Discounted Cash Flow

The awkward part of pricing a company is that almost all the justification sits in the future. What exists today is office furniture and a bank balance, and an investor pays $20 million because of what they believe the company will earn later.

Adding future money up as-is produces nonsense, though. A million dollars arriving five years from now cannot be worth the same as a million dollars today. Take it today and you have five years to put it to work, and waiting five years carries the chance that the plan misses and nothing arrives at all. Neither the cost of waiting nor the risk of being wrong shows up in a plain sum.

In an investment meeting that gap looks like this. The company writes down that it earns $5 million a year by year three, one side prices the company off that figure as written, and the other side says it's year-three money that may never arrive and cuts it in half. With no rule behind the size of the cut, the negotiation turns into a contest of nerve. Discounted cash flow (DCF) fixes the size of that cut as a single number so both sides run the same arithmetic.

You write down the cash expected in each of the next several years, divide each year by a set rate to convert it into today's value, and total the results. It takes five steps.

1. Choose the projection window, usually five years. Stretch it further and the later figures drift toward fiction, which costs more credibility than it adds precision.
2. Project the cash for each year. Cash, not revenue — what came in from operations minus what went out for equipment and development, grounded in the history laid out in the cash flow statement. For a subscription business, how fast annual recurring revenue (ARR) grows and how many customers leave is what produces these numbers.
3. Set the discount rate, which combines the cost of waiting with the risk of being wrong. Something as safe as a government bond runs a few percent; an early-stage startup gets something around 30 percent. The higher the chance of failure, the harder future money gets marked down.
4. Divide year by year. At a 25 percent rate, $1 million arriving a year out becomes $800,000 after dividing by 1.25, and $2 million arriving two years out becomes $1.28 million after dividing by 1.25 twice. The divisor grows with distance, so later years shrink fast.
5. Add the value of everything past the window. The company doesn't close in year six, so the cash beyond the window gets calculated as a single lump and added on. That lump is the terminal value, and it often accounts for more than half the total.

The sum is what the business itself is worth; subtract the debt to be repaid and what remains is the shareholders' share. In practice this gets built as a spreadsheet with one row per year, wired so that changing the discount rate cell alone recalculates everything.

It fits where future cash is reasonably visible: subscription businesses with long contracts and customers who rarely leave, profitable companies whose growth has flattened out, assets like a building with set rent. The less room there is for the projection to be badly wrong, the more the result means.

For an early-stage startup it's the least trustworthy of the methods, for two reasons.

1. Miss the year-five revenue by a factor of two and the answer is off by a factor of two. For a company with almost no revenue today, the only support for a year-five figure is the founder's plan.
2. Choosing 25 percent instead of 35 percent changes the present value of year-five cash by about a third. There is no correct answer for that rate, which is how people end up picking the conclusion first and tuning the rate until it appears.

Check the weight of the terminal value too. If $13 million of a $20 million total comes from year six onward, the precise-looking five-year model explains only a third of the answer.

So the method never runs alone. Comparable analysis, which prices a company off what similar companies recently traded for, runs alongside it to produce a range. When comparables point at $14 million, multiples at $18.5 million and DCF at $20 million, the negotiation happens inside that band. Its real use is exposing assumptions rather than producing one number. Work backward to find which growth rate and which discount rate yield the price the other side named, and you can see what belief that price rests on.

"I ran a DCF, but the terminal value comes out at 70 percent."

Most of the answer comes from estimates past year six, so nobody should lean on this result hard.

"What discount rate did you use?"

Checking the assumption before the conclusion, usually leading into a question about how that rate was chosen.

"Let's treat the DCF as reference and negotiate off multiples."

The projections can't carry the weight at this stage, so another method should be the primary basis.

This is exactly what an investor asked Pagedeck for ahead of its next round. Grace couldn't hand over the revenue table from the existing business plan: revenue and cash are different numbers, and with annually billed customers in the mix, the moment money arrived no longer lined up with the moment revenue was recognized.

The five-year table got rebuilt on a cash basis. The discount rate was agreed with the investor at 30 percent, and growth was split into optimistic, base and pessimistic cases so all three results went out together. Bring one number and the argument is about whether that number is right; bring three and the conversation moves to which assumption you believe. Whatever gets agreed ends up as a line in the term sheet and gets checked during due diligence.

Because the arithmetic is elaborate, the result is read as exact. There's a table by year and the answer lands on a decimal, so it looks sturdier than a number someone called out from instinct. But the method only calculates the assumptions you feed it, and when those are wrong the result is wrong to the decimal. Mistaking precision for accuracy turns a range into a single figure and leaves you no room to move in a negotiation.

Treating the discount rate as a fixed market number is the other common error. It feels like a rate that must be published somewhere, but it's a judgment about how precarious this particular company is. The company and the investor using different rates is normal, and the same company's rate falls as it matures. Treat it as given and you never ask why the other side chose theirs, handing over the lever that moves the answer most.