Research skills
The craft: reading filings, the ratios that matter, forensic signals.
Learn
Learn to research,
from the source.
Practical guides for reading public companies from their own filings — the 10-K and the ratios, the forensic signals, and how to run it all in AnalystBook. Our voice, our data — no listicles. (How the product works lives on Technology; the findings live in Intelligence Reports.)
Insider buying: when it actually means something — and the three filters that separate signal from noise
Executives buy and sell their own stock constantly, and most of it tells you nothing. Here's how to read Form 4 filings like an analyst: why single purchases are noise, why scheduled trades don't count, why selling is nearly meaningless — and why a cluster of insiders buying in the same two weeks is the one pattern worth your attention.
Read →The Altman Z-Score, plainly: what the bankruptcy number really says — and the formula trap most screeners fall into
The Altman Z-Score condenses a company's balance-sheet strength into one number with three zones: safe, grey, distress. Here's what the zones mean, why there are actually three different formulas, how using the wrong one branded Salesforce a bankruptcy risk, and how to read a 'distress' reading without panicking.
Read →How to analyze a bank in five ratios
Banks don't work like normal companies — revenue is a spread, inventory is loans, and leverage is the whole point. Five ratios do most of the work: net interest margin, the efficiency ratio, loan-to-deposit, the provision rate, and return on assets. Here's what each one tells you and what a healthy reading looks like.
Read →ROE and ROA, and what actually drives them (the DuPont view)
Return on equity can look great for the wrong reason — borrowed money, not real profitability. The DuPont breakdown splits ROE into margin, asset efficiency, and leverage so you can tell which one is doing the work. Here's how to read ROE and ROA together, and why a high ROE isn't always good news.
Read →Accruals vs. cash: the earnings-quality test that predicts reversals
Two companies can report the same profit while one is backed by cash and the other by accounting entries that quietly reverse. The gap between net income and operating cash flow — the accrual — is the single best measure of earnings quality. Here's how to read it, why high accruals tend to unwind, and the six things that erode quality.
Read →The Beneish M-Score, plainly: how to tell if a company is massaging its earnings
The Beneish M-Score is a formula that reads eight numbers from a company's own filings and estimates how likely it is manipulating earnings. Here's what the score means, where the -1.78 line comes from, why it flags healthy growth companies by mistake, and how to read it without jumping to conclusions.
Read →How to read a 10-K in 30 minutes (without missing what matters)
A practical method for reading annual reports fast: read the diff not the document, let computed signals point you to the sections that changed, go verbatim only where it matters, and capture your judgment where you'll find it next quarter.
Read →Using AnalystBook
The same skills, run as workflows in the product.
A filing just dropped: read what changed, not the whole thing again
Experienced analysts don't re-read a 10-K from page one — they read the delta: which numbers moved, which sentences were edited, which risks appeared or quietly disappeared. Here's that workflow, and how AnalystBook computes the delta for you the moment a company you follow files.
Read →From a ticker to a read you can trust: a 10-minute workflow in AnalystBook
A practical walkthrough of researching a company in AnalystBook — start from the computed record, check the forensic signals, verify any number against the filing in one click, ask Ana the questions that remain, and keep your read where it compounds. Built so you never have to take a number on faith.
Read →