Nobody in a pipeline review makes a different decision because activity volume was up 12% this week. As a sales rep, the only question I care about with any number is: would I do something differently if it moved? If the answer is no, it's not a metric, it's decoration.

Before your next pipeline review, run your dashboard through six checks: write a decision next to every number, count qualified pipeline created, strip the placeholder amounts out of pipeline value, split win rate by source, follow one month's deals through the stages, and close out the deals that are already dead.

1. Write a decision next to every number, and delete the blanks

Add a column to the slide. Next to each number, write who does what differently when it moves. Not "we keep an eye on it." A person and an action: the manager coaches a rep, a rep calls a deal, leadership moves budget. If the cell stays empty, the number comes off.

What you'll probably find: the fastest-moving numbers on the slide are mostly the ones nobody acts on. Calls made, emails sent, open rates. They change every day and change nothing.

Meetings booked is the one to fix rather than delete. Time with customers is the point of the job, so count meetings held, and log what each one produced: a new qualified deal, a deal moved a stage, or nothing. A meeting that produced nothing is a calendar event, not a sales outcome.

2. Put qualified pipeline created where the activity counts were

Qualified pipeline created is the number of new deals per rep per week that have passed discovery. It tells you what's coming, and it's the number I watch closest with new reps. Someone can be busy, hit their call quotas, book meetings, and still not be creating real pipeline.

Define qualified and hold everyone to it: discovery has happened, a named person has said the problem out loud, and there's a next step with a date. Count deals, not value, which at this stage is a guess. If a rep drops below their own recent average, talk to them this week, not at quarter end.

What you'll probably find: a rep whose activity is up while their pipeline is flat. That's an aim problem, not an effort problem: the wrong accounts, or the wrong message. Coach the targeting, not the hours: go through last week's accounts and opening line with them.

3. Take the placeholder amounts out of pipeline value

A placeholder amount is the default value a deal gets before anyone knows what it's worth. On plenty of teams, every new deal gets one the day it's created. Add those up and your pipeline total is a sum of guesses: a number that looks reassuring and means nothing. Weighting doesn't rescue it. A placeholder times a stage probability is still a guess, only smaller.

Report value only on qualified deals with a real amount, set after discovery, not the default. Show everything else as a count, so it can't pass for money. The total drops, sometimes a lot. The smaller one is the first pipeline figure you can plan a quarter around.

4. Split win rate by source, because the average hides a failing channel

Win rate by source is your win rate calculated separately for each place deals come from: outbound, inbound, partners, referrals, events, your own network. A blended win rate is an average of a channel that works and one that doesn't. Blending sources into a single win rate is how you keep investing in something that quietly stopped working. For each source, divide deals won by deals won plus lost, and put the deal count next to the rate. Two wins out of three partner deals isn't a 67% win rate, it's three deals. Don't move budget on three deals. Move the hours toward the source that closes.

5. Follow one month's deals to find the stage where they die

Conversion by stage is the share of deals that make it from one stage to the next. Win rate tells you how many deals you lose; conversion by stage tells you where. Take every deal that reached "qualified" in March and follow that group: how many reached "proposal sent", and how many of those closed?

Look for the biggest drop. In my experience, deals that die between qualified and proposal never had real pain, or you were talking to the wrong person. The fix is in discovery: stop letting deals through the qualified gate on a maybe. Deals that die after the proposal usually lose on price, to a competitor, or because nobody on their side could say yes. Find out who signs and what you're up against before the price goes out.

6. Close out stale deals: past close dates and too long in a stage

Run two filters on your open deals. First, every deal with a close date in the past. In HubSpot, Salesforce, Pipedrive or Attio, that's one saved view: close date before today, deal still open. Every deal on it is a number on your dashboard that is wrong today. Second, every deal that has sat in its current stage longer than your won deals usually spent there. Say won deals leave "proposal sent" within two weeks, and one of yours has sat there for six. Call it today, or close it as lost.

Stale CRM data isn't random. It's optimistic.

A deal that moves forward gets logged the same day. A dead one waits for the quarterly cleanup. So a stale dashboard doesn't just blur the numbers, it flatters them. Pipeline looks bigger, and win rate looks better, because the losses haven't been marked yet.

I used to be the person leaving a trail of stale deals behind me, not because I didn't care, but because I'd walk out of four back-to-back calls and the CRM was the sixth thing on my mind, not the first. Updating a deal meant stopping, opening the record, and reconstructing what happened from memory. That's a process problem, not a discipline problem, and asking people to try harder doesn't fix it.

The sales metrics worth keeping, and who acts on each

Here's the whole slide after the six checks: seven numbers, each with the person who acts on it.

KeepWho acts when it moves, and how
Open deals with a past close dateEach rep fixes theirs before the meeting starts
Qualified pipeline created, per rep per weekThe manager talks to any rep below their own average this week
Meetings held, and what each producedThe manager coaches the meetings that end without a next step
Pipeline value on qualified deals onlyLeadership plans the quarter on it
Win rate by source, with deal countsLeadership moves hours and budget toward the source that closes
Conversion by stage, one month's dealsThe team tightens discovery, or finds the signer before pricing
Deals in a stage longer than won deals stayedThe rep calls today, or closes the deal as lost
The ones tied to a decision someone makes when the number moves. For most B2B teams that means qualified pipeline created per rep, meetings held and what they produced, pipeline value on qualified deals only, win rate by source, conversion by stage, and time in stage on open deals. Each one has an owner and an action. Numbers without one belong off the slide.
A vanity metric is a number that moves when people are busy but changes no decision. Calls made, emails sent and open rates are the usual examples, along with pipeline totals padded with placeholder amounts. The quick check: if it moved, who would do what differently? If nobody can answer, it's decoration, however precisely it's tracked.
Only the ones someone acts on: qualified pipeline created per rep per week, pipeline value on qualified deals only, win rate by source with the deal count beside each rate, conversion by stage on one month's deals, and time in stage against how long won deals stayed. Start the review with open deals that have a close date in the past, because each one is a number that is wrong today.
A new deal that has passed your qualification gate: discovery has happened, a named person has said the problem out loud, and there's a next step with a date. Count those deals per rep per week rather than summing their value, and add value only once the amount stops being a placeholder. It is not the total pipeline already sitting in the CRM.
Weekly. Every Monday, before the dashboard goes up, each rep clears open deals with a close date in the past and deals that have sat in a stage longer than won deals usually do. Run the full six checks once to cut the slide down, and again each quarter.
It makes them optimistic. Deals that move forward get logged quickly, while dead ones wait for a cleanup, so stale data inflates pipeline and makes win rate look better because the losses aren't marked yet. Count open deals with a close date in the past: each one is a number that is wrong today. Current data won't rescue a bad metric, but it makes the good ones believable.
Delete the numbers nobody acts on, then make the rest specific: split win rate by source with deal counts, follow one month's deals through the stages to see where they die, and report pipeline value only on qualified deals. Fix stale data first. A dashboard full of past-due close dates flatters every number on it.

Start with one filter this Monday

Don't run all six checks at once. Start with the one that makes the rest worth trusting.

This MondayBefore the dashboard goes up, every rep filters their open deals for a close date in the past. For each one: move the date, move the stage, or close it as lost. Nothing stays on the list.
Next MondayAdd the decision column to the slide. Any number without a name and an action next to it comes off.
Every Monday afterBoth filters from step 6 run before the meeting, not during it.

If clearing that list takes a rep more than a few minutes, your problem isn't the metrics. It's how deals get updated, and no dashboard fixes that.