Is the forecast real?

A forecast is a stack of claims

The number in the forecast cell looks like math. It is actually a pile of assertions: this person can buy. That conversation happened. The budget exists. Nothing has changed since someone last checked. The roll-up adds the assertions together and hands you the sum, with the evidence stripped off.

So “is the forecast real?” is never answered by re-reading the roll-up, adjusting the weightings, or asking reps to re-affirm their numbers in a bigger meeting. It is answered by auditing the claims — and the claims live in exactly two places: the people around each deal, and the record that describes it.

---

The commit call. A deal sits at 90%, closing in twelve days, a number everyone in the room has repeated to someone above them. It has one contact role filled.

Nobody in the room has met the budget owner.

1. What the percentage claims

What is the number in the cell actually made of?

Two failures hide inside every fake forecast number, and they compound.

The first is a people failure: the deal is single-threaded. One enthusiastic contact stands in for a buying committee that was never mapped, so the “90%” is really one person’s optimism, weighted by nothing. If that person goes quiet, changes jobs, or turns out to hold no budget, the deal doesn’t slip — it evaporates.

The second is a record failure: the CRM says things nobody verified. The stage was advanced after a good meeting; the close date was set to fit the quarter; the “budget confirmed” note is old enough that it’s really a rumor with a timestamp.

Either failure alone distorts a deal. Together they make the number unfalsifiable — the committee that could contradict the record was never engaged, and the record that could expose the gap was never checked. That is why forecast inspection has to test both. A committee test without a record test trusts stale facts; a record test without a committee test trusts one voice.

2. The committee test

Who has to agree before this closes — and how many have you met?

Every deal of size has a buying group: someone who owns the problem, someone who owns the budget, someone who can veto on security or legal, people who will live with the choice. The committee exists whether or not your CRM knows about it — the only question is whether you found it before or after it voted.

So the test is blunt. For each committed deal: who are the people this purchase actually requires, and with how many of them has a two-way conversation happened? Not identified. Not emailed. Engaged — they have discussed the initiative with you, in their own words.

Be suspicious of the word champion while you run it. A champion is built and confirmed, not declared; the first friendly voice in an account is a contact, and calling them a champion is how one relationship gets mistaken for a committee.

3. The record test

When did a human last verify this deal?

Freshness is not recency. A note added yesterday can restate a three-month-old assumption; a six-week-old note can still be the last time anyone actually confirmed anything. The question is not “when was this record touched?” but “when did a person last verify its load-bearing claims in a conversation?”

Three fields carry most of the weight: the close date (whose date is it — the buyer’s, or the quarter’s?), the stage (advanced on evidence, or on momentum?), and the economic claim (who said there’s budget, to whom, when?). If the answer to “when was that last true-checked” is measured in months, the field isn’t data. It’s sediment.

Illustrative deal

The deal at 90%

A commit-week audit of one forecast row — the two tests run on the same deal.

• Stage90%
• Closes in12 days
• Value$180K
• Contact roles filled1

Operations — owns the problem

• Director of OperationsEngaged — "the champion"
• VP Operations · decision makerIdentified

Finance — owns the budget

• CFO · budget ownerNamed, never met

IT & end users

• Security lead · can vetoIdentified
• Two team leads · will live with itRoles unknown

Relevant buyers engaged

1 of 6

Functions engaged

1 of 4

Budget owner

Never met

Last verified by a conversation

47 days ago

4. Audit one deal

What does auditing one row take?

About ten minutes, once the two tests are questions instead of instincts. Read the deal above the way its buyer would: a real initiative, a genuinely enthusiastic Director — and a 90% built entirely on that one relationship, resting on facts nobody has checked in seven weeks.

Notice what the audit did not conclude: kill the deal. The initiative may be real and winnable. The audit converts a false certainty into two named moves — earn the VP conversation, get introduced to the CFO — and the forecast conversation stops being about a percentage and starts being about closing: who else has to believe, and what has to be verified, before this number deserves the room’s trust.

Run that on every committed row and the roll-up changes character. Deals don’t get smaller; claims get named. Some numbers go down and — just as often — a real deal that looked thin gets the second thread that saves it.

5. The ratios that survive

Which measures predict the commit?

You don’t need a scorecard with nine tabs. Four ratios, run only on committed deals, catch most of the fiction:

• Buyer penetration: relevant buyers engaged ÷ relevant buyers identified, per deal. The single-thread detector.
• Functional coverage: functions with at least one engaged buyer ÷ functions the decision requires. Enthusiasm in one department is not a deal.
• Decision access: is the person who can actually say yes identified and engaged — and if not, what introduction is missing?
• Record freshness: days since a human verified the close date, stage, and budget claim in a conversation.

None of these has a universal target; the right buying group depends on the initiative. Their value is direction and honesty — a committed deal whose four numbers you can’t produce isn’t committed. It’s hoped.

6. Audit with a spreadsheet

Can you run the truth test without a system?

Yes. One sheet, one row per committed deal, the four ratios as columns, filled by hand before each forecast call. The first pass is uncomfortable and worth it: the sheet makes the missing conversations undeniable, and the forecast meeting starts producing next moves instead of re-affirmations.

Then the limit arrives on schedule. Buying groups shift, verification dates age the moment they’re typed, and hand-auditing thirty deals weekly is a job nobody was hired for. The truth decays faster than a spreadsheet gets updated.

Systematize the evidence, keep the judgment. Narrative AI maintains the buying-group map and flags what’s stale, and its CRM suggestions come with the source and reason attached — the rep approves every AI-suggested change before anything writes, and every message a seller sends stays rep-initiated. What reaches your forecast call is the same roll-up, standing on claims somebody actually checked.

The percentage was never the forecast. The claims underneath it are.