supported