You send your deck to a VC you actually respect. Two days later, the reply lands: "Love the team, love the product. Send me the model." And your stomach drops, because what you have is a spreadsheet you built at 1am with a hockey-stick revenue curve and a prayer.
I know that feeling. The first time an investor asked me for my financial projections, I sent back a tab called "Revenue" with a single formula dragging a number from 10 to 400. He was kind about it. He did not invest. That rejection taught me more about how to create financial projections for startup investors than any template ever did, and I want to save you the same wasted month.
Here is the honest version: investors are not checking whether your forecast is accurate. Nobody can predict three years into a startup's future, and they know it. What they are checking is whether you understand your own business mechanics well enough to survive when the plan goes sideways. The spreadsheet is a test of your thinking, not your optimism.
Key takeaways
- Projections are a reasoning test, not a prediction contest. Vague optimism reads as inexperience.
- Build bottom-up from concrete drivers (conversion, churn, headcount) instead of assuming a growth rate.
- Always present a base, a high, and a low scenario. Investors trust founders who show the downside.
- Your cash flow statement and your raise amount must tell the same story. This is where most models break.
- Know your runway and burn before anyone asks. It is the first number a serious investor looks for.
Why investors ask for financial projections at all
Most founders treat the model as a box to tick. Wrong framing. When a partner at a fund opens your spreadsheet, they are running one question through their head: if I gave this person money, does the money have a job to do, and can they tell me what that job is?
The projections are where that question gets answered. Not in the pitch, not in the market-size slide. In the boring tab where you list what each hire costs and when you run out of cash.
What a financial model is not
It is not a sales document. I have seen founders paste a 40% month-over-month growth rate because "that is what SaaS companies do." The investor asks one follow-up question — where does that rate come from? — and the whole thing collapses. A model that cannot survive a single "why" is worse than no model.
It is also not a forecast you are bound to. If you hit the plan exactly, you probably aimed too low. The document's purpose is to make your assumptions visible so a smart person can pressure-test them with you.
What are the 7 steps of forecasting?
The sequence matters, because building a model out of order creates circular logic you will spend hours untangling. Here are the seven steps I now follow every time, in this order.
- Revenue — built from drivers, not a growth percentage. Units sold × price, customers × subscription fee, whatever fits your model.
- Variable costs — the costs that move with each sale. Payment processing, hosting, delivery, the commission you pay a reseller.
- Fixed costs — rent, software subscriptions, insurance. Predictable, boring, easy to forget.
- Payroll — split by role, with a start month for each hire and full employer cost, not just salary.
- Working capital — the gap between when you pay and when you get paid. This quietly kills profitable companies.
- Capital expenditure — equipment, development you capitalize, anything you buy once and use for years.
- Cash flow and scenarios — pull it all into a monthly cash statement, then run a high and a low case on top.
That order is not arbitrary. Revenue drives variable costs. Payroll drives your burn. Working capital and capex feed the cash statement, which is the only output investors genuinely care about. Build cash first and you are guessing at everything upstream.
Why the order saves you hours
When revenue is a formula from drivers, changing the conversion rate automatically flows through to costs and cash. One cell, whole model updates. When revenue is a hardcoded number, every assumption change means a manual rebuild. I rebuilt a model from scratch twice in one week because I built it the lazy way. Do the drivers first.
Building assumptions an investor will actually believe
The single most common failure I see is the same one I made: revenue that appears from nowhere. An investor can spot an unsupported number from across the table. The fix is to tie every revenue line to a driver you can defend with something you have actually observed.
The drivers that matter
For a subscription business, that means CAC (what it costs to acquire a customer), ARPU (what each customer pays you), and churn (how many leave each month). For a marketplace, it is take rate and transaction frequency. For a physical product, it is units per channel and gross margin per unit.
The investor's real question is not "is this number right?" It is "did you pull this from a real signal?" I once based a conversion assumption on 20 beta users and told the investor exactly that: 20 users, this rate, and here is why I expect it to hold. He did not challenge the number. He challenged the sample size, we discussed it, and the conversation got better. Transparency about where a number comes from is worth more than a polished figure.
On growth rates
Every model has a growth assumption. The mistake is picking the rate because it looks good rather than because you can justify it. If you are pre-revenue, base it on comparable public data or a pilot, and say so. If you have traction, your last three months of actuals are your best evidence — and if they are noisy, show the noise instead of smoothing it into a clean line. Investors respect a jagged real curve far more than a smooth imaginary one.
Scenarios, sensitivity, and the break-even point
A one-line forecast is a red flag. It tells the investor you have not thought about what happens when the world does not cooperate. Bring three versions of the same model.
| Scenario | What changes | What it signals to investors |
|---|---|---|
| Base | Your realistic expectations, driven by current data | You understand your own momentum |
| High | Faster conversion, lower churn, one extra hire | You see the upside without betting on it |
| Low | Slower sales, higher churn, delayed expansion | You know your survival plan |
The low case is the one most founders skip, and it is the one that wins trust. Here is what happens if we only reach half our target and we cut hiring is a sentence that tells an investor you will not run the company into a wall. When I started including a downside scenario, the tone of my investor conversations changed completely. Fewer gotchas, more strategy.
Find your break-even
Break-even is the month when revenue covers your costs. For many companies that is later than the current funding round allows, which is fine — but you need to know the gap and how much cash fills it. Run a sensitivity check: what happens to your runway if churn is 20% higher, or if your biggest customer leaves? If the answer is "we die in six weeks," you have found a problem worth fixing before the meeting, not during it.
The mistakes that kill a projection in the room
I have sat on both sides of this conversation, and the failures repeat with almost comic regularity. Here are the ones I see most.
- Cash flow that contradicts the raise. You ask for a round that lasts 18 months, but your own model runs dry at month 11. Investors spot this instantly.
- No burn rate or runway stated. These are the first two numbers a serious investor looks for. Make them prominent.
- Headcount that appears without dates. A row of salaries with no start month is not a plan, it is a wish.
- Optimism with no anchor. If your growth assumption is far above what you have ever achieved, you need a reason, not a hope.
Notice the pattern: none of these are about being wrong. They are about being internally inconsistent. A model that contradicts itself is the fastest way to lose credibility, and it is entirely avoidable.
Should you use a template or build from scratch?
Templates are a fine starting point and a terrible finishing point. I have used free financial projection templates and Excel sheets I found online, and every single one needed surgery to fit my actual business. A generic template assumes a generic company. Yours is not generic.
My advice: download a template to see the structure, then rebuild the revenue and cost tabs around your own drivers. Keep the format clean and readable. Nobody wants to reverse-engineer your logic through a maze of colored cells.
Quick checklist before you send the model
- Does revenue trace back to drivers you can name?
- Is every hire dated and fully costed?
- Do the cash flow and the raise amount agree?
- Are there three scenarios, including a downside?
- Is your runway and burn rate stated plainly?
The real test is not the spreadsheet
Here is what I wish someone had told me before my first investor meeting. The model is not there to prove you will hit the numbers. It is there to prove you can think, adjust, and survive when you do not. Every experienced investor knows a three-year forecast is a work of disciplined fiction. What they are buying is the discipline.
So build it bottom-up. Show the downside. Know your runway cold. And when they ask why a number is what it is, have a real answer ready. That answer is the actual investment case.
