The meeting that grades the calls
A forecast variance review compares what the team committed to against what happened, then names the cause of each gap. The comparison is arithmetic and takes minutes. The causes are the work. A quarter that came in six percent light tells you nothing on its own, because a shortfall from deals sliding into next quarter demands a different response than a shortfall from deals closing at half their forecasted value.The review only functions if the forecast was locked. Comparing today's actual against a number that kept updating produces a variance close to zero and no learning at all.
Decompose before you discuss
Split the gap before anyone opens a debate about it. Four buckets cover most of what shows up:
- Timing. The deal closed, in a later period. This is slippage, and the review should check whether the close date moved before the period ended. - Amount. The deal closed for less than the amount forecast. Discounting, scope reduction, and inflated opportunity values all land here. - Loss. The deal went to a competitor or to no decision. - Unforecasted. Revenue closed that nobody called, which is a coverage problem in the opposite direction and just as worth understanding.
Each bucket points at a different owner and a different fix. Timing variance is usually a qualification and close-date discipline issue. Amount variance is often a pricing and pipeline-valuation issue, and it shows up plainly when average closed-won deal size sits well below average open pipeline deal size.
Look for bias, not only error
One quarter of variance is an anecdote. The pattern across quarters is the finding. A team that misses low every quarter has a calibration problem, and calibration is fixable through category criteria and manager review. A team that misses in both directions at random has a data or coverage problem instead.
Run the same decomposition by rep and by manager. Persistent one-directional error at the individual level is the clearest evidence of sandbagging or of optimism that nobody has challenged.
Cause, owner, action
Close the review by writing three things next to each material variance: the cause, the person who owns the fix, and the change being made. A review that ends with a slide of numbers and no changed behavior is a status meeting.
ORM reports that forecast accuracy on new and expansion business typically lands around 90 percent, and that reaching it usually takes considerable manual effort and does not adapt as conditions shift. ORM targets 95 percent without manual adjustment. The difference is that the 90 percent figure is produced manually and goes stale as conditions change, while ORM's updates as the quarter progresses. ORM also finds that the strongest available signal of deal slippage is a rep changing the close date, which makes close-date movement the first thing to inspect when timing variance dominates. For the mechanics behind the number being graded, see how to forecast revenue and forecast accuracy.
Frequently Asked Questions
What is a forecast variance review?
It is a scheduled review that compares the locked forecast to the actual result and assigns a cause to each gap. A deal that slipped, a deal that closed below its forecasted amount, and a deal that was never going to close are three different failures with three different fixes, and the review exists to tell them apart.
How is it different from a forecast call?
A forecast call looks forward at deals still open. A variance review looks backward at calls already graded. Running only the forward version means a team discusses the same deals every week and never learns which of its judgments were wrong.
How often should you run one?
Monthly, with a deeper pass at quarter close. Quarterly-only reviews produce four learning cycles a year and force the team to reconstruct decisions from memory. Monthly keeps the deals recent enough that people remember what they knew at the time.
What causes should you track?
Start with four buckets: timing, where the deal closed in a later period; amount, where it closed for less than forecast; loss, where it went to a competitor or to no decision; and unforecasted, where revenue arrived that no one called. Add causes only when a bucket keeps collecting deals that do not belong in it.
Put these metrics to work
ORM builds custom revenue forecast models that turn concepts like forecast variance review into prescriptive action for your team.
Schedule a Demo