Fair value is an estimate of what a share ought to cost — the price at which it would be neither a bargain nor an overpayment given what the business produces. It is the reference point the quoted price is measured against; without one, “cheap” and “expensive” are moods rather than statements.
There are two roads to it, and they answer different questions. Absolute valuation builds the figure from the company’s own cash flows and ignores the market entirely — that is intrinsic value. Relative valuation asks what multiple this business has normally traded on and applies it to what the business earns now: if a company has usually been priced at 18 times earnings and is priced at 11 today, its earnings imply a higher share price than the one on the screen. The second road is the one the word “fair value” most often means in practice, because it needs no forecast.
The figure per share is earnings, free cash flow, revenue, book value or EBITDA; the reference multiple is typically the median of the company’s own history over a chosen window, or the median of its peers. For a dividend the arithmetic inverts — fair value is the dividend divided by the reference yield.
Suppose a company earns $4.00 per share over the last twelve months, and over the past ten years its shares have typically been priced at a median of 18× earnings:
Against a quoted price of $60 that implies roughly 20% of upside — on the single assumption that this company goes back to being priced the way it has usually been priced. Run the same arithmetic on free cash flow, on revenue and on book value and you get three more answers; the spread between them says how much the conclusion depends on which denominator you picked.
Read it as a comparison, never as a verdict. A fair value built from a company’s own history assumes the past decade was normal and that the business has not changed shape; a fair value built from peers assumes the peer group is fairly priced. Both assumptions are worth stating out loud before acting on the number.
Use several denominators and look at the spread. Where earnings, cash flow and book value broadly agree, the reading is about the company. Where one of them is far from the rest, it is usually the one whose denominator the business has outgrown.
In the app, the Fair Value tab on a company page runs seven of these readings side by side and blends them into a median, with each lens judged by its valuation percentile against its own history.
Relative fair value is at its strongest on a mature business with a long, boring record: enough history for a median to mean something, and no structural break in the middle of it. It is also the quickest sanity check on an intrinsic-value estimate — if a discounted cash flow says a share is worth three times what every multiple in its history implies, the forecast rather than the market is the thing to re-examine.
A company’s own history is a poor reference when the company has changed: a business that has moved from hardware to subscriptions is not the business whose median multiple you just computed. The method is also circular in a way absolute valuation is not — it assumes the market was right on average in the past in order to say the market is wrong now. And a multiple collapses whenever its denominator does: negative earnings, a year of heavy capital spending or a suspended dividend make the corresponding reading meaningless rather than merely low.