What Is Pipeline Management? Stage Definitions, Stalled Deals, and Forecast Examples

Published:
Last Updated:
Category: CRM, LTV & Customer Management, Marketing DX
Authors: Shusaku Yosa
Pipeline management keeps opportunities, customer evidence, next actions, and forecasts current. Adding stage names is not enough. Define what counts as a deal, what moves it forward, and which period the forecast covers before interpreting the totals.
Define the Deals and Forecast Horizon in Scope
Separate a contact or enquiry from a qualified opportunity. Otherwise, a larger list may look like pipeline growth without representing more buying intent. Define the deal ID, customer, buying need, and duplicate rule. Keep the opportunity key distinct from the marketing initiative ID.
Choose a forecast horizon: orders expected by this quarter-end differ from potential opportunities over the next six months. Bookings, recognised revenue, and cash receipts are separate measures. Product types and buying procedures can also differ within one stage, so avoid relying on a single average for every deal.
Even a small pipeline may need a register when values, customer promises, or handoffs matter. There is no universal deal-count threshold below which management is unnecessary.
Build Stage Exit Criteria Around Customer Evidence
Move stages using customer evidence, not just a seller’s completed activity. Sending a proposal is different from the customer agreeing to evaluate it. These three stages are illustrative, not a recommended number for every business.
Stage | Customer evidence | Record | Next agreement and owner |
|---|---|---|---|
Needs confirmed | Customer confirms the problem and evaluation purpose | October 3 meeting note | Requirements meeting October 12; salesperson A |
Evaluating | Both parties agree requirements and evaluation process | Evaluation sheet v2 | Evaluation response October 20; salesperson A |
Terms under discussion | Budget, decision route, and remaining issues confirmed | Customer confirmation email | Terms discussion October 28; salesperson B |
Write concise entry and exit criteria, required evidence, and the exception decision-maker. Record missing evidence rather than forcing progress. Agree where different products or customer groups need a different route.
Record Next Actions and Reasons for Stalls
Include stage entry date, last customer check, next action, promised date, and blocker. “In progress” cannot distinguish a customer waiting for a document from an internal quotation delay. For an approval wait, name the responder, required material, and next check.
An age above the historical average is a reason to investigate, not proof of a lost deal or poor ownership. Check the buying cycle, renewal timing, deal size, and the next customer agreement. Averages based only on finished deals or excluding long cases may be biased.
A review should turn a stall into support: who will supply the evaluation material, and by when? Requiring an explanation without addressing the blocker does little to improve the record or the customer’s decision.
Separate Deal Counts, Value, and In-Period Win Probability
This hypothetical quarter forecast uses consistent JPY deal values and a shared cutoff. The 40% and 70% assumptions concern winning by this quarter-end, not eventual lifetime win rates.
Deal | Deal value JPY | Assumed probability | Expected close | Weighted current-quarter JPY |
|---|---|---|---|---|
A | 2,000,000 | 40% this quarter | Current quarter | 800,000 |
B | 1,000,000 | 70% this quarter | Current quarter | 700,000 |
C | 3,000,000 | 20% next quarter | Next quarter | Outside current-quarter scope |
Weighted open bookings = 2,000,000 × 0.4 + 1,000,000 × 0.7 = JPY 1,500,000. Adding separately recorded won bookings of JPY 500,000 on the same basis gives JPY 2,000,000 expected for the quarter. Do not include deal C in that total. This is an expected bookings value, not confirmed revenue, cash, or a guarantee for an individual deal.
In practice, inspect mature comparable histories, sample size, horizon, and movements between periods before assigning probabilities. State uncertainty for small samples and compare forecasts with outcomes to recalibrate. Retain unweighted value and deal counts so that changes in composition remain visible.
Handle Holds, Losses, and Reopened Deals with History
A hold differs from a confirmed loss. Automatically marking a budget-delayed deal lost and reopening it under a new ID may distort deal creation and conversion measures. Preserve the history, as in this hypothetical example.
Deal ID | State | Reason | Next check | History and forecast treatment |
|---|---|---|---|---|
D-014 | On hold | Customer budget pending | November 10 | Keep the ID; exclude from this current-quarter forecast |
D-021 | Lost | Customer confirmed competitor selection | None | Retain reason and confirmation date |
D-014 | Reopened | Budget confirmed | November 15 | Record hold start/end and revised close date |
A genuinely new buying project may warrant a new ID. Record its relationship to the earlier deal and the rule used. Preserve past states while ensuring the current view does not count the same opportunity twice.
Run Weekly Reviews Across Sales and Marketing
- Update customer evidence, next dates, values, and expected close dates.
- Select review cases using agreed ageing criteria and missed commitments.
- Explain new, won, lost, changed-value, and moved-in or moved-out deals separately.
- Assign help for missing material or decisions with an owner and due date.
- Feed customer reasons for qualification or rejection back into campaign improvement.
Salesforce’s official material distinguishes pipeline changes by period and close date. That product example does not establish another system’s stage or forecast capabilities. A lower stage conversion rate also does not identify messaging as the sole cause: inspect audience mix, deal mix, delivery capacity, and data changes. More exits than entries in one week is not a complete forecast of future pipeline value.
References (checked October 10, 2026)
Related reading




