Win/loss themes for the quarter

HubSpot + Gong

Who it's for

Sales leaders and RevOps who want to know why deals were won or lost last quarter, with counts and examples, before a QBR or planning session. HubSpot can be swapped for Salesforce, Attio or Pipedrive, and Gong for Fireflies or Granola.

What you get

Q3 win/loss themes, with counts and evidence. Sample output: every company and person in it is fictional.

Q3 2026 (Jul 1 to Sep 30): 46 closed deals in HubSpot, 19 won and 27 lost.
Analyzed in Gong: 22 deals you confirmed (all 14 lost deals over $20K, plus 8 won deals).

Why we lost (14 analyzed)
1. Stayed with the status quo: 6 deals, $188K
   - Mossgate Labs (fictional), $42K. Gong: the buyer said the current spreadsheet process was good enough for this year.
   - Fernhill Software (fictional), $31K. Gong: budget moved to a hiring freeze in August.
2. Never reached the economic buyer: 4 deals, $126K
   - Quarry Lane Analytics (fictional), $48K. Gong: finance joined no calls; procurement stalled after the proposal.
3. Lost to Tidewell Metrics (fictional): 3 deals, $77K
   - Larkspur Freightworks (fictional), $29K. Gong: Tidewell offered a free pilot.

Why we won (8 analyzed)
1. Security review done in under two weeks: 5 deals
2. A champion in operations who ran the internal pitch: 3 deals

Where HubSpot and Gong disagree: 3 lost deals have 'Price' as the lost reason in HubSpot, but Gong points to timing or budget freezes.
No Gong data: 1 lost deal (Brindle Harbor Co., fictional) had no calls in Gong. It's left out of the themes.

Small sample: themes are based on 21 deals with Gong data. Treat counts as directional.
No changes were made in HubSpot.

Time and cost

Estimate About 20–30 minutes, most of it confirming the deal list and reading the themes.

Assumptions: One quarter with about 45 closed deals, of which you confirm about 20 for Gong analysis.

Exact credit costs depend on your plan with each vendor. Check the tool pages before a large run.

MCP servers

Every server is confirmed in: ClaudeChatGPT

Set up in a chat app

Every server in this recipe is confirmed for Claude and ChatGPT. Connect each one, then paste the prompts below into one conversation.

Developer setup

Auth setup and mcp.json for Claude Code and Cursor
  • HubSpot: Sign in with OAuth in your client (run /mcp in Claude Code). Do it once interactively, and the client stores and refreshes the token.
  • Gong: Sign in with OAuth in your client (run /mcp in Claude Code). Do it once interactively, and the client stores and refreshes the token.

Put this in .mcp.json (Claude Code, project root) or ~/.cursor/mcp.json (Cursor). Swap the example keys for your own, and don't commit real ones. Remove the // comment lines if your client requires strict JSON.

{
  "mcpServers": {
    "hubspot": {
      "url": "https://mcp.hubspot.com"
    },
    "gong": {
      "url": "https://mcp.gong.io/mcp"
    }
  }
}

Step-by-step prompts

  1. Find all HubSpot deals closed won or closed lost between {quarter_start} and {quarter_end}. Show me a table with deal name, company, amount, owner, won or lost, and the lost reason if one is filled in. Give me the totals. Don't change any records.
  2. Suggest which deals to analyze in Gong: every lost deal over {min_amount}, plus up to {n_won} won deals from different owners. Show me the list and how many Gong calls that means, and wait for me to confirm it.
  3. For each deal I confirm, use Gong ask_deal once to ask why it was won or lost, which competitors came up, and the main objections. If Gong can't find the deal, use ask_account on the company and say so. Give me one line per deal.
  4. Group the reasons into themes, separately for won and lost. For each theme show the deal count, the number analyzed, the total amount, and two or three examples with the Gong answer or HubSpot field behind them. List deals with no Gong data, and every deal where the HubSpot lost reason and Gong disagree.
  5. Write a five-line summary I can paste into my QBR deck. Keep it here as text. Don't write anything to HubSpot.

Run it on a schedule (optional)

Where this goes wrong

Gotchas

Download

Get the prompts, guardrails and expected output as one Markdown file. Paste it into a Claude Project or ChatGPT project's instructions, or save it as CLAUDE.md for Claude Code.

Download win-loss-themes.md

Vendor sites: HubSpot. If we join an affiliate program, these links will be labeled as affiliate links.