Stop 7 of 8 · Weekly
Pipeline and forecast
The HubSpot export becomes a stale-deal list, a next-step audit and a Friday forecast with one honest confidence line per deal.
Forty deals, none of them honest
Dalia's pipeline says the same thing every quarter: a hopeful total, close dates that have been pushed twice, and a stage column that describes what she meant to do rather than what happened. She knows which eight deals are real. She cannot defend that to her manager without an hour she does not have.
The export fixes it. HubSpot gives you a CSV of open deals with amount, stage, close date, last activity and owner. That file plus your call notes is enough for Claude to do the reading you keep postponing.
The three passes
| Pass | What Claude looks for |
|---|---|
| Stale | No two-way contact in twenty-one days |
| Next-step gaps | No named event, or an event with no date |
| Date drift | Close date moved more than once |
Run all three before the forecast, because the forecast is only as honest as the rows underneath it. A deal with a pushed date, no reply in a month and no next step is not a best case deal, whatever the stage says.
Try this
Attached is my HubSpot open-deals export: [attach the CSV]. Three tables. One: deals with no inbound contact in twenty-one days, oldest first, with the last thing that happened. Two: deals with no next step, or a next step with no date. Three: deals whose close date has moved more than once, with how far it has moved. Then give me the five deals to spend Monday on and why each one made the list. Do not total anything yet.
Same export, plus my call notes for these deals: [paste or attach them]. For every deal with a close date this month, give me one row: company, amount from the file, the named event that has to happen for it to close, whether that event has a date, and a confidence of commit, best case or unlikely with one sentence of evidence. Total the commit deals only, using the amounts in the file. Do not adjust or estimate any amount. Then list the two best case deals closest to commit and what I would have to do this week.
Here are the notes and emails on the Harborline deal: [paste them]. Tell me, using only what is here, what actually stopped after the walkthrough, who went quiet, and the three most likely reasons. Rank them and say what I would ask to test the top one.