Most sales forecasts are not predictions. They are hopes with a spreadsheet wrapped around them, and everyone in the room quietly knows it.
The test of a forecast is simple: does the number hold? If the quarter keeps landing 20 or 30 percent below what you called, that is not bad luck. It is a forecast built on optimism instead of evidence. For a founder-led or small B2B tech team, and especially one selling offshore where you cannot read the room in person, a forecast you can trust is the difference between hiring with confidence and guessing.
You are not alone in distrusting the number. Gartner found that fewer than half of sales leaders and sellers have high confidence in their own forecasting accuracy (Gartner, February 2020). And the bar for good is higher than most teams assume: Fullcast puts an acceptable B2B forecast at 80 to 85 percent accuracy, good at 85 to 95 percent, and world-class at 95 percent or better (Fullcast, November 2025). If you have never measured yours, assume it sits below that floor.
The fix is not a new tool. It is six habits.
1. Define stages by what the buyer did, not how you feel
A deal is not "50 percent done" because you like the champion. Each stage should be defined by a buyer action you can point to: they took the demo, they shared it internally, they gave you access to the person who signs, they asked for pricing in writing.
If you cannot name the action, the deal has not reached the stage. This one change strips most of the fiction out of a pipeline before you do any maths.
2. Weight with real conversion rates, not gut feel
Once stages are defined by action, you can measure how often each one actually converts. If deals at "pricing requested" have closed 40 percent of the time across your last 20 deals, then 40 percent is your weighting, not the 70 percent you feel today.
Ten deals at a real 40 percent rate is four expected wins, not the seven optimism wants to book. No history yet? Start recording it now, and use conservative placeholders until the real rates arrive.
3. Separate commit, best-case and total pipeline
One number hides three very different things. Commit is what you would stake your reputation on. Best-case is what happens if the maybes break your way. Total pipeline is everything, most of which will not close this period.
Report all three, and never let total pipeline masquerade as the forecast. Boards and offshore investors trust a team that shows the range, not one that quotes a single hero number.
4. Inspect deal age and slippage, not just the headline value
A NZ$200k deal that has sat in the same stage for four months is not worth NZ$200k. It is a warning. Age and movement tell you more than value ever will.
Track how long deals sit and how often they slip to the next quarter. If the same deals keep sliding, they were never commit deals, and every slide quietly widens the gap between your forecast and reality.
5. Every deal needs a single next step and a real close date
A deal with no scheduled next step is not active, it is a wish. A close date that always happens to be the last day of the quarter is not a date, it is a placeholder.
Make the rule simple: no confirmed next step and no realistic, buyer-agreed close date, and the deal does not count in the forecast. It can still live in total pipeline, but it stays out of the number you commit to.
6. The founder's job is to kill weak deals early
The instinct is to carry every deal, because each one is hope. But the discipline that fixes accuracy is subtraction: kill the deals that fail the tests above quickly, so your energy goes to the ones that can actually close.
The cost of skipping this is real. Even after cutting quotas by 13 percent, 77 percent of sellers still missed their number (Fullcast, November 2025), usually because the pipeline was never as solid as it looked on the board.
So the real question is not how big your pipeline is. It is how much of it you would actually bet on. How much of yours would survive that test?
How predictable is your revenue engine, really?
Start with my free Growth Scorecard - about 6 minutes, self-scored - to see where your revenue engine is strong and where it leaks, from first response to close. And when you are weighing a move offshore, the Global Growth OS runs eleven expansion diagnostics across your business, including how ready your pipeline and process are to travel.
Get your free scorecard → See the eleven Global Growth OS diagnostics →Sources. Forecast confidence: fewer than 50 percent of sales leaders and sellers have high confidence in their forecasting accuracy (Gartner, "Gartner Says Less Than 50% of Sales Leaders and Sellers Have High Confidence in Forecasting Accuracy," gartner.com, 12 February 2020). Accuracy benchmarks and quota data: an acceptable B2B forecast runs 80 to 85 percent accurate, good 85 to 95 percent, and world-class 95 percent or better; and 77 percent of sellers missed their number even after quotas were lowered by 13.3 percent (Fullcast, "Forecast Accuracy Benchmarks," fullcast.com, 20 November 2025). The conversion example (ten deals at a 40 percent close rate producing four expected wins) is an illustration of the method, not a cited statistic. The stage, slippage and deal-hygiene practices draw on my own experience helping NZ tech companies build forecasts their boards trust, not a single study.