Your projections will be wrong, and their real job quietly changes as your company grows.
You open the template. There’s a cell labeled “Year 5 Revenue,” and you have no idea what to type. You’re pre-revenue, or close to it. You’ve read enough startup advice to know that whatever you put there will be wrong. And yet an investor asked for it, or the accelerator application demands it, so here you are, manufacturing a number you don’t believe to satisfy someone who probably won’t believe it either.
This is one of the more honest forms of founder anxiety, and almost nobody addresses it directly. The how-to guides teach you to build the model. The contrarian blogs tell you hockey sticks are nonsense. Neither answers the question underneath: if the numbers can’t be accurate, what is the point of making them at all?
The short answer: projections don’t have one purpose. They have a purpose that changes as your company grows. Build the wrong kind for your stage and you’ll either look naive or leave money on the table. So let’s untangle it.
The numbers really will be wrong, and that’s normal
First, let’s put the anxiety on solid ground, because it’s justified. Even mature sales organizations with real history struggle to forecast accurately, which tells you something about how hopeless precise prediction is for a company that doesn’t exist yet.
Forecast confidence is the exception, not the rule. Gartner found that fewer than 50% of sales leaders and sellers have high confidence in their organization’s forecasting accuracy. And per SiriusDecisions, 79% of sales organizations miss their forecast by more than 10%. That’s not a startup-incompetence problem. These are established companies with quarters of data, and they still miss. Forecasting the future is simply hard.
Now notice where the early-stage founder lives: the five-year model. The exact horizon where numbers are least trustworthy as predictions is the one you’re being asked to fill in.
And it’s not just first-time founders who looked unimpressive on paper early on. Even category-defining companies grew slowly in absolute terms at first. By one well-known account, Facebook and Google each took about four years to reach hundreds of millions in revenue, and Amazon was only at tens of millions in its third year. Set that against the typical pitch deck, where a brand-new company routinely projects nine figures by year three, and you can see how much of what gets typed into these models is fiction.
So the founder’s instinct is correct. As a prediction, your early projection is close to worthless. The mistake is concluding that the projection itself is worthless. It isn’t. It’s just doing a different job than you think.
Projection vs. forecast: a distinction that matters
People use the words interchangeably, but the difference is exactly what trips founders up.
A forecast is your best estimate of what will actually happen. It’s a prediction, and it should be judged on accuracy. A projection is a modeled outcome under a stated set of assumptions: if these things are true, this is what the numbers do. It should be judged on the quality of its logic, not its accuracy.
As one investor-side framing puts it, a projection is not simply a prediction of where a company might end up; it is a framework that explains how the business behaves as it grows. The same piece makes the consequence explicit: investors are not buying the forecast, they are underwriting the mechanics. That quietly dissolves most of the anxiety. You are not being asked to predict the future. You are being asked to show that you understand the machine you’re building.
Which raises the real question: what should that machine show at your stage? Because here’s where almost every guide goes flat. It treats a pre-revenue founder and a Series B founder as if their projections do the same work. They don’t.
Job #1 (early stage): express ambition and create a shared language
At pre-seed and seed, before you have meaningful revenue, your projections have two real jobs, and accuracy is neither of them.
The first job is to express ambition. A startup raising venture money is implicitly promising a particular kind of outcome. Investors who back you are underwriting the possibility of a large return, which is precisely why they ask for the hockey stick in the first place. As one well-known takedown of the format admits, “no savvy investor ever believes those projections,” and yet they demand them anyway, because the projection signals whether you’re chasing a large enough opportunity to justify the bet. Your year-five number isn’t a promise. It’s a statement of how big you’re trying to go.
The second job is to create a shared language. This is the part the debunkers miss entirely. Without numbers, you and an investor literally cannot have the most important conversation in the room: what kind of company is this? A comfortable lifestyle business throwing off cash for the founders? A venture-scale company aiming for a meaningful exit? A generational, category-defining business? These are radically different paths, and they imply different funding, different ownership expectations, different definitions of success.
You cannot have that conversation in adjectives. “Big” and “ambitious” and “scalable” mean nothing on their own. The moment you put a model on the table, even a wrong one, you’ve given everyone a concrete object to point at. The investor can say “your year-three number implies you’ll need three more rounds, are you comfortable with that dilution?” You can say “this assumes a 4% conversion rate, here’s why I think that’s reachable.” Now you’re actually talking. There is no “right” answer to what kind of company you’re building. But without the numbers, you can’t even have the discussion.
And here’s the subtle thing the model reveals: whether you understand your own business. Early-stage investor Phil Nadel argues that financial projections are “the window to a company’s soul,” giving investors deep insight into how the founders think about their business. Unrealistic numbers don’t sink deals by being inaccurate, since everyone expects inaccuracy. They sink deals by tipping off inexperience. A founder who claims EUR 50M in year two with no explanation of how the funnel produces it has revealed they don’t understand the funnel. A founder whose EUR 4M year-two number is built on honest, traceable assumptions about pricing, conversion, and churn has revealed the opposite, even if the EUR 4M never materializes.
So at this stage, stop optimizing for precision. Optimize for internal coherence and honest assumptions. The questions that matter: Do my numbers connect to each other? Can I defend every major assumption? Does the trajectory honestly reflect how big I’m trying to build? That’s a model fit for your stage. As one pre-seed-focused writer puts it, forecasting is useful at the beginning as a reasoning exercise, not as a crystal ball.
Job #2 (as you de-risk): prove predictability
Now fast-forward. You’ve got revenue, paying customers, a few quarters of data, maybe a Series A behind you. The job of your projections flips.
The investor is no longer underwriting a dream. They’re underwriting a machine, and machines are judged on whether they behave predictably. The implicit promise changes from “this could be huge” to “put in EUR 10M, get EUR 25M out, and here’s the mechanism.” Now accuracy, unit economics, and variance genuinely matter, because the investor is paying for a system that converts capital into growth at a known rate.
The same writer who calls early forecasting a reasoning exercise stresses that no forecast survives first contact with reality, so projections should be living documents that you re-forecast as you learn. The further you get, the more those living numbers are expected to hold. The thing being tested moves steadily from can you articulate where you’re going toward can you prove the economics are real and repeatable. Ambition gives way to predictability.
Practically, this means the later-stage model carries different weight:
- Unit economics have to be real, not assumed. CAC, payback period, gross margin and retention all have to be grounded in your actual data, not industry hand-waving.
- The near term has to be tight. Conventional practice is a three-year view where year one is broken down monthly and, beyond year two, the forecast can shift to quarterly. Distant years stay directional. But as you mature, the near years had better be sharp, because now you’ll be measured against them.
- Knowing your numbers cold becomes the credibility test. At seed, fumbling your churn number is forgivable. At Series B, it signals you’re not running the machine you’re asking someone to fund.
The format barely changes between stages. What changes is what people are actually checking when they read it.
What this means for the model you build today
If you’re early, here’s the permission you came for: stop manufacturing fake precision. A five-year forecast accurate to the euro is not the goal and was never achievable. Build a model that is honest about its assumptions, internally consistent, and ambitious enough to reflect the company you’re genuinely trying to build. That’s not a lesser model. It’s the correct model for your stage.
If you’re scaling, the opposite warning applies: don’t keep pitching ambition when investors have started buying predictability. A hand-wavy “we’ll 10x” deck that worked at pre-seed reads as a red flag at Series A, because the audience has moved on to “show me the economics hold.”
This is also where a structured valuation earns its place. A multi-method valuation isn’t a magic accuracy machine, since nothing makes the future knowable. What it does is turn your projections into a shared, defensible language with investors. The qualitative methods, the Scorecard Method and Checklist Method, capture the parts of an early-stage company that no spreadsheet can: team, market, traction, the strength of the idea. The financial methods, DCF with Long-Term Growth, DCF with Multiple, and the Venture Capital Method, translate your projections into a value, with their influence growing as your numbers become reliable.
That weighting is built to track the same shift. Early on, when projections are about ambition, the qualitative methods carry more weight. As you mature and your projections start proving predictability, the financial methods take over. The valuation does at each stage precisely what your projections are for, drawing on a methodology that is public, IPEV-aligned, and benchmarked against 160,000+ valued companies and 30,000+ public-market comparables across 90+ countries and 136+ industries.
So stop treating “my numbers will be wrong” as a failure and start asking the only question that matters: what job should these numbers do at my stage? Early, they express ambition and give you a shared language for deciding what kind of company you’re building. Later, they prove the machine is predictable.
If you want projections investors can actually engage with, build your valuation on Equidam — and see how the methodology weights them by stage.