── SKILL
closed-lost-analysis✓ APPROVED

Categorize your losses and name the fix.

Sorts lost deals into no-decision, price, competitor, timing, or fit, pulls competitive intel, and splits preventable from structural.

Inside: The loss taxonomy, the competitive intel pull, and the fix per theme.

01 / HOW TO USE

Install it in one line, or paste it in.

1
In Claude Code (one command)
Copy the install line, paste it into your terminal, and restart Claude Code. The skill installs itself to ~/.claude/skills/ and runs automatically when it is relevant.
2
In Claude, ChatGPT, or a Project (no terminal)
Open the file, then upload it to your chat or paste its contents in. A skill is just a markdown file of instructions, so any capable AI can follow it.
New to skills? A skill is a plain-text file that teaches your AI a workflow. Point any capable assistant at it and it follows the steps, on your data.
02 / MAKE IT YOURS

Connect your context. Set it to your motion.

CONNECT YOUR CONTEXT · AND WHY IT HELPS
a CRM

it reads all of the above automatically, across every lost deal in the window you pick, including the raw loss notes.

a meeting or transcript tool

reads what the buyer actually said before they went dark, so "no decision" gets a real cause.

a product-analytics tool

on lost trials, shows whether they ever reached first value, which separates a product gap from a sales gap.

SET IT TO YOUR MOTION

This was built for a B2B SaaS org running a staged pipeline. Set these to your stack:

Run any loss taxonomy you like. The skill sorts your losses into your buckets and names the fix per bucket, so point it at your fields, not anyone else's.

Set thisWhat it isDefault / Example
LOST filterhow you mark a lossStage = Closed Lost
WINDOWthe loss-date range to analyzetrailing 4 quarters
REASON fieldyour loss-reason picklistClosed Lost Reason
NOTES fieldfree-text loss notesloss descriptionnext-step notes
COMPETITOR fieldwho you lost tocompetitor on the deal
STAGE_AT_LOSS fieldhow far it got before dyingstage when marked lost
SEGMENT fieldshow you group accountsindustryemployee countregion
MIN_SAMPLEfewest losses before a theme counts5 (raise if you lose a lot)
03 / THE FULL SKILL

Everything the skill does, in full.

── WHAT THIS DOES

Reads your closed-lost deals and tells you why they died, in buckets you can act on: no decision, price, competitor, timing, or fit. It pulls the competitive intel out of the losses, then splits them into two piles that matter more than the total: the losses you could have prevented, and the ones that were structural and never yours to win. Each theme ends with the fix. The point is not to feel bad about the number. It is to stop losing the same deal twice.

── THE METHOD
  1. 1
    Loss categorization (clean buckets)

    Sort every lost deal into one primary reason: NO_DECISION (they picked nothing, status quo won), PRICE (budget or value gap), COMPETITOR (they picked someone else), TIMING (real but not now), or FIT (never a match). When the reason field is blank, read the notes or the transcript to place it. Report the share and the lost deal size behind each bucket, so a common reason and an expensive reason are both visible.

  2. 2
    Competitive intel pull

    For every COMPETITOR loss, pull who won and, where the notes say so, why. Roll it into a per-competitor view: how often they beat you, at what stage, in which segment, and the reason that keeps repeating. This is the sheet your reps wish they had before the next competitive deal.

  3. 3
    Preventable vs structural split

    Split every loss into two piles. PREVENTABLE: the deal was winnable and something in the motion lost it (single-threaded, slow follow-up, wrong persona, weak business case, late competitor entry). STRUCTURAL: it was never yours (no budget, no real need, wrong segment, a feature you do not build). This split matters more than the raw count, because only one pile is yours to fix.

  4. 4
    The fix per theme

    Every theme ends with a specific, do-this-next fix, not a platitude. NO_DECISION driven by no compelling event points to earlier qualification of the trigger. PRICE losses that cluster at one tier point to packaging. A competitor winning late points to multi-threading sooner. Name the change, not the feeling.

── QUALITY GATES
  • No theme reported below MIN_SAMPLE. A single bad-luck loss is not a trend.
  • Every loss lands in exactly one primary bucket, so shares add up and nothing is double-counted.
  • Preventable is only claimed with a reason from a field, a note, or a transcript, never a hunch.
  • Competitive claims name the source. No invented win reasons for the other side.
── OUTPUT (EXAMPLE)
EXAMPLE OUTPUT
CLOSED-LOST ANALYSIS · 41 losses, trailing 4 quarters

Loss reasons
Bucket         Share   Lost $     Note
No decision    37%     $520K      Status quo, no compelling event
Competitor     27%     $410K      Lost late, after Solution Validation
Price          20%     $300K      Clustered at the mid tier
Timing         10%     $140K      Real, revisit next cycle
Fit            6%      $70K       Wrong segment, disqualify earlier

Preventable vs structural
  Preventable: 54%  (no-decision + late competitor losses)
  Structural:  46%  (timing, fit, no-budget)

Who we lose to
Competitor     Losses   Wins late?   Segment
Rival A        8        Yes, 6 of 8  mid-market
Rival B        3        No           enterprise

Fixes:
  1. No decision. Qualify the compelling event before Stage 3, or park it.
  2. Rival A. Multi-thread by Stage 2. They win by out-flanking a single champion.
  3. Price at the mid tier. Revisit packaging, not discounting.
── WHERE THE NUMBERS COME FROM

MIN_SAMPLE (5) and the trailing-4-quarter window are defaults, not laws. They suited a mid-market SaaS cycle with steady loss volume. If you lose a small number of large deals, widen the window so the buckets are not built on three data points. The preventable-vs-structural split is always grounded in your own notes and fields, so the fix list is yours, not a benchmark's.

── EXAMPLE CHAT
EXAMPLE CHAT
You
Here's our Q2 closed-lost list, 22 deals. Why do we keep losing?
CLAUDE
Read's done. Your biggest bucket isn't the one your team talks about. Bucket Share Lost $ Note No decision 41% $290K Status quo, no compelling event Competitor 23% $180K Lost late, after demo Price 18% $140K Clustered at entry tier No decision is your real problem, not the competitor you keep post-morteming in deal reviews. 9 of 9 no-decision losses had no compelling event on file at any stage. That's a qualification gap, not a competitive one. One sharpener: add a loss-notes field to your CRM so the next run can tell "no decision, ghosted" apart from "no decision, picked a spreadsheet."
You
Can you split those no-decision losses into preventable vs structural?
CLAUDE
Yes. Of the 9, 6 had an active champion past Stage 2, meaning the deal was alive and died anyway. That's preventable. 3 never had budget confirmed, that's structural. Preventable: 6 of 9 (67%) - champion present, no compelling event pushed Structural: 3 of 9 (33%) - no budget confirmed at any stage The fix for the preventable pile: qualify the compelling event by Stage 2 or park the deal. One sharpener: connect your meeting tool so the next run reads what the buyer actually said before going dark, instead of inferring it from stage history alone.
04 / GO FURTHER

Where an operator takes this next.

The read is step one. Here's where an operator takes it once the manual version proves out.

You built the read once; now it runs itself.

01
Run it every quarter without asking

Point a scheduled Claude task at Salesforce closed-lost exports and drop the themes into Slack before the next deal review.

02
Feed the fix list to the field, not just the deck

Push the "qualify compelling event by Stage 2" rule into a Gong scorecard so reps get flagged live, not in a postmortem.

03
Close the loop on the competitor pile

Route every COMPETITOR loss into a competitive-battlecard update automatically, so the sheet reps wish they had actually gets built.

05 / PART OF A BIGGER PLAY

One skill is the on-ramp.

── GET NEW SKILLS IN YOUR INBOX

One operator-built skill a week. Free.