Your pipeline is only as honest as your stages

Open a CRM nobody agreed on and you find the same thing every time: a dozen stages, half of them meaning the same deal, and a “status” field fighting the stage for the same job. Three teams are in there too, each certain their version is the real one.

That CRM can’t tell you anything. The data is all there; it just disagrees with itself. Every report averages four definitions, so the output is nonsense carried out to a decimal point.

A stage is a promise, not a bucket

A stage should answer one question: how close is this deal to becoming money? Prospecting, qualified, proposal, negotiation, closed. Each one is a promise that specific things are true and specific things have happened before a deal is allowed to sit there.

When stages stop meaning that, they turn into buckets - somewhere to drop a deal so it’s off your list. “Nurturing” and “on hold” both become graveyards, and the pipeline number the board sees is built partly out of them.

Stage and status are two different questions

People collapse stage and status into one field and then wonder why nothing reconciles. They answer different questions. Stage is where the deal sits on the path to closed. Status is what’s happening to it right now - active, stalled, waiting on the customer, waiting on us.

Salesforce makes the split obvious if you let it. A Lead has a Status: Open, Working, Nurturing, then Qualified or Unqualified. An Opportunity has a Stage, and each stage carries a probability and a forecast category behind it. The field is different and so is the object. A Lead becomes a Contact, an Account and an Opportunity the moment it converts. Treat Lead Status like an Opportunity Stage and you’re pushing two things through one number, and the number ends up reporting neither of them properly.

Nothing lives in a temporary stage forever

Every stage needs a shelf life, a smoke alarm that goes off when a deal has sat there too long.

A lead that’s been “qualified” for ninety days is really a decision nobody got around to making. A deal parked in “proposal” for a quarter is usually a no that hasn’t been said out loud. The temporary stages are the worst offenders, “on hold,” “follow up later,” “nurturing”, because they feel like progress, and they’re where deals go to die.

Put a clock on each stage. When a deal sits longer than the stage should ever take, something moves it: a task fires, or it gets demoted back to reality. The automation itself matters less than what it forces into the open, which is that a stuck deal stops being invisible. A forecast built on deals that haven’t moved in months is optimism with a date on it.

Agree where it came from before you argue who gets credit

Attribution fights are about credit, not data, whatever people claim in the meeting. Marketing says the webinar sourced it. Sales says they’d been working the account for months. It’s the same deal with two stories, and no way to settle it once the deal has closed.

The only fix is to agree the model while the deal is still being created, well before it closes. First touch, last touch, or weighted across the whole path: pick one, and make the CRM record it the same way every time. Salesforce will run multi-touch with Campaign Influence if you set it up. It will also happily let a rep overwrite Lead Source with whatever they typed last. Campaign Influence gives you a model; the overwrite habit gives you a guess wearing a field name. The source has to be captured where it happens, on the Lead, and carried through conversion to the Opportunity, or it’s gone. Miss that, and the deal can’t tell you where it came from, so it can’t tell you what to do more of.

Design it for the people who live in it

None of this holds if the people in the CRM every day don’t recognise it as theirs. A stage model built in a strategy meeting and handed down gets ignored by the reps clicking it forty times a day.

So first you watch how they really work. Where they hesitate, what they skip, which field they fill with junk because it’s mandatory and means nothing to them. Then you design the stages, statuses and source fields so filling them in correctly is the easy path rather than a tax on top of the real job. That’s the difference between a CRM people use and one they work around.

Adoption comes from building a system people don’t need training to use, not from drilling them on a bad one. Everyone in the cycle, marketing, SDRs, AEs, success, has to read the same stage and picture the same thing.

A CRM only starts giving you insight once everyone has agreed on what it’s saying.