Capital Copilot
Capital
  Copilot
sell Pricing person_addlogin
article
arrow_backCapital Copilot

How to choose a valuation model for any company

eventNovember 9, 2026schedule5 min readvisibility43 views

Three questions about the company, the models that fit each type, and what to do when none does

Every valuation model answers the same question, what is this business worth?, but each one makes a different assumption about the business. Pick a model whose assumptions match the company, and the answer means something. Pick the wrong one, and you get a precise number that means nothing.

This guide shows how to match a model to a company. It also covers what to do when no model seems to fit.

Start with three questions about the company

  1. How does it return cash to owners? Through dividends, buybacks, or by reinvesting everything?
  2. Are its earnings steady or cyclical? Does profit grow smoothly, or swing with an industry cycle (chips, energy, autos, housing)?
  3. What drives its value? Cash flow, assets (banks, insurers, real estate), or future growth that hasn't shown up in profit yet?

Your answers point to the models that fit.

Which model fits which company

Company type Models that fit Why Use with care
Steady, mature cash producer (consumer staples, utilities, many industrials) DCF, P/E vs its own history, dividend discount model (if it pays a steady dividend) Its past is a good guide to its future, so history and steady-growth models work. —
Fast-growing and profitable (large software, platforms, leading chip designers) DCF on free cash flow, reverse DCF, forward P/E It's valued on future cash. The reverse DCF shows what growth the price already assumes. PEG and P/E vs its own history: both assume today's growth or yesterday's company.
Growing, but little profit yet (reinvesting nearly everything) Reverse DCF, price-to-sales with margins in view, earnings power before reinvestment Reported profit is near zero by choice, so profit-based models break. Anything built on P/E.
Cyclical (memory chips, commodities, autos, airlines) DCF with a full cycle, mid-cycle P/E (average earnings across good and bad years) One year's earnings can be a peak or a trough. A P/E on peak earnings looks cheapest just before earnings fall.
Asset-heavy or financial (banks, insurers, real estate) Price-to-book, dividend discount model, Graham number Value sits on the balance sheet. A DCF, since "free cash flow" is hard to define for a bank.
Mature dividend payer Dividend discount model, dividend yield vs history Dividends are how owners get paid. A DDM for a company that buys back stock instead (it will read near zero).

An example: Nvidia (Nov 6, 2026 close).

  • Fit: the cash-flow DCF and the reverse DCF.
  • Partly fit: forward P/E, which depends on next year's estimates.
  • Didn't fit:
    • the dividend model: it pays almost no dividend and returns cash through buybacks
    • the Graham number: its value isn't in its book
    • PEG: its growth rate can't last
    • P/E against its own history: today's Nvidia is a much larger, higher-margin business

The full walk-through: What is NVDA's current value?

Use more than one, and read the range

No single model is "the" value. Run the two or three that fit, and read the range they give.

  • If the models agree, you can be more confident about the value.
  • If they disagree, find out which assumption causes it. It is usually the growth rate or the discount rate, and that tells you what you're really betting on.

Capital Copilot suggests a model for each company

On a company's Valuation page, Capital Copilot runs the DCF, reverse DCF, dividend discount model, Graham and EV/EBITDA and multiples models. Each one shows its value against today's price, and Capital Copilot suggests which models fit that company. For example, for Reddit it suggests the DCF, or P/E against growth.

Use the suggestion as a starting point, then check it against the three questions above.

When no model seems to fit

Sometimes every model gives an answer you don't trust. Then:

  1. Start from the price. Use the reverse DCF to see what growth the price assumes, and ask whether that's believable given what the company has done. You don't need to forecast to do this.
  2. Look at a range of assumptions, not one. The Sensitivity tab shows the value across discount rates and growth rates, so you can see how wide your uncertainty really is.
  3. Compare with peers. A company that's hard to value on its own is often easier next to two or three similar businesses.
  4. Check earnings quality. One-off gains, such as selling a business or gains on investments, can flatter profit. Compare free cash flow with net income.

More methods professionals use

These methods go further for specific kinds of companies. We're looking at which ones to add to Capital Copilot next:

  • Scenario DCF: separate bear, base and bull growth paths, including a down year for cyclical companies, weighted by how likely you think each one is.
  • Mid-cycle margin DCF: values a company at its average margin across a full cycle, not at today's peak or trough.
  • Customer-spending model: for suppliers like chipmakers, revenue = what their customers plan to spend × the supplier's share.
  • Sum of the parts: values each business line separately, for companies that are really several businesses in one.

Which ones would you use? Tell us in the comments on the video.

Education, not investment advice. Models are tools for thinking about value, not predictions.

Comments

Loading comments

rocket_launchCapital Copilot

Want your own AI research briefings?

Capital Copilot turns market research into audio you can listen to. Free plan, no time limit.

More from Capital Copilot

Lesson 3: Stocks

Free beginner course, lesson 3: what a share is, common and preferred stock, dividends, splits and buybacks

Nov 19, 20268 min read

A 16% AI bond isn't free money. Here's what the yield is telling you.

What Microsoft, Alphabet, Amazon, Meta, Oracle and CoreWeave pay to borrow, what that price says about risk, and why interest rates matter more than you think

Nov 18, 20266 min read

Google Cloud doubled. Alphabet's free cash flow went negative. Is the AI spending paying off?

What the filings show about cloud revenue, capital spending and backlog at Google, Amazon and Microsoft, and the signals to watch

Nov 18, 20266 min read

Mastercard earns 93% on its capital. Its shareholders got 13% a year over five years. Here's why.

What return on equity and return on capital really measure, and the two numbers that decide your return

Nov 18, 20264 min read

Research like this, on tap

Briefings, reports and audio TL;DRs — free, with no time limit.

© 2026 Perpetuator LLC · Terms · Privacy · Your Privacy Choices · Educational, not financial advice