The forecast nobody scores

Four in five sales and finance leaders missed a quarterly forecast in a single year, and more than half missed two or more. That's from Xactly's 2024 benchmark survey of 405 North American sales and finance leaders.

A separate Gong survey of 2,015 US and UK business leaders landed in the same place. 81% of their companies missed a sales forecast in at least one quarter between early 2021 and late 2023. Only 19% hadn't missed once in seven quarters.

So the miss is normal. Markets move. Deals slip. A buyer goes quiet in week eleven. No forecasting method removes that.

What shouldn't be normal is what happens after.

Used every week, scored almost never

Most leadership teams I've worked with used the forecast every week and almost never scored it once the quarter closed. The call ends, the quarter ends, the number either landed or didn't, and the next quarter's forecast starts from a blank page.

That's strange when you think about it. We score reps on attainment. We score marketing on pipeline sourced. The one number that drives hiring plans, spend and board conversations often has no scorecard at all.

And a missed forecast isn't a spreadsheet problem. In Gong's survey, 42% of US respondents said they'd frozen hiring because of missed forecasts, and 28% of all respondents said they'd had to let people go.

The trap: the total can land while everything inside it misses

I owned BI and analytics for more than $850M in revenue across four regions. The question that changed how I ran it was a boring one: how accurate were we last quarter, and where exactly were we wrong?

The honest answer is usually uncomfortable, because the total hides the errors. Say two regions of the same size land 20% over and 20% under. The roll-up reads "on target." Leadership sees a clean number. Nobody sees that the three biggest deals in commit all moved.

A forecast that hits in total and misses on the deals that mattered isn't accurate. It's lucky. Luck doesn't repeat, and it teaches the organization nothing.

Sales and finance often don't even describe the miss the same way. In Xactly's survey, 52% of sales leaders said their forecasts miss by 10% or more, while 66% of finance leaders said theirs usually land within 9%. Two functions, two forecasts, two accuracy stories, and in my experience nobody reconciles them until the quarter is gone.

Three numbers, every quarter

Scoring a forecast doesn't need a new tool. It needs three numbers that you calculate the same way every quarter and put on the first slide of the next forecast call.

1. Total error. The forecast you called in week one against what actually closed. One number, no adjectives. Pick the snapshot date once and never change it, or the number becomes negotiable.

2. Error by region or segment. This is where a lucky quarter shows itself. If the total landed because two regions missed in opposite directions, you'll see it here, and you'll know which part of the forecast process needs work.

3. Commit-to-close. Of the deals in commit at the start of the quarter, how many closed that quarter? This number tells you whether "commit" means anything. If a third of commit slips every quarter, the category is a feeling, not a forecast.

Track all three for four quarters and you'll learn more about your forecast than a new model could tell you. You'll also have something most teams don't: a baseline. Without one, "we're getting better at forecasting" is a claim nobody can check.

Why this lands on RevOps

A few weeks ago I wrote that the RevOps seat is shifting from scorekeeper to operator: the function that used to measure the number is now asked to own it. The forecast is where that shift shows up first.

The function that used to build the forecast now answers for it. When the forecast misses, leadership doesn't ask for a better dashboard. It asks why the number was wrong and what you're doing about it. The only good answer to that question starts with a scored history.

If you're stepping into a bigger RevOps seat, this is the first thing I'd build. It's cheap, it's fast, and it moves the forecast conversation from opinions to evidence in a single quarter.

If you own the forecast: could you say today how accurate it was last quarter, deal by deal, and not just in total?