18 pieces
Fundamental research is the work of forming a view on a business by studying the business — what it sells, to whom, at what margin, against which competitors, under which management. It sits opposite the approach that treats a share price as a pattern to be traded rather than an ownership stake to be understood.
The method is old and not especially glamorous. Philip Fisher described much of it in 1958: before buying into a company, talk to the people who genuinely know it — customers, suppliers, former employees, competitors — and weigh what they say against what management claims. What has changed since is not the framework but the evidence available to apply it. Earnings-call transcripts, regulatory filings, review archives and job postings now sit in public view in volumes no analyst could ever read unaided, and machines can now read all of them.
The essays below cover both halves. How the judgment gets made: what thousands of earnings calls reveal when read together, why a great company is not automatically a great investment, and why this work matters most exactly when prices are moving on news that has nothing to do with the businesses underneath them. And how the evidence is gathered now that the reading is no longer the bottleneck.
A devil's advocate in decision making is a named assignment to defeat a case, not a personality or a mood in the room. An analysis deserves trust only after someone with standing has tried to kill it and failed.
A private investment's valuation is a mark, not a price — the last investor's terms applied to every share that never traded. WeWork carried $47 billion until a real buyer had to pay.
AI checks its work only when it is built to and when the people relying on it insist. The simple tests — admitting a figure is missing, citing the line, disagreeing with the asker — outrank any benchmark.
A quality of earnings report tests whether reported profit is real, repeatable and collected in cash. It protects a buyer only while the findings can still change the price, as Caterpillar's Siwei write-off showed.
Segment reporting is read by setting the company-wide totals aside for the note that splits a business into its parts, then asking of each part how fast it grows, what margin it keeps, and which costs were left out.
Municipal bonds are not a safe category; each one is a claim on a specific stream of cash. Detroit's bankruptcy left water and sewer bondholders with their principal while general obligation holders took losses.
Adjusted EBITDA is profit with selected costs removed, and no rulebook says which ones. It is fair when those costs will not return, and misleading when they are the cash a business must spend to stay open.
When an earnings call and a 10-K disagree, believe the filing. The call is what management chose to say; the 10-K is what it was required to disclose, and the deciding fact usually sits in the notes.
The engine is sold near cost, and the money arrives over the following twenty-five years in mandated overhauls, spare parts and service contracts. A thin front end is often what a long, durable cash stream looks like.
Revenue that never converts into repeatable, owner-controlled cash is a story rather than growth. Three plain questions separate the businesses whose cash actually compounds from the ones with a good chart.
A competitive advantage is only real once it shows up in the cash a business keeps after paying to stay alive and grow. Treat the word as a question to be tested over years, not as an answer.
Reported profit is an accountant's considered opinion; cash is closer to a fact. An earnings collapse can be almost entirely an accounting artifact while the cash the business generates barely moves.
Researching a business through direct conversation with its customers, suppliers and former employees is not a new skill for someone who has built a company. It is the same habit, pointed outward.
SpaceX went public at more than a $2 trillion valuation on its first day of trading — and whether it is a great company and whether it is a great investment at that price are two entirely different questions, with a disciplined way to tell them apart.
In a week where a ceasefire moved the Dow 1,300 points and oil swung 16% in a single session, the investors who thrive are the ones who already did the work.
Every quarter, thousands of public companies deliver earnings calls. Most investors listen to a handful. What happens when you read thousands of them systematically — not for the numbers, but for the language?
AI in investing usually means algorithmic trading. The more interesting shift is happening on the other side of the industry — in the messy, judgment-heavy work of understanding businesses.
The most durable edge in investing isn’t in the data everyone can see — it’s in the qualitative signals most models can’t process.