Why does one ruler produce two answers? The same company: one person reads the price-to-earnings ratio and says cheap; another reads the same statements and says expensive. The problem is rarely in the eyes; it is in the ruler. Take a chemical company sitting at the top of its cycle — its profit belongs to the past, its capacity to the present, and the cheapness computed from last year’s profit measures a move already played out. The ruler is not broken; it was matched to the wrong asset.
One valuation framework puts this bluntly: classify the asset first, then match the yardstick — cyclical, growth, value, each with its own ruler. What to measure is decided by what the asset is; only then does the how carry meaning. The ruler follows the asset; the asset does not bend to the ruler.
The cyclical ruler measures capacity and the supply cycle. Nonferrous metals, chemicals, shipping: profit follows price, price follows supply, and while supply is still expanding, today’s profit is tomorrow’s reversion to the mean. A statically low earnings multiple often stands exactly at the cycle’s peak — the cheapness it measures is false cheapness. So the cyclical ruler carries no static-earnings marking at all — not because the ruler is inaccurate, but because that marking should not exist. The question is not how much the company earns now; it is whether capacity is still being added: has supply been cleared, has the industry structure settled. Only when those answers change does the quality of the profit change.
The defense ruler measures injections and orders. How much quality asset the group still holds for injection, and how stable the long-term orders are — that sets the floor of the business. AVIC and CASC: the injection space on the ledger is a marking on this ruler; the current quarter’s reported profit is secondary. Orders are profit that has not happened yet; injections are assets not yet on the books. Measured with numbers already on the books, only half of them shows.
The consumer and pharma ruler measures brand, licenses, and the installed base of customers. A brand brings repeat purchases; a license keeps others outside; the installed base is the future already banked. All three are accumulated — a little each year, for a decade — and destroyed in a single stroke. Their valuations run naturally high, because what is being measured is not this year’s revenue flow but the stock of accumulated equity. Measured with a flow ruler, the verdict is always the same: too expensive. What is expensive is actually a different scale.
First objection: one ruler for every company is simple and saves effort. The saving is real, and so is the error. The unchanging side of each asset class differs by nature — for cyclicals it is the law of supply; for defense, the group and its order structure; for consumer and pharma, the accumulated stock. When the anchors differ, uniform standards are the mistake. And when the measured bargain is false, the effort saved is paid back with interest.
Second objection: different standards for different assets is subjectivity. The order is reversed. Standards follow anchors; anchors differ, so rulers differ — that is not subjectivity, that is correspondence. The real subjectivity is gripping one ruler in advance and combing the market for companies that fit it. Letting the ruler pick the assets is what deserves the word.
Third objection: brands and licenses are the old way; the current way is traffic and growth. A traffic platform is a different asset class with a different ruler — users, retention, monetization. That ruler is not wrong; what is wrong is forcing it onto cyclicals. A ruler matches an asset class, not an era. Stripping the old ruler of its outdated label and raising another misapplied one is not an update; it is measuring wrongly in a new direction.
The pity is for those who pick the wrong ruler: watching the cheapest reading, they catch the most expensive asset.
So the order runs the other way: first ask what kind of asset this is, then ask what it is worth. The asset fixes the ruler; the ruler fixes the reading. With the reading wrong, the more precisely the rest is computed, the more thoroughly it errs. These rulers govern the choice of scale, not the timing of trades; when to act is another subject. The next time a bargain appears, before computing the discount, ask first: which ruler measured it?
