How to Audit a Sales Forecast Before It Goes to the Board

A forecast that reaches the board unaudited is a set of rep opinions with a company logo on it. The audit is the pass that separates what the data supports from what the organization hopes. It takes about two hours a quarter once the queries exist, and it catches the errors that cost credibility.

Run it against an archived snapshot, not a live CRM view. Deals moving underneath you while you work make the findings unreproducible.

What is a forecast audit checking for?

Seven specific failure modes, each with a query behind it. The audit is not a judgment call about whether the number feels right. Work the list in order. Stale record removal changes the denominators for everything after it, so doing it last means redoing the rest.

How do you strip stale pipeline correctly?

Define a touch as a change in stage, close date, or amount, then exclude anything untouched for twelve months. Logged calls and emails do not qualify. Activity logging is a rep behavior, and it drifts toward whatever the CRM dashboard rewards.

Across ORM customers, more than 10 percent of open pipeline sits in this bucket. Those records inflate every ratio built on top of them, including the coverage number that leadership uses as a comfort signal. Removing them usually moves coverage by enough to change the conversation, which is exactly why the check goes first.

The twelve-month rule works for most businesses. If your average cycle runs much longer, calibrate against your own close curves rather than borrowing the default. Opportunity groups carry different close curves, and while a few extend beyond a year, most of the closing expectation lands well inside week twelve.

How do you test for value inflation?

Compare average pipeline deal size against average closed-won deal size for the same segment and the same period. One query, and it explains a large share of most historical misses.

A pipeline carrying an $80,000 average deal size that produces closed-won deals averaging $40,000 is telling you every weighted total is inflated by a factor you can compute. Apply that ratio to the carry-over portion of the forecast before anyone presents it. This is not conservatism, it is correcting an arithmetic error that has been running for years.

Watch the spread over time as well as the level. A widening gap between pipeline value and closed value is one of the earliest signs of pricing pressure, usually from a new competitor, and it shows up in this ratio before it shows up in win rate. Our weighted pipeline guide covers how to fold the correction into the model.

What does the concentration check catch?

A quarter that depends on one deal. Set a concentration threshold and hold to it. Fifteen percent of the number is a workable starting point, and it should be set against your own deal size distribution. Any single opportunity above the threshold turns the quarterly number into a binary bet, and the board deserves to know that before the quarter rather than after.

Report the forecast twice. Once as submitted, once with the largest deal removed. If the second number is materially below plan, the concentration is the story, and the mitigation plan is what the board should be discussing.

The same check applies at the rep and segment level. A segment whose number rests on two deals owned by one rep who is also new to the territory is a different risk than the same dollar amount spread across twenty deals.

How should slipped and silent deals be treated?

Slipped deals get downgraded unless someone re-evidences them. Silent deals get downgraded automatically.

The strongest available signal that a deal is at risk is the rep changing the close date. Once a deal slips from one quarter to the next it is less likely to close, even when it stays in commit. A commit that has already slipped once should not carry the same weight as a commit that has held its date since creation, and the audit should flag every one of them.

The earliest signal is quieter and easier to miss. It is the absence of any signal at all. No stage movement, no amount change, no reply to email, no meeting booked. From a seller's side, a buyer who stopped returning calls is a bad sign long before anything in the CRM changes. Query for committed deals with no meaningful activity inside the silence window you set, and require evidence or a downgrade. Fourteen days is a workable starting point, and it should be calibrated against your own cycle length. See deal slippage for the detection logic.

Why does composition belong in the audit?

Because a forecast total hides how the quarter has to happen. Split it into three sources and report each: deals already in pipeline on day one expected to close this quarter, deals that will be created and closed inside the quarter, and deals pulled forward from future periods.

Pull-forward is the line boards care most about and see least often. Accelerating future deals into the current quarter usually costs discount and borrows revenue from the next period. A quarter that hit plan on the back of significant pull-forward is a different result than a quarter that hit plan on in-quarter execution, and the following quarter's forecast needs to know which one happened.

Reporting composition also protects you. Around 20 percent of the pipeline carrying in-quarter close dates on day one actually closes in that quarter. Presenting a total without composition invites the board to assume the visible pipeline is the plan, which sets an expectation the data does not support.

What do you hand over at the end?

Findings, not edits. The audit produces a short document: which records were excluded and why, the value inflation ratio applied, the concentration exposure, the count of slipped and silent commits, and the composition split. Sales leadership decides what moves.

Keep the separation. An auditor who can change the number is a second forecaster, and you lose the independent read that made the exercise worthwhile.

For a reference point on what the audited number should be capable of, accuracy around 90 percent on new and expansion business is typical of a heavy manual process, and it degrades as conditions shift. ORM targets 95 percent without manual adjustment and holds from day one to day ninety of the quarter. An audit will not get you there on its own, but it will tell you which of your inputs is standing between you and that range. Start with forecast accuracy and pipeline coverage for the definitions the audit relies on.