I built my first revenue forecast the way most founders do: I picked a market size, guessed at a capture rate, and backed into a number that looked good in a fundraising deck. It was directionally useless the moment I tried to use it to decide whether to hire a second AE.
That's the mistake underneath most bad startup forecasts: using the wrong method for the decision you're actually making. Top-down and bottom-up forecasting aren't competing philosophies, they answer different questions, and using the wrong one is how founders end up either underselling their startup to investors or overhiring against a number their pipeline can't support.
The two methods answer different questions
Top-down forecasting starts with a market number, your total addressable market or a company-wide revenue target, and works inward to a capture assumption: "if we win 1% of a $2B market, that's $20M." It answers the question "how big could this get," and it's genuinely useful for that question. It's fast to build, doesn't require any deal history, and gives investors the market-size narrative they're listening for during a raise.
Bottom-up forecasting starts from the ground up: open pipeline, weighted by stage, multiplied by realistic conversion rates. It answers a completely different question: "what will actually close this quarter." It's slower to build and needs real pipeline data to be worth anything, but it's the only one of the two that should ever inform a hiring plan, a spend commitment, or a board update on near-term revenue.
Confusing the two is the actual failure mode. A top-down number dressed up as a near-term forecast is how a founder ends up hiring three reps against a $2M "projection" that was really a market-share assumption, not a pipeline commitment.
The 3-question test for which one you need
Run this before you build either model.
- Who reads this number? A board deck or an investor update can carry a top-down market story. A headcount plan or a cash runway model cannot, it needs bottom-up pipeline math or it will be wrong in the direction that costs you money.
- Do you have 90 days of deal history? Under that, your stage win rates aren't statistically real yet, so a bottom-up model is really a bottom-up-shaped guess. Use top-down for anything beyond the current quarter and bottom-up only for the pipeline you can already see.
- What decision does this number unlock? If the answer is "whether to sign a lease, hire a rep, or extend an offer," you need the granular, deal-level accuracy that only a bottom-up forecast provides. If the answer is "whether this market is big enough to raise a Series A into," top-down is not just acceptable, it's the right tool.
A worked example: same startup, two numbers, two different answers
Take a startup selling a $6,000/year tool into a $3B addressable market. Top-down: capture 0.5% of that market over five years and you're at $15M ARR, a fine story for a seed deck. Bottom-up, this quarter: 24 open deals, weighted by stage win rate, comes out to $180,000 in expected new ARR.
Both numbers are correct. They're just not interchangeable. The $15M story justifies raising a round. The $180,000 pipeline number is what should actually determine whether you can afford to add a third AE this quarter. Founders who hire off the top-down number are hiring off a market opinion, not a revenue commitment.
Why the strongest forecasts use both, on purpose
Companies that pair both methods rather than picking one are meaningfully more likely to consistently hit their revenue targets than teams relying on either one alone. The reason isn't magic, it's that each method catches the other's blind spot: top-down alone hides how thin your actual pipeline is, and bottom-up alone can't tell you if you're chasing a market too small to matter.
In practice, that means running two numbers side by side in every board deck: a top-down market-share line for the multi-year story, and a bottom-up weighted-pipeline number, built the way we've laid out in how to forecast sales with no historical data, for the number you're actually accountable to this quarter. Never let the first one substitute for the second when a hiring or spending decision is on the table, since we've also covered what a bad sales forecast actually costs once you've spent against a number pipeline can't support.
What to do this week
Pull up your next board deck or hiring plan and check which forecast is doing the talking. If a top-down market number is quietly standing in for a near-term revenue commitment, replace it with a weighted-pipeline number before anyone signs an offer letter against it. Keep the top-down story for the parts of the deck that are actually about the market, not about what closes next quarter.
Frequently asked questions
Which forecasting method should a pre-revenue startup use?
Top-down, out of necessity, since there's no closed-deal history to weight a bottom-up model against. Treat the number as a market-opportunity estimate for investors, not a near-term revenue commitment, and switch to bottom-up the moment you have real pipeline.
Can I use a top-down forecast to plan hiring?
No. Top-down forecasts describe market opportunity, not committed near-term revenue. Hiring decisions should be sized to a bottom-up, weighted-pipeline number, and ideally the low end of that number's range.
How much pipeline history do I need before bottom-up forecasting is reliable?
Roughly 15 to 20 closed deals to calculate real stage-level win rates. Below that, use public seed-stage benchmarks as a placeholder and replace them with your own data as deals close.
Do investors expect a top-down or bottom-up forecast?
Both, at different points in the deck. A top-down market-size narrative shows the size of the prize; a bottom-up pipeline number shows investors you actually understand your near-term growth levers instead of just the market.
Neither method is wrong. The mistake is letting the wrong one answer the wrong question, especially the one that decides whether you can afford the next hire.