Most project forecasts I've reviewed are a list of hopes with percentages next to them. Stage three, sixty per cent. Stage four, eighty. The numbers look precise. They're guesses wearing a suit.
It isn't just me saying it. A Gartner survey of sales operations found only 45% of sales leaders and sellers had high confidence in their organisation's forecasting accuracy. That's across sales in general. Stretch the cycle to 18 months, put an EPC contractor, a consultant and an operator between you and the purchase order, and I'd expect that confidence to drop further.
I've sold engineered equipment for thirty years, across fifteen industries. The forecasts that held up were never the ones with the cleverest stage weightings. They were the ones built on evidence I could point at.
The problem with a stage based forecast is that it tracks what you've done. You sent the proposal, so the deal moves to stage three. You had a good meeting, so it moves to stage four.
None of that tells you whether the project is moving. On a long cycle job the buyer's timeline runs on budget approval, FEED, the main contract award and the site schedule. Your proposal date has nothing to do with any of it.
So the first change is simple. Every deal in the forecast carries the buyer's milestones, not yours. When does the budget get signed off? When does the EPC award the package? When does the equipment need to be on site? If you can't answer those three, the deal isn't in the forecast. It's in the pipeline, which is a different list. I wrote about building that list in a pipeline that survives an 18 month sales cycle.
Five things. Each one is something the buyer did, not something I hoped.
If you can't point at what the buyer did, it isn't a forecast. It's a wish list with a probability column.
On 22 Bishopsgate in London I was Global Business Development Director at Brannan, selling pressure gauges and thermometers into a landmark tower. The order landed around August 2019, worth around £150,000.
I didn't forecast that one on a stage. I forecast it on who was saying our name. WSP made the initial decision to specify us. Multiplex, the main contractor, needed confidence to use the product, and being a UK manufacturer with more than 100 years behind it helped. Hotchkiss, the MEP contractor, had to go through Wolseley, our partner, who had the relationship. I made sure every one of them, and the developers, knew exactly who Brannan was.
By the time I asked for the order, I already knew we had it. Several people upstream had told me they were using our products. That's the signal I trust most in any long cycle forecast: other people in the chain saying your name before you ask.
The other thing worth tracking sits above the deal level. Look at your number by channel, year on year.
At Hughes, the Middle East business ran through a distributor and hit roughly the same figure almost every single year. It got to a level and stopped. On paper that looked like a dependable forecast. In reality it was a ceiling. Once I replaced the distributor that wasn't fit for purpose, got close to the EPC contractors and started bidding alongside fire safety players, Middle East sales went up 167% in 2024.
A forecast that's accurate because nothing ever changes isn't good forecasting. It's a warning. I covered the same pattern from the other side in nobody lost that deal to a competitor.
This is the uncomfortable part. A forecast is only as good as the person who owns it. If the number gets built by the MD on a Sunday night from the CRM, or passed up by a distributor, nobody's actually testing it. Gartner points to poor data quality as one of the main contributors to inaccurate forecasts. In my experience that comes down to ownership far more than software.
Someone senior needs to own the deals, the evidence and the call on what goes in. In a lot of the manufacturers I work with, that seat's empty. That's the gap a Fractional Commercial Director fills: someone who owns the pipeline, tests every deal against what the buyer actually did, and tells you the truth about the number.
If your forecast keeps slipping a quarter at a time, book a 20 minute call and we'll go through what you're tracking and what you should be.
Sources: Gartner Says Less Than 50% of Sales Leaders and Sellers Have High Confidence in Forecasting Accuracy, Gartner, 12 February 2020. Checked 18 November 2026.
Qualification is a paid call that gives you an honest yes or no on your market, direct with Scott. The $250 deposit is credited if you go ahead, and refunded in full if you do not.