Here's the idea. Every B2B SaaS company past Series A is running three forecasts in parallel, and nobody on the operating team can tell you with a straight face that they agree. Finance maintains one. Sales maintains another. The board sees a third, usually a hybrid stitched together the week before the meeting. The three are close enough that nobody panics, far enough apart that nobody trusts the number, and the gap between them is where most of the operational chaos in the next quarter actually lives.
What I would recommend, before we get into the structural fix, is to walk through what each of these forecasts is actually doing, because they look the same on a slide and they are not the same on a spreadsheet.
Forecast 1: What finance is forecasting
Checkpoint GTM finds finance forecasts cash and recognized revenue, not bookings: the deferred-revenue waterfall, renewal book, and billing schedule. This makes finance the most accurate on a six-month horizon and the least accurate on a thirty-day one, because usage expansion, un-notified churn, and last-fortnight deals stay invisible until sales reports them.
Finance is forecasting cash and recognized revenue. Not bookings. Not pipeline. They are answering the CFO's question, which is: what will hit the P&L next quarter, and what will land in the bank account. Their inputs are the deferred-revenue waterfall, the renewal book, the billing schedule against existing contracts, and an assumption about new ARR that they got from sales months ago and have not refreshed.
The thing about finance's forecast is that it is the most accurate of the three on a six-month horizon and the least accurate on a thirty-day one. Deferred revenue is a known waterfall. Renewals that have been paper-signed are basically known. The pieces that are unknown: usage-based expansion, churn that hasn't been formally notified, deals that close in the last two weeks of the quarter, are exactly the pieces that move the number, and finance has no real visibility into them until sales tells them.
Forecast 2: What sales is forecasting
Checkpoint GTM finds sales forecasts weighted bookings off CRM pipeline, call notes, and a manager-judgment overlay. That makes sales the most accurate on a thirty-day horizon and unreliable past sixty days, where pipeline is hope dressed up as pipeline. Sales never speaks to recognized revenue, so bookings and recognized revenue diverge in meetings nobody enjoys.
Sales is forecasting bookings, weighted by stage probability, on a thirty-, sixty-, and ninety-day horizon. Their inputs are the pipeline in CRM, the rep-by-rep call notes, the manager judgment overlay that adjusts every weighted number up or down by some unspoken instinct, and the qualification methodology, MEDDPICC or MEDIC or whatever the team is nominally running.
Sales' forecast is the most accurate on a thirty-day horizon and gets unreliable past sixty. The reason is structural: the stuff that closes in the next month is mostly in a stage where the rep can see it, the buying committee is identified, the legal cycle is in motion. The stuff that's supposed to close two quarters out is hope dressed up as pipeline. And sales' forecast doesn't speak to recognized revenue at all, bookings are not bookings-recognized, and the gap between the two is where finance and sales start arguing in meetings nobody enjoys.
Forecast 3: What the board is forecasting
The board's forecast is the narrative one. It is not a model: it is the founder or the CEO standing in front of seven people who have invested money, telling a story about why the next quarter will look a certain way. The supporting math comes from finance and sales, but the framing comes from whoever is presenting and from what got promised in the last board meeting.
The board's forecast is the least precise but the most consequential. It is the number that gets compared to plan. It is the number that triggers the conversation about extending runway, hiring, or pulling spend. And it is the number that, when missed by 10%, makes the next funding conversation harder. To be fair, board forecasts move in narrative chunks for a reason, boards are not operators and should not be reading the same forecast a sales manager reads. But the slide they see should agree with the system underneath, and most of the time it does not.
Why do your three revenue forecasts never reconcile?
Checkpoint GTM finds the three forecasts never reconcile because nobody owns the gap full-time. Finance, sales, and the CEO each hold a piece part-time, so missed renewals surface in finance before sales, expansion stays invisible until after booking, and the board sees a number no operator would sign off on.
So you have three forecasts: finance's recognized-revenue model, sales' weighted bookings pipeline, and the board's narrative. The reconciliation between them is somebody's job. In most Series A and Series B companies I have worked with, that somebody is the CFO part-time, the head of RevOps part-time if there is one, and the CEO when the gap becomes a fire.
The issue is that none of those people has the cycle time, the authority, or the operating-system view to actually reconcile in real time. The CFO is doing close. The head of RevOps is firefighting CRM hygiene. The CEO finds out about the gap on the call where it becomes a problem. The result is predictable: missed renewals show up in finance before sales sees them. Expansion ARR is invisible to the forecast until after the booking. Deals that the rep called as commit slip into the next quarter and finance finds out on the close report. The board sees a number that nobody on the operating team would have signed off on if they had been asked.
How do you get finance, sales, and the board to forecast one number?
Checkpoint GTM fixes this structurally, not analytically: one operator gets end-to-end ownership of the reconciliation, run weekly instead of quarterly. That operator maps every commit deal to finance's revenue view, runs the stale-pipeline diagnostic before the call, and preps the board roll-up off the same source of truth the operating team uses.
The fix is not analytical. You do not solve this with a better dashboard. You solve it by giving one operator end-to-end ownership of the reconciliation, and you give that operator the cadence and the system to run it weekly instead of quarterly.
What that operator actually does is unglamorous. They sit in the weekly forecast call and they map every commit deal to the finance system's view of what would land in revenue if it closed. They run the stale-pipeline diagnostic before the call so the weighted number is not lying. They reconcile the usage-based expansion forecast against product instrumentation, not against last quarter's growth rate. They prep the board roll-up off the same source of truth the operating team is using, with the narrative layer added on top, not replacing, the numbers.
The thing that matters here is that this person does not own the forecast. Sales still owns the bookings forecast. Finance still owns recognized revenue. The CEO still owns the board narrative. The operator's job is to make sure the three of those agree, and to surface the gap to the right person at the right time before it becomes a number on a slide that nobody is happy about.
The three-step prescription
If you are looking at this and recognizing your own company, the work to fix it falls into three steps. Don't overcomplicate it.
- Pick the operator. Not the CFO. Not the head of sales. Somebody whose only job is reconciliation across the three. Could be a head of RevOps with the right scope, could be a fractional operator brought in for 90 days, could be a director of strategic ops if you are big enough to have one. What matters is that this person sits between finance and sales structurally, not on top of either one.
- Lock the source of truth. Decide what the system-of-record is for each piece: CRM for pipeline, finance system for deferred revenue, product instrumentation for usage signals, and stop reconciling in spreadsheets. The operator's first ninety days should be wiring those three systems together so that the reconciliation is a query, not a Slack thread.
- Run the cadence weekly. Monday morning, the operator pulls the three forecasts, runs the diff, and walks into the forecast call with the gap already mapped. The forecast call stops being a recitation of rep numbers and starts being a conversation about where the three views disagree and what we do about it before the close of the week.
That's the work. None of it is exotic. The reason most companies don't do it is not that they don't see the problem, they see the problem every quarter when the board call gets uncomfortable. The reason they don't do it is that nobody on the existing team has the bandwidth or the structural position to own the reconciliation, and hiring a full director of strategic operations is a year-out decision even when it's the right one. So the gap persists, and the forecast keeps being three different numbers in three different rooms.
What I would recommend, if you can't hire that person tomorrow, is to bring in an embedded operator on a ninety-day engagement to build the architecture and run the first cycle with you. Stand up the reconciliation, document the cadence, train whoever is going to inherit it. The forecast does not become defensible to the board because somebody buys you a tool. It becomes defensible because somebody sits behind it and runs the reconciliation week after week. That is the change.
Sources
- Lemkin, Jason. "Sales Teams Aren't Great at Forecasting. Here's How to Fix That." Harvard Business Review, March 2019. hbr.org
- Sinha, Pranava and Shastri, Sukriti. "How CFOs and CROs Should Be Aligning On Forecasting." SaaStr, 2024. saastr.com
- Poyar, Kyle. "Why Most B2B Forecasts Are Wrong, And What To Do About It." Growth Unhinged, 2024. growthunhinged.com