How do you value a loss-making tech company?

Most valuation tools assume the company makes money. P/E needs profit. EV/EBITDA needs positive EBITDA. A DCF needs cash flows you can forecast.

So what happens when a company loses money — sometimes a lot of money — and is still worth billions?

This question separates candidates who learned a process from candidates who understand why the process exists. The wrong instinct is to say it cannot be valued. The right one is to explain which tools break, why, and what you would use instead.

Why a loss-making company can be worth a lot

Start here, because it is not obvious.

Many technology companies lose money on purpose.

Imagine a software business. It costs $1,000 in advertising and sales effort to win one customer. That customer then pays $100 a month, and typically stays for four years.

So each customer costs $1,000 and eventually brings in $4,800. That is a good trade.

But the $1,000 is spent immediately, and the $4,800 arrives slowly over four years. If the company signs up thousands of customers this year, it will report an enormous loss — even though every single customer is profitable.

The faster it grows, the worse the loss looks.

That is a completely different situation from a company that loses money because its product costs more to deliver than customers pay for it. One is investing. The other is broken.

Your job in this question is to tell those apart.

Step 1: Say what breaks and why

Open by naming the problem.

• P/E needs positive earnings. There are none.
• EV/EBITDA needs positive EBITDA. Often there is none.
• A DCF still works mechanically, but if the next few years of cash flow are negative, nearly all the value sits in the terminal value. That turns the valuation into an argument about what margins the business will earn a decade from now.

Saying this clearly shows you understand the tools rather than just knowing their names.

Step 2: Revenue multiples, adjusted for quality

The usual starting point is EV/Revenue — enterprise value divided by annual revenue. Revenue is positive even when profit is not.

But a raw revenue multiple treats every company the same, which is obviously wrong. A business growing 60 percent a year with 80 percent gross margins is not comparable to one growing 10 percent with 30 percent margins, even if both have the same revenue.

So you adjust.

The Rule of 40

A simple shorthand widely used for software companies.

Revenue growth rate + operating profit margin

If a company grows 50 percent and has a −10 percent margin, that is 40. It passes.

If it grows 20 percent with a −25 percent margin, that is −5. It does not.

The idea is that a business should be delivering either growth or profit, and ideally some combination. Losing money while growing slowly is the bad quadrant.

The Rule of 40 is a screening tool, not a valuation method. But it tells you quickly whether a high revenue multiple is deserved.

EV/Gross Profit

Sometimes better than EV/Revenue. Gross profit is revenue minus the direct cost of delivering the product.

If one company keeps 80 cents of every revenue dollar and another keeps 30 cents, those revenue dollars are not equally valuable. Comparing on gross profit corrects for that.

Step 3: Unit economics

This is where you show you understand the business rather than the multiple. These are the numbers investors actually argue about.

Customer acquisition cost (CAC). What it costs in sales and marketing to win one new customer.

Lifetime value (LTV). The total gross profit that customer generates before they leave.

LTV to CAC ratio. How much you get back for each dollar spent winning a customer. Above 3 is generally considered healthy.

CAC payback period. How many months until a customer has repaid what it cost to acquire them. Under twelve months is strong.

Gross margin. What percentage of revenue is left after the direct cost of serving the customer. Software is usually 70 to 85 percent.

Net revenue retention (NRR). What last year's customers pay this year, including cancellations, downgrades and upgrades. Above 100 percent means existing customers spend more each year — so revenue grows even with no new customers. This is one of the most valuable metrics in software.

If the unit economics are strong, the losses are an investment decision. If they are weak, the losses are the business model.

Step 4: Cohort analysis

A cohort is a group of customers who joined in the same period.

Instead of looking at the company as a whole, split customers by when they signed up and track each group separately.

If older cohorts are profitable and the overall company is not, the loss is being caused by spending to acquire new customers. The business works; growth spending is hiding it. Stop growing and it would be profitable.

If cohorts never turn profitable, no multiple will save the valuation. The problem is structural.

This one analysis often answers the whole question.

Step 5: A path-to-profitability DCF

You can still build a DCF. You just have to do it differently.

Project further out than usual. Maybe ten years rather than five, so you reach a point where the business has actually matured.

Normalise to a target margin. Decide what operating margin the business will earn once it stops spending heavily on growth. Justify that number using mature companies in the same sector.

Present a range, and say what drives it. The answer is enormously sensitive to that end margin. Show what happens at 15 percent, 20 percent and 25 percent.

Being explicit about that sensitivity is a strength, not an admission of weakness.

Step 6: Check the balance sheet

Two things that get forgotten.

Cash runway. How long until the money runs out at the current burn rate? A company with eight months of cash is in a very different position from one with four years.

Dilution. A company burning cash will raise more money. That means issuing new shares, which means every existing share owns a smaller slice. Valuing today's share count and stopping there overstates value per share.

Step 7: Consider the parts separately

Sometimes a company has a profitable core business funding a loss-making new venture.

Consolidated numbers hide both. A sum of the parts valuation values each segment separately and adds them up. It often reveals that the profitable business alone is worth more than the whole company's market value.

The stock-based compensation question

Worth having a view on, because interviewers increasingly ask.

Many technology companies pay staff partly in shares. That does not use cash, so it gets excluded from "adjusted" profit figures, which makes the company look far more profitable than it is.

But it is a real cost. Those shares dilute existing owners. Someone is paying — just not in cash.

The defensible position is to treat it as a real economic cost, and to be sceptical of adjusted EBITDA figures that strip it out. Saying so, and being able to explain why, is a genuine analytical view rather than a memorised line.

Saying it out loud

"The usual tools break here, so I would start by saying why. There is no P/E without earnings and no EV/EBITDA without positive EBITDA, and a DCF would put nearly all the value in the terminal value.

So I would start with EV/Revenue, adjusted for growth and margin — using something like the Rule of 40, or EV/Gross Profit where gross margins vary a lot across the peer group.

Then I would look hard at the unit economics: customer acquisition cost, lifetime value, payback period, and net revenue retention. And I would run a cohort analysis, because if older cohorts are profitable then the losses are growth spending rather than a broken model.

If the cash flows are predictable enough, I would build a longer DCF to a normalised margin, justified against mature peers, and present it as a range.

Finally I would check the runway and the likely dilution, because a company still burning cash will raise again."

What the interviewer is really checking

Can you think when the process is removed?

Most technical questions have a standard method. This one deliberately takes it away. What is left is judgment.

Do you know the vocabulary?

Net revenue retention, CAC payback, Rule of 40, burn rate. These are the terms used in practice, and using them correctly signals that you follow the sector rather than having read a guide once.

Can you tell types of loss apart?

Losing money to win customers who pay back in a year is an investment. Losing money because gross margins are thin is a problem. The valuation approach differs completely.

Do you handle stock-based compensation honestly?

Whether you treat it as a real cost is a genuine analytical judgment. Interviewers like hearing a candidate take a position and defend it.

Do you give a range?

A single point estimate for a business like this is overconfident, and saying so is part of a good answer.

Common mistakes

Saying it cannot be valued.

It can. The question is which method and what assumptions.

Going straight to a DCF with a hockey-stick forecast.

Technically allowed, analytically weak, unless you can defend the terminal margin against real mature comparables.

Using EV/Revenue with no adjustment.

Comparing a 60 percent grower to a 10 percent grower on the same multiple is exactly the error the question is designed to catch.

Ignoring stock-based compensation.

Excluding it flatters margins substantially. Treating adjusted EBITDA as if it were cash earnings is a real analytical failure, not a technicality.

Ignoring dilution and runway.

A company burning cash will raise again. Value per share based on today's share count overstates the answer.

Treating all losses as the same.

The distinction between investment losses and structural losses is the substance of a good answer.

Forgetting the comparables might be wrong too.

If the whole peer group is priced on sentiment, a relative valuation inherits that sentiment. One sentence of scepticism is worth including.