How to read this site

1 The big idea

A company's official numbers only come out every three months. In between, lots of free public data hints at how the business might be doing. More cardboard boxes being bought can mean more packages shipping. Fewer people in airport lines can mean emptier planes.

Professional investors pay for data like this; it's called alternative data. We pull the free public versions, compare each number with its own recent past, and explain what it might mean. These are clues, not predictions.

2 A signal card

Each card on a stock page is one piece of offbeat data. Here's a real one from Amazon's page, with its parts numbered:

1. The signal's name and what exactly is being measured.
2. The latest number, often an average of the last few weeks or months so one odd day doesn't throw it off.
3. The change tag. Green if the number went up compared with its baseline, red if it went down.
4. The reading: tailwind, headwind, or neutral for the company.
5. The source and "data through" date, so you can check it yourself.

3 Green/red vs tailwind/headwind

These are two different questions, and mixing them up is the most common confusion:

• Green / red answers "Did the number go up or down?" compared with its baseline, which is its own recent past (for example, the last 28 days vs the 28 before, or this year vs last year).
• Tailwind / headwind answers "Is that move good or bad for this company?" A tailwind pushes in the company's favor; a headwind pushes against it.

Usually they match. But some signals flip: more unsold inventory or more owner complaints are bad news even though the number went up. Live examples from the site right now:

Small moves inside a signal's normal range are called neutral. They show gray badges no matter which way the number moved.

4 The lean, strength, and time window

• The lean adds up all of a stock's signals. More tailwinds = Lean tailwind. More headwinds = Lean headwind. A tie = Mixed. Nothing moving = Quiet.
• Strength (mild, moderate, strong) says how loud the lean is, based on how many signals moved a lot versus a little. It is not a price target and doesn't say how much a stock will move.
• "Weeks to a quarter" is the usual time frame for this kind of clue: data like this is often discussed as hinting at results over the next few weeks to about three months (one quarter). It's a rough range, not a clock, and the market may already know.

5 The summary

On curated stocks, a short paragraph recaps the signals. When it's labeled AI summary, it was written by Google's Gemini using only the numbers on that page. If the AI's text sounds like investment advice ("buy", "price target", "will rise"), we throw it out and use a plain template instead. Treat it as a recap, not an extra source of truth.

6 Sources and "data through" dates

Every card links to where the number came from, like the Federal Reserve's FRED database, the TSA, or the SEC. The "data through" date is how recent the newest number is. We check for new data every 4 hours, but some agencies only publish monthly, so a number can be a few weeks old. If a source fails, the card says so instead of guessing.

7 Has it worked?

That's why the right way to use a lean is as a conversation starter: a reason to look closer at a company, ask why a number moved, and read more. It isn't a prediction. See the full track record →

8 Words to know

Ticker

A stock's short code, like AMZN for Amazon.

Alternative data

Information about a company that doesn't come from its own financial reports.

Baseline

The "normal" a number is compared against, usually its own recent past.

Noise band

How much a number usually wiggles. Moves inside it are called neutral.

Tailwind / headwind

A push in the company's favor / against it.

Quarter

Three months. Companies report results once a quarter.

S&P 500

An index of about 500 big US companies, often used as "the market".

Backtest

Checking how a method would have done in the past, using only data available back then.

Correlation vs causation

Two things moving together doesn't mean one causes the other.