Review the sales pipeline every week

When to use

Every week, on a fixed slot, while the company has open deals. Not for deciding whether one specific deal is real — that is sales.qualify-lead. Not for chasing a sent offer that has gone quiet — that is sales.follow-up-offer. Not for the deep look at one large account — that is sales.quarterly-account-review. This playbook is the weekly heartbeat that keeps all of them honest.

Before you start

• Last week's snapshot exists. If this is the first run, take the current state as the baseline and say so in S1.
• The period goal (monthly bookings or quarterly revenue) is set in the goals module.
• The owner has 30 minutes blocked and will actually attend; a review without the owner produces no decisions.

What a run requires1

The trail8 steps

1. S1

   Take the snapshotagent

   Pull every open deal with stage, value, expected close date, owner and last activity date. Freeze it as this week's baseline row.

   Done when the snapshot row exists with a timestamp and deal count.

2. S2

   Check hygieneagent

   List every deal that has: no next action, no close date, a close date in the past, no activity in 14 days, or a value of zero.

   Done when the list is complete — including deals that should be closed as lost.

3. S3

   Compute movementagent

   Compare against last week: deals added, won, lost, slipped (close date moved), changed in value. Weight by stage using the company's standard stage probabilities.

   Done when the movement table shows each number next to last week's.

4. S4

   Flag and proposeagent

   For each deal from S2 and each slip from S3, write a one-line proposal: advance, push the date (to when), discount (by how much), or close as lost (why).

   Done when every flagged deal has exactly one proposal and the proposals are consistent with the period goal.

5. S5

   Review and decideownerneeds approval · owner

   Approval · S5 · owner — the run stops until a named person records the decision

   Walk the movement table and the flagged list. For each flagged deal the owner names one decision and a date. The owner states the forecast for the period in one sentence.

   Done when every flagged deal has a decision with a name and a date next to it, and no flagged deal is left "we'll see".

   ⛔ A review that ends with "keep an eye on it" on three or more deals has not happened — re-run S5 on those deals.

6. S6

   Record decisionsagent

   Update each deal's stage, close date and next action per the decisions. Add the follow-up tasks with due dates (bc tasks add). Close lost deals with a reason.

   Done when the live pipeline matches the decisions — no flagged deal keeps its old state.

7. S7

   Update the forecast and goalsagent

   Recalculate the weighted forecast for the period and record it where the money side reads it; it is an input to money.cash-forecast-13w. Check the period goal in the goals module against the forecast and note the gap (bc goals add only for a corrective target the owner asked for).

   Done when this week's forecast figure and the gap are written down with the snapshot.

8. S8

   Reportagent

   Send the owner the one-page summary: movement, decisions taken, forecast, gap to goal, deals with no next action (target zero).

   Done when the summary is sent and filed with the snapshot row.

Checks — how we know it worked

• Read back three random decisions from S6 against the live records — stage, date and next action match.
• Zero flagged deals remain unflagged: every deal has a next action and a date.
• This week's snapshot total minus won/lost equals last week's total plus adds — the arithmetic closes.
• The forecast figure in S7 matches what S8 reported — one number, not two.

If it goes wrong

What each step leaves behind

1. S1the snapshot row exists with a timestamp and deal count.
2. S2the list is complete — including deals that should be closed as lost.
3. S3the movement table shows each number next to last week's.
4. S4every flagged deal has exactly one proposal and the proposals are consistent with the period goal.
5. S5every flagged deal has a decision with a name and a date next to it, and no flagged deal is left "we'll see".
6. S6the live pipeline matches the decisions — no flagged deal keeps its old state.
7. S7this week's forecast figure and the gap are written down with the snapshot.
8. S8the summary is sent and filed with the snapshot row.

Evidence to keep

Snapshot row with timestamp · movement table · the flagged list with proposals · the owner's decisions with dates (who decided, when) · the forecast figure and gap · the sent summary. Next week's run starts by reading last week's — that is what makes this a review and not a meeting.

How this playbook improves

After every 8 runs ask: how much owner time did the review take, and did S5 finish inside it? What share of deals left the review with a dated next action? Was forecast error shrinking or flat? Did any deal slip its close date twice without a decision? A new version changes the step that leaked — usually S4's proposals or S5's decision discipline — and says so in its change note.