How to build revenue projections that survive diligence.

Startup revenue projections are useful only when they survive contact with the evidence. This guide builds the framework that does.

Every Series A pitch includes a revenue projection. Most projections break the moment an investor builds their own model. The reason is that founders typically project revenue top-down (we will capture X percent of the market) when investors evaluate bottoms-up (this many customers at this price with these conversion rates). The two methodologies produce different numbers, and the bottoms-up version is the only one that survives diligence.

A bottoms-up revenue model starts with the acquisition funnel and works forward to revenue. The components are: monthly new lead volume by channel, conversion rate from lead to opportunity by channel, conversion rate from opportunity to closed customer by channel, average contract value by segment, expansion rate within existing customers, and gross retention rate.

The math compounds forward. If you generate 1,000 leads per month, convert 5 percent to opportunities, convert 30 percent of opportunities to customers, at $10K ACV, you produce 15 new customers per month worth $150K in new ARR per month. With 90 percent gross retention and 110 percent net retention, the existing base grows by 10 percent annually. After twelve months of compounding, the model produces a specific ARR number that investors can stress-test by adjusting any input.

The discipline of building this model forces every assumption to be explicit. Investors who have built their own bottoms-up models on hundreds of companies can immediately spot when an input is fabricated. The signature of a fabricated input is one that does not have a sourced reference or a data point from the company's own funnel.

The most-fabricated input in startup revenue models is conversion rate. Founders project lead-to-customer conversion rates between 20 and 40 percent, when industry benchmarks for cold leads are typically 1 to 5 percent and benchmarks for inbound leads are typically 5 to 15 percent. The high projected conversion rate produces a revenue number that looks healthy and a CAC number that looks efficient, both of which break when investors apply realistic conversion rates.

The fix is to use actual conversion rates from your existing funnel where available, and conservative industry benchmarks where not. According to HubSpot's 2025 sales benchmark report, the median B2B SaaS company converts 13 percent of marketing-qualified leads to opportunities and 27 percent of opportunities to customers, for a total funnel conversion of approximately 3.5 percent. Any model projecting higher than this for cold or moderately warm leads needs specific evidence to support the projection.

The second common failure is projecting growth from channels that have not yet been proven. A founder with 100 percent of current customers acquired through referrals projects that 60 percent of next-year customers will come from paid acquisition. The implicit claim is that paid acquisition will work at the same unit economics as referrals, which is rarely true.

The fix is to project new channels at substantially worse unit economics than proven channels, and to constrain the share of new channels in year-one projections. A reasonable rule: any channel that has not produced at least ten customers should not contribute more than 20 percent of year-one projected revenue. Channels that have produced zero customers should not appear in the projection at all.

Revenue projections often assume that customers acquired in month one produce revenue in month one. This is false for any company with a sales cycle longer than the billing cycle, or with annual contracts that recognize revenue ratably. A B2B SaaS company with a sixty-day sales cycle and annual prepayment recognizes $1 of revenue in month one for a customer who signed at the end of month one, not the full $10K ACV.

The fix is to model revenue recognition based on the actual contract structure. Annual prepayments produce cash up front but ratable revenue. Multi-year deals require more careful unwinding. Monthly contracts produce revenue close to the booking date. Investors will adjust your projection to reflect these dynamics, and the projection you present should already account for them.

The most-requested addition to founder revenue projections is sensitivity analysis. Investors want to see what happens if conversion rates are 30 percent lower, if churn is 50 percent higher, or if the average contract value is 20 percent below projection. The sensitivity table is small (three to five inputs, three scenarios each) and the output is the range of plausible ARR outcomes at month twelve and month twenty-four.

A founder who presents only the base case signals to investors that they have not considered the downside scenarios. A founder who presents a sensitivity table signals that they understand the risk profile of their own business. The sensitivity table also makes the conversation easier in IC, because partners can see the range and form their own judgment about where the actual outcome will land.

The revenue model should connect directly to the headcount plan, the marketing budget, and the gross margin assumptions. If your model projects $5M new ARR in year one, the marketing budget needs to be sized to generate the lead volume to support that ARR at your projected conversion rates. If you project 15 new customers per month at $10K ACV, the sales team needs to be sized to close 15 customers per month at your projected sales productivity.

Founders who present a revenue projection that is not consistent with the headcount and marketing plan have a credibility problem. The disconnects are easy for investors to spot. The fix is to build the operating plan from the revenue model rather than building them separately and reconciling them.

A revenue projection that survives diligence has three properties. Every assumption is sourced or comes from the company's own data. The sensitivity analysis shows the downside scenarios as well as the base case. The operating plan (headcount, marketing budget, gross margin) is consistent with the revenue projection. Investors can rebuild your model in thirty minutes and arrive at numbers close to yours. If they cannot, the projection is not credible regardless of how detailed it looks.

Bottoms-up structure. Conservative conversion assumptions. Realistic channel mix. Proper revenue recognition. Sensitivity analysis. Operating plan that ties to the revenue model. These six elements distinguish projections that survive diligence from projections that get marked down by investors who build their own version. The work to make a projection defensible is the same work that makes you a better operator, so doing it well has compound value beyond the fundraise. For related diligence preparation, see our startup due diligence checklist and how to present research to investors.