Most marketing teams build forecasts in two ways. They take last quarter's number and add 20%. Or they wait until the CFO asks, then panic.
Neither is a forecast. Both are aspirations.
A real forecast: marketing-attributed pipeline by source × close rate by stage × historical conversion timing × budget-to-pipeline yield curve. Built monthly. Reconciled to the variance from the prior month. Updated when something material changes — channel shift, ICP narrowing, new sales hire. It is the output of a real marketing analytics practice, not a slide built the night before the board meeting.
The kind of forecast a CFO can drop into a board deck without rebuilding the math. The kind a Head of Sales actually trusts when they're staffing for the quarter. It is also the planning layer companies need to actually scale to $10M without re-forecasting from scratch every quarter.
How It Actually Works
Pipeline forecast by source
Channel-by-channel attribution × close rate × historical timing. LinkedIn produces $X in pipeline at Y% close rate over Z days. Google produces a different shape. Inbound is its own curve.
Updated monthly, not quarterly. Reconciled to actuals before the next forecast goes out. The forecast for May reflects what April actually shipped, not what April was supposed to ship — the same weekly optimization rhythm applied at the planning layer.
Scenario modeling
Best case, base case, floor case across budget allocations. So you know what's actually at risk if you cut a channel. So you know what "break the quarter" looks like before it happens.
The CFO needs three numbers to decide budget: the floor, the base, and the upside if the bets pay off. Most marketing teams give the CFO one number, in confidence, and miss by 30% in either direction. Three scenarios is the minimum.
Budget-to-pipeline yield curve
How much pipeline does $1 of paid spend produce, by channel, by ICP segment? If you can't answer that, you're guessing — and most marketing teams are guessing.
The yield curve is built from 12+ months of attribution data. It exposes saturation points (where adding budget stops producing pipeline), efficiency floors (where the channel won't perform regardless of spend), and the right blend across channels at any given budget level.
Monthly variance analysis
What did the forecast predict? What actually happened? What changed? Built into the rhythm — not bolted on after a board meeting where the forecast missed.
Variance analysis catches the drift early. If May was off by 12% on pipeline-from-LinkedIn, that's data for June's forecast and June's budget allocation — not a story to tell the CFO three months from now.