The Assumptions Hidden in Your Revenue Forecast

Boards are now asking harder questions about revenue forecasts. After two years of miss-and-reset cycles, many CFOs have become skeptical of the numbers they receive. That skepticism is reasonable, but it often lands on the wrong part of the problem.

Most of these weaknesses leave evidence in the CRM well before quarter-end, but they are not visible in the standard board forecast. They surface after the quarter closes, when the post-mortem reconstructs what the data was saying all along.

Three gaps appear most consistently.

When coverage ratios overcount

Boards routinely receive pipeline coverage ratios as a planning input. A 3.2x pipeline-to-quota ratio looks conservative. The problem is that the ratio includes opportunities that should not count.

Across the mid-market SaaS pipeline reviews I have conducted, typically involving several dozen late-stage opportunities at a time, around 20 to 30 percent of nominally active deals showed no recent buyer activity. No meetings scheduled, no emails returned, no documented advancement. Those deals sat in the funnel at their last-touched stage, and nothing in the CRM triggered a flag. The specific field that reveals this is Days in Stage: pull it, sort by longest, and the list of deals that have quietly stopped moving is usually immediate.

At a 30 percent stall rate, a reported 3.2x coverage ratio drops to roughly 2.2x. Boards making capacity, hiring, or investment decisions on the higher number are working from data that has not been pressure-tested.

The adjustment is not complicated: run a quarterly audit requiring documentation of the last verified buyer action on each opportunity above a defined ACV threshold. Remove or re-stage deals with no confirmed advancement in 30 days, or 45 to 60 days for enterprise cycles with longer procurement timelines, defined before the audit rather than applied after the fact. Report both the gross and the adjusted ratio. One practical question always comes up: what counts as verified buyer activity? A booked meeting is clear. A LinkedIn message is not. Consistent enforcement is the harder problem, particularly when a manager has to challenge a rep's read on a deal they both want to close.

When the commit label exceeds the evidence

In one mid-quarter review I ran, a $180K deal had remained in Commit status for almost four weeks. There had been no recorded buyer contact, no updated close date, no mutual action plan. The rep was still relying on a verbal indication from three weeks earlier. The manager knew the deal was thin. Moving it from Commit, though, meant explaining to the VP why it had ever been there in the first place. So it stayed, through the final week of the quarter, and then slipped.

The forecast process collected what it was designed to collect: the rep's read, and the manager's reluctance to have a harder conversation. The issue persists because most forecast processes capture the seller's conclusion without capturing the buyer evidence behind it.

Experienced sellers can often tell the difference between a real signal and an optimistic interpretation, but the standard forecast process rarely asks them to articulate it. They submit forecasts they are not fully confident in, because the process gives them no mechanism to flag the gap between what they believe and what the CRM supports.

A commit call should require some form of buyer-side evidence: a documented conversation with the economic buyer, a confirmed legal timeline, or a mutual action plan with active buyer engagement. When none of those exist, the deal belongs in best-case. Making the assumption visible to the reviewing manager is more useful than a confidence rating that obscures it.

The reconciliation problem between Finance and Sales

Finance and Sales often avoid this conversation until the quarter has closed, when the difference between their numbers becomes impossible to ignore.

Sales reports against ACV. Finance models against recognized revenue, which accounts differently for expansion ARR, multi-year prepays, and deals that reclassify between periods. A deal Sales counts as closed in Q3 may not appear in Finance's Q3 recognized revenue, depending on contract structure and recognition timing.

When the quarter closes below expectations, Sales starts with the submitted forecast, Finance starts with the revenue model, and the board starts with the number it was given. Each team had built their model on different assumptions that nobody made explicit before the quarter started.

A pre-quarter alignment session covering which deals are in which number and which metric each function reports against takes under an hour the first time, and less than 30 minutes once the definitions are settled. Without it, the post-quarter explanation the board receives is a reconstruction, not a report. Finance cannot model what the forecast process does not expose.

What this means for the board

The standard board forecast package does not ask any of these questions. A CFO who does will usually find answers that change the risk picture before the quarter closes.

Dana Therrien's 2016 Forrester study, "The Definitive Way to Measure and Grade Sales Forecast Accuracy," found that 79% of B2B sales organizations miss their forecast by more than 10%. Most boards have absorbed that as a condition of the business. The question worth asking is how much of that variance was preventable.

None of these changes requires a new forecasting platform. They can be introduced through operating definitions, review criteria, and board reporting standards. The goal is not a perfect forecast. It is to know, before the quarter closes, which assumptions the current one is resting on.