Three people will tell you what last quarter closed. They won’t agree.
The VP of Sales gives you the number he committed to on the forecast call. The finance manager gives you the figure her model recognizes as revenue. Roll up what the agencies reported to the factory, and you get a third. Same quarter. Same projects. Three numbers.
None of them is wrong. None of them is the same.
The instinct is to blame accounting, sandbagging, or a CRM that needs cleaning. It’s none of those. The problem is that the word won means a different event in each of the three rooms. The VP counted it verbally. Finance counts it as recognized revenue. The agency counts it when the order ships and the commission is earned, which, on a project closeout, can be a full quarter later.
Nobody decided this. The word drifted, and every number built on top of it drifted with it.
When the definitions drift, the dashboard doesn’t lie loudly. It lies quietly, in a language everyone in the room still believes they share.
The Drift Nobody Decided to Install
Drift doesn’t happen by design. It accumulates. A rep starts calling an introduction a “qualified lead” because the quarter is thin and the word is forgiving. A regional leaves dead projects in the active pipeline because removing them looks like shrinking the funnel. The factory logs a job as won the day the PO clears the distributor, while the agency that fought for the spec won’t count it as won until the order ships.
Each person is being honest. Each is using a shared word to describe an unshared event. No dashboard reconciles the gap.
The data bears this out. Gartner’s State of Sales Analytics research found that 47 percent of sales leaders cited inconsistent definitions of metrics and KPIs as a primary barrier to data quality, making it one of the most common reasons different functions can’t share data or align on analytics output. That’s not a technology problem. It’s a vocabulary problem. And the mechanism doesn’t need a survey to be visible: anyone who has watched two regionals describe the same deal in two different pipeline stages has already seen it.
A forecast is only ever as honest as the least-defined word inside it.
Six Words That Need Written Definitions
The fix isn’t another dashboard. It’s six definitions—written down, agreed across the manufacturer and the agencies reporting into it, and anchored to something observable rather than to a salesperson’s read of the room.
| Qualified | A specifier has taken a concrete next step: accepted a cut sheet, requested photometric files, or agreed to a basis-of-design conversation. A meeting taken or a card traded is presence, not qualification. Anchor the word here, and a thin quarter reads as thin. Leave it loose, and coverage reports a pipeline that was never built. |
| Active | The project has a dated next step and a known buyout horizon—a bid date on the calendar, a scheduled action with the distributor or contractor. No date, not active. When this word slips, coverage multiples report three times the quarter, while a third of the projects underneath haven’t moved since spring. |
| At-Risk | A state with triggers, not a feeling, a rep may or may not mention. Triggers include: a competitor named in an addendum, a value engineering signal at the contractor level, a distributor pushing an alternate, or a specification loose enough that several products comply. Without this word doing real work, the lost order arrives as a surprise, when the loss was visible weeks before the buyout. |
| Won | One trigger, agreed across every room: the released purchase order tied to the specified product. Not a verbal commitment. Not recognized revenue. Not the shipped order. Pick one event, name it, and require every party to report against it. When you do, commission disputes stop being arguments about money—because they were always arguments about when a thing happened. |
| Lost | Recorded at the moment of confirmed substitution at buyout, with a reason code tied to the erosion path that caused it—competitor named in addendum, value-engineered out, distributor-pushed alternate, or specification broken. Without recorded losses, spec hold rate cannot be calculated, and a specification-driven business that can’t calculate its spec hold rate is flying without its most important instrument. |
| Dormant | No dated activity inside a defined window, or a buyout date passed with no order captured. The project moves to dormant, not dead, but no longer propping up the active funnel. Coverage deflates to the truth, and the relationship becomes something to deliberately recycle rather than a number holding up the forecast. |
The common thread: anchor every definition to a buyer’s behavior or a channel event, not to a salesperson’s confidence. Behavior is comparable across reps and across the contract. Confidence is not.
The Honest Objection
The strongest version of the pushback is true. The channel runs on relationships, every project carries nuance, and a good principal reads a job in ways no taxonomy fully captures. Over-define the words, the argument goes, and you’ve handed a green rep a checkbox to game—while stripping something real from the veteran who could feel the deal.
Grant all of it. The objection still misses what a definition actually does.
A definition doesn’t replace the read. It records the read in a form that the next person can act on. The principal who knows a job is soft is still free to know it. The discipline asks only that the softness show up as a missing next step or a triggered flag—so the regional, the factory, and the board are looking at one project instead of three private versions of it.
Judgment survives the discipline. It just stops being unauditable.
The Sixty-Minute Exercise
None of this requires a consultant in the room to begin.
Put your commercial leadership team around a table for an hour. Take the six words one at a time. Before anyone speaks, each person writes a one-line definition on their own—no conferring. Then read them aloud.
The distance between the definitions is the drift. On won and active, it’s almost always wider than the room expects.
Once the gaps are visible, write one definition per word that hangs on an observable event. Then test each against the last ten real deals. A definition that can’t cleanly sort ten jobs whose outcomes you already know isn’t finished.
The exercise takes an hour. The argument it settles has usually been running for years, unspoken, inside every pipeline review.
Why Now
LED economics, controls integration, the migration of supply reps into lighting, and a fast-moving wave of channel M&A are all redrawing who sells what to whom. Definition drift accelerates exactly when markets move—because every new line, new hire, and new combination arrives carrying its own private vocabulary.
A team that never wrote down what “qualified” means won’t agree on it under that kind of motion. A team that did not absorb the change without watching the dashboard quietly part company with the business.
Six words. Written down. Anchored to behavior. Reconciled across every room that reports a number.
Until that work is done, the pipeline review keeps producing a different answer depending on who’s presenting, and the quarter keeps closing at three numbers that each belong to whoever said them.
