Our lead investor asked one question in the diligence call that stopped the round cold: "Walk me through your last three forecasts versus what you actually closed." I didn't have an answer, because I'd never tracked it.
For two quarters I'd been telling our board a single number pulled straight from gut feel: total pipeline value, rounded down a little to sound cautious. It always came out confident. It was also wrong almost every time, and the gap between what I promised and what we actually closed had quietly become the biggest risk sitting in our Series A data room.
The forecast that looked fine until someone asked to see the math
Going into that quarter, our pipeline showed $340,000 in open deals. I forecasted $220,000 closing, because that felt like a reasonable haircut. We closed $95,000. The quarter before that wasn't much better. The investor doing diligence pulled both numbers side by side and asked the obvious question: why was I consistently overconfident by more than double?
The honest answer was that I'd never separated deals by how real they were. A prospect who'd taken one discovery call counted the same as a prospect with a verbal commitment and a signed pilot agreement. Both sat in the "pipeline" column at full value. Worse, deals that had gone quiet for two or three months were still sitting in the total, because closing them out as lost felt like admitting failure, so I just left them there, inflating every forecast that came after.
The three days I spent rebuilding it live, mid-diligence
With the term sheet stalled on this exact question, I didn't have the luxury of researching the "right" way to forecast. I built something crude, fast, and honest instead.
- Listed every open deal and assigned it an honest current stage: discovery, qualified, proposal sent, verbal commit.
- Pulled win rate per stage from our 11 closed deals to date, thin data, but real data. Where a stage had too few closed deals to trust, I used a conservative published seed-stage benchmark instead and flagged it as such.
- Multiplied the value at each stage by that stage's win rate, then summed the results instead of using the raw pipeline total.
- Marked anything untouched for 60+ days as stalled and pulled it out of the active total entirely.
The new number: $118,000, against a raw pipeline that still read $340,000. That is a brutal gap to show an investor mid-diligence. But it was the truth, and more importantly, it came with a method attached instead of a vibe.
Why showing the method mattered more than the number
No investor doing seed or Series A diligence expects an early-stage forecast to be accurate. Pipelines are too thin and deals are too lumpy for precision. What they're actually testing is whether the founder understands their own number well enough to know when it's wrong, and has a system that gets more honest over time instead of one that just gets repeated with more confidence each quarter.
Once I walked the investor through the stage-weighted math instead of a single confident figure, the conversation changed completely. It wasn't about defending $220,000 anymore. It was about showing that $118,000 was a real, defensible floor, and that the system producing it would only get sharper as we closed more deals and replaced the benchmark win rates with our own. The term sheet moved forward two days later, with one added condition: a weekly one-page pipeline breakdown by stage, sent through the rest of diligence.
What changed after the round closed
The habit didn't stop once the wire hit. We kept the weekly pipeline review and added one rule: no rep could report a deal as "committed" without pointing to the piece of evidence that earned that stage, a signed order form draft, a Slack thread with the buyer's legal team, not just "they sounded excited." Six months later, our forecast-to-actual gap had narrowed from being off by more than 100% to being off by roughly 15-20%, which is close to what most Series A investors consider normal for an early-stage pipeline.
The bigger shift was in hiring. Because the weighted forecast was now something the board trusted, it also became the number we used to decide when to add a second sales hire, instead of guessing off a headcount plan. A forecast that's honest about its own uncertainty turns out to be useful for a lot more than the board deck.
What I'd tell any founder before this question lands on you
This isn't a fundraising problem. It's a habit you either have or don't have well before a term sheet is on the table.
- Even a rough spreadsheet works. The pattern across quarters is what investors ask about, not any single miss.
- Even with a handful of closed deals to calculate rough win rates from, a modest weighted number beats a confident raw one.
- If nothing has moved in 60 days, pull it out of the active forecast. Leaving it in only makes the next quarter's miss look worse.
- Handing it over unprompted signals more discipline than any line on a pitch deck can.
The math itself took me an afternoon to build once I actually sat down and did it. The three months of forecasts I'd given without it are what nearly cost us the round. Build the discipline now, while the only person grading your forecast is you.