A valuation earns its keep before the raise, as your anchor. Building one backwards to justify a number someone else set inverts the whole point
You sit across the table. You said ten. They said twenty. After some back and forth you shake hands on fifteen. The deal is good, everyone’s happy, and now there’s one slightly awkward task left on your list: produce a startup valuation report that “shows” the company is worth fifteen. So you open a model, tweak the growth assumptions, nudge the margins, adjust the discount rate, and back out a set of cash flows that happen to land on the agreed figure.
A founder asked us recently whether Equidam had a feature for exactly this: type in a target valuation, have the model work backwards and fill in the supporting numbers. It’s a completely reasonable question. He’d seen it done plenty of times, and from the trenches it looks efficient. The price is settled. Why not just generate the paperwork to match? The trouble is that a valuation built this way stops doing the only job a valuation report is good for.
We don’t have that feature, and we left it out on purpose. Not out of stubbornness, but because reverse-engineering a report to fit an agreed price quietly destroys the one thing a valuation is actually for. The number isn’t the deliverable. The reasoning is. And once you work backwards, the reasoning is fiction you’ve reverse-fitted to a conclusion someone else handed you.
Pull this apart, though, because “reverse-engineering a valuation” means two very different things, and only one of them is a trap.
The good kind of reverse math (and the trap that looks like it)
Investors reverse-engineer valuations constantly, and there’s nothing wrong with it. A seed investor often starts from two things they already know: how big a cheque they want to write, and how much ownership they need to make the fund’s math work. From there the “valuation” falls out as arithmetic. Jason Mendelson’s worked example, shared by SVB, puts it bluntly: founders ask for $4M, the investor needs a 25% stake, so “on paper the company is worth $16 million” ($4M divided by 0.25). The post-money number is an output of the deal mechanics, not a reading of the company’s intrinsic worth.
That’s fine. It’s how a big share of early-stage pricing actually gets set, and pretending otherwise would be naive. The same guides that teach this also tend to agree on the framing: as one widely-read valuation guide puts it, “valuation isn’t a formula; it’s a negotiation anchored in your next investor’s thesis… It’s worth what you and the investor agree it’s worth, within a range.” Within a range. Hold onto that phrase.
This is where the two meanings split. Backing a post-money number out of cheque-size and ownership is back-of-envelope deal math, and it’s honest about being that. Building a full valuation report (projected P&L, cash flows, discount rate, comparables) to justify a price that was already agreed is a different animal. The first is a quick sketch everyone knows is a sketch. The second dresses up a negotiated outcome as if it were an independent analysis that produced it. One is a tool. The other is theatre.
And the theatre has a tell. A model bent to hit a predetermined number stops being able to tell you anything you didn’t already decide. You’ve removed its only useful function, which is to surprise you.
Why the sequence is the whole game
A startup valuation is worth the most before you walk into the room, and worth almost nothing if you assemble it afterwards to ratify what happened in the room.
Built first, it does real work. It gives you a defensible range to anchor on, so when they open at twenty (or you open at ten), the conversation has a center of gravity that isn’t just two people’s gut feel. It forces you to confront your own assumptions privately, where it’s cheap to be wrong, instead of discovering mid-negotiation that you can’t actually explain why your growth curve bends the way it does. It hands you a support document you can put on the table, walk an investor through, and use to defend the parts of your story that deserve defending. That’s leverage. Quiet, unglamorous leverage, but leverage.
Built afterwards, it does the opposite. It launders a number you didn’t derive into a number that looks derived. It teaches you nothing, because you started from the answer. And it carries a real downstream cost if the price you retrofitted was too high, because the model will obediently “support” whatever you point it at, including a valuation the company has no realistic path to growing into.
That downstream cost isn’t hypothetical. When a round gets priced above what the business can later justify, the correction shows up as a down round at the next raise. Those have been a meaningful slice of the market: Carta’s State of Private Markets shows down rounds reaching 23% of new fundings in Q1 2024, the highest in over five years, before receding. The retrofit feels free in the moment because the deal is already done. The bill arrives twelve to eighteen months later, when you have to raise again from a number you can no longer reach.
So the sequence matters more than the inputs. Valuation before the raise is decision support. Valuation after the raise, bent to fit, is paperwork pretending to be analysis. Same model, opposite value, entirely because of when you built it.
Your last round is story, not the input to the next one
A separate version of the same mistake shows up around prior rounds. Founders ask, reasonably, whether last round’s valuation sets this round’s. It feels like it should, the way a previous salary anchors the next offer.
It does, but only as story, not as a formula. Your previous price is context. It sets expectations, and clearing it cleanly is a nice signal that things went the way you said they would. As FounderCatalyst points out, prior investors set a precedent, and an aggressive early number raises the bar for everyone after: price a round high and miss the growth needed to justify it, and the next round comes in flat or down. The previous valuation shapes the conversation. It doesn’t compute the new number.
What computes the new number is how the company is actually doing now. New traction, new revenue, new evidence that the risks have shrunk. If you’ve genuinely de-risked the business since last time, the new valuation should reflect that, and the prior round is a pleasant footnote. If you haven’t, no amount of pointing at last year’s headline price will hold the number up, because the next investor is underwriting the company in front of them, not the one in the old term sheet.
This is the same trap wearing a different hat. Reverse-engineering from an agreed price, and anchoring rigidly to a past price, both substitute a number from somewhere else for the work of valuing the company as it stands today.
The same logic, by the way, is why a SAFE valuation cap isn’t your valuation. A cap is a conversion threshold, not a price tag. To illustrate the mechanics with round numbers of our own: a $2M raise on a $10M post-money SAFE cap leaves the SAFE holders with roughly 20% of the company at conversion, which can be close to double the stake the later Series A price alone would imply. (The arithmetic there is ours, not a figure the source reports; the point it backs up, that a cap is a threshold rather than a valuation, is theirs.) The cap influenced the deal. It was never a reading of what the company was worth. Treating it as one is the same error from yet another angle: borrowing a number from a prior negotiation and mistaking it for the present-day truth.
What a valuation is actually for
Strip all of this back and there’s one idea underneath it. A valuation is not a number you produce to satisfy someone. It’s a piece of reasoning you build to understand your own company well enough to defend it.
That’s why the sequence is non-negotiable. You build it first, with honest inputs, so it can tell you something. You use it as your anchor and your support document in the negotiation. You let the agreed price land where negotiation lands it, which may be above or below your range, and that’s fine. Then, if a regulator needs a formal certificate, you get the right registered professional to produce one, fully aware it’s a separate, after-the-fact exercise.
Equidam is built for the first job and only the first job. The methodology runs five valuation methods, two qualitative ones that price the risk in your specific situation and three financial ones that price the return, weighting them toward the qualitative side early on (when projections aren’t yet reliable) and toward the financial side as the company matures. The qualitative scoring and the comparables are benchmarked against 160,000+ valued companies and 30,000+ public-market comparables, across 90+ countries and 600+ industries, so the range you walk in with reflects companies like yours rather than a number you talked yourself into. The point is to hand you a defensible estimate and a shared language for the conversation, not to certify a figure and not to ratify one you already agreed to.
We’ll be honest about the limits. No valuation, ours included, is “accurate” in the sense of naming the one true price. Markets set prices; methodology sets a defensible estimate and explains it. What a built-first valuation gives you isn’t certainty. It’s the ability to walk into the room knowing what you’re asking for and exactly why, which is worth far more than a tidy report assembled afterwards to match a number you didn’t choose.
If you’re heading into a raise, build the valuation before you need it, not after you’ve shaken hands. You can run yours on Equidam and use it the way it’s meant to be used: as the anchor you bring to the table, not the paperwork you produce once you’ve left it.