What I've Learned Reviewing 50+ Startup Financial Models as an Investor
Over the course of reviewing deal flow and working directly with portfolio companies on their financial infrastructure, I have looked at more than fifty financial models from early-stage B2B founders. Some were built in Excel with multiple interconnected sheets and complex scenario logic. Others were single-tab Google Sheets with a few years of revenue projections. The quality of the tool matters less than people assume. What I have found is that the patterns separating the models that instil confidence from the ones that raise concern are consistent, predictable, and fixable.
The models that work have one thing in common
The financial models that build genuine investor confidence all share a single quality: you can trace every number back to an activity or an assumption. The revenue growth is not just a percentage, it is the product of specific customer acquisition activities at specific conversion rates. The burn rate is not a lump sum, it is the sum of specific hires, specific tools, and specific spend commitments, each with an attached purpose.
When I can trace a number back to its source, I can evaluate whether the assumption is reasonable. When I cannot, the number is essentially unverifiable and I discount it accordingly.
I am not trying to poke holes in your model. I am trying to understand whether you understand your own business. The model is how you show me that you do.
The most common mistake: spend without outcome expectations
The pattern I see more than any other is a cost model that lists spend categories without attached outcome expectations. Marketing spend appears as a line item with no corresponding assumption about what that spend should produce. Headcount appears as a salary cost with no modelling of the contribution that each hire is expected to make.
This is not just a modelling problem. It reflects a deeper strategic gap: the founder has not connected their spending decisions to their financial outcomes. They are operating on faith rather than on a feedback loop.
The best founders I have worked with can tell you, for every significant line of spend in their model, exactly what it is supposed to produce and what they will do if it does not.
What I see in the revenue assumptions
Revenue projections in early-stage models tend to fall into one of three patterns.
The flat percentage: revenue grows at X percent per month for the entire projection period. This is almost always too simple.
The hockey stick: flat or modest growth for several months followed by a sharp inflection. This pattern is not inherently wrong, but it requires a clear explanation of what causes the inflection.
Activity-driven revenue modelling: Instead of projecting a growth rate, the founder projects the activities that generate revenue: inbound leads by channel, outbound sequences, partnership referrals, conversion rates at each stage of the funnel, and average contract values by customer segment. The revenue is the output of these activities, not an assumption in itself. This approach is more work to build, but it is far more defensible and far more useful as an operational tool.
The headcount models that concern me
There are two headcount patterns that raise flags consistently. The first is the team that is too small for the ambition. The second is the team that is too senior too early: a seed-stage startup planning to hire a VP of Sales, VP of Marketing, and Head of People before they have reached product-market fit.
The headcount models I find most credible are ones where each hire is tied to a specific milestone that the previous hires are expected to reach. There is a logical dependency chain, not just a spending plan.
On financial projections and reality
Every investor knows that financial projections will not be met precisely. That is not the point of the model. The point is to demonstrate that you have thought carefully about your business, that your assumptions are grounded in something observable, and that you will know quickly when reality deviates from the plan and why.
The founders who earn my confidence are not the ones with the most aggressive projections. They are the ones who can say: here is what we expect, here are the specific assumptions underlying that expectation, and here is how we will know within thirty to sixty days whether those assumptions are holding.
A practical checklist before you send your model to an investor
- Every revenue line is driven by a specific activity with a specific conversion assumption, not a growth percentage.
- Every cost line has an attached purpose. You can explain what each spend is expected to produce.
- Your unit economics are calculated correctly and you can explain every component from memory.
- You have modelled a conservative scenario where growth is 20 to 30 percent slower than your base case.
- Your runway calculation uses net burn, updated to last month's actuals, not a static figure.
- Your projections extend to 24 months and connect to the milestone that would justify the next raise.
Ready to apply this to your business? Get early access to YourCFO at yourcfo.tech.