Why a 60% Stage-4 Deal Closes Less Than Half the Time

Open almost any CRM and you'll find a number attached to each pipeline stage. Discovery is 20%. Proposal is 60%. Negotiation is 80%. Multiply by deal value, sum the column, and you have a forecast.

It's clean and instant. It's also mostly fiction.

Where the number comes from

The weight on a stage isn't a measurement of a specific deal. It's an average. Sometimes that's a real historical average. More often it's a round number someone typed in during the CRM setup three years ago and never revisited. It says: across all the deals that have ever sat in this stage, roughly this fraction closed.

That's a useful fact about your pipeline in aggregate. It is almost useless as a statement about the deal in front of you. Two deals can sit in "Proposal, 60%" on the same Tuesday and have nothing in common.

One has an engaged economic buyer, a champion who's already sold it internally, and a signed evaluation plan. The other has a single mid-level contact who went quiet eleven days ago and asked for pricing before you understood their budget. The CRM calls them both 60%.

A forecast built on that math will tell you you're fine right up until the quarter ends and you're not.

The three ways it lies

It hides slippage inside healthy-looking stages. A deal can sit in stage four for ninety days, untouched, and still contribute its full weighted value to the forecast every week. The number doesn't decay. Nothing about "60%" knows that the last meaningful contact was in March. So the deal looks alive in the roll-up while it quietly dies in real life. You find out when the rep finally marks it closed-lost the week before the board meeting.

It rewards stage inflation. Reps learn the system. Pushing a deal from "Discovery" to "Proposal" raises its weighted contribution and gets the manager off their back in the pipeline review. So deals migrate forward on optimism rather than evidence. The exit criteria get skipped, because the CRM rewards the move regardless. Your forecast inflates precisely where your process is weakest.

It treats every rep as identical. Take a deal in "Negotiation" run by someone whose best-case calls convert 30% of the time. It is not the same as that deal run by someone who lands 70%. Stage weights flatten the difference completely. Your most optimistic rep and your most reliable rep contribute identical numbers, and you find out which was which only in hindsight.

What actually predicts the close

If you want a number that means something, stop asking what stage the deal is in. Start asking what's true about it. A handful of signals do most of the work:

• Stakeholder coverage. Is this single-threaded, or have you reached the economic buyer and the technical evaluators alongside the champion? Single-threaded deals are the ones that evaporate when one person changes jobs.
• Playbook completion. Were the exit criteria for each stage actually met, with evidence, before the deal advanced? Or did it jump forward because the rep felt good after a call?
• Engagement recency, read against stage. Fourteen days of silence means something very different in early discovery than in late-stage negotiation. The same gap can be neutral or alarming depending on where the deal sits.
• Competitive signal. Is a known competitor in the room, and has the buyer's language shifted since they showed up?
• Who's selling it. A rep's own track record of turning best-case calls into closed deals is one of the most predictive inputs you have, and almost nobody uses it.

None of these are in the stage weight. Yet all of them are knowable from the data you already generate: the calls, the emails, the meeting notes, the engagement history.

Make the number defensible

The fix isn't a better static weight. It's a score that adapts to the deal's actual context. The stage is a starting point, then adjusted up or down by the signals above, then scaled by the reliability of the person forecasting it.

The discipline that makes this work is auditability. A weighted-stage forecast can't be questioned, because there's nothing underneath it to question. A contextual score should open up on one click. Here's the base for the stage. Here's the credit for reaching the economic buyer, the penalty for nineteen days of silence, the rep's multiplier and why. When a number can show its work, managers stop arguing about gut feel in pipeline reviews. And reps stop gaming a system that now reads evidence instead of stage position.

You'll forecast a smaller number, and it will be a truer one. The deals the old math was hiding (the stage-four ghosts that contributed full value while going nowhere) finally show up while there's still time to do something about them.

The weighted-stage forecast survives because it's easy and it's always been there. Neither is a reason to keep trusting it with your quarter.

Related questions

• Why are my B2B sales forecasts always wrong? The 90-second answer to the question this post unpacks.
• How do you measure progress before deals close? What we score that moves before revenue does.
• Why do most sales calls go uncoached? A related execution failure: 94% of calls live unwatched. Deep dive.
• How do I get my reps to actually follow the playbook? A real signals-based forecast depends on playbook execution. Deep dive on why playbooks die in three months.
• Why multi-threaded enterprise deals win 34% more often Stakeholder coverage is the single strongest leading signal a real forecast can read.
• Three sales tools your team could actually cut this year Standalone forecasting tools visualise bad inputs. The cut order, and what reads the inputs cleanly.
• The 2010s B2B sales stack is falling apart Why the instrumentation tools that were supposed to fix forecasting never could.

Or score a real deal of your own with the same model in two minutes.