How it works
A ranking you can take apart
PickCompanion is a research publication, not an adviser. One model, run on a schedule, produces one list for everyone. This page explains what goes in, what comes out, and the rules we hold ourselves to.
What gets ranked
US companies listed on the NYSE or Nasdaq with at least three fiscal years of audited filings, a recent filing, at least $50 million in revenue, a market value above $100 million, and enough daily trading to buy and sell without moving the price. Funds, trusts, warrants, and duplicate share classes are excluded. That leaves roughly 1,900 companies, re-checked every run.
How a score is built
Each company is measured on about thirty factors. Every factor is compared with other companies in the same sector, so a retailer’s margins are judged against retailers, not software. Factors are grouped, groups are weighted, and the result is a 0 to 100 score. A risk penalty of up to 10 points is subtracted for things like very high debt, heavy losses, or extreme price swings.
| Group | Measures | 1-year | 3+ years |
|---|---|---|---|
| Growth | Revenue, profit, and cash flow growth over one and three years. | 20% | 25% |
| Quality | Margins, return on capital, cash conversion, and consistency. | 20% | 30% |
| Balance sheet | Debt relative to earnings and equity, interest cover, liquidity. | 10% | 15% |
| Valuation | Price against earnings, sales, cash flow, and book, sector-relative. | 20% | 15% |
| Momentum | Price trend over six and twelve months, and distance from highs. | 20% | 5% |
Two horizons, two weightings. The 1-year view leans more on price trend and valuation. The 3-year-plus view leans on quality and growth, because over long periods the business matters more than the entry point. If a company is missing data for a factor, that factor is left out and the others are re-weighted; a confidence figure shows how complete the picture was.
What the labels mean
Labels describe where a score ranks, nothing more. Strong Buy is the top 3 percent with high confidence, Buy the next 12 percent, Avoid the bottom 15 percent, and Hold everything in between. A high-risk flag can cap a label regardless of score. The labels are the model’s reading of the numbers; they are not instructions to you.
Every score comes with its reasons
Under each score is a list of the factors that helped and hurt, in plain English, each one showing the figure, the filing it came from, and the filing date. These reasons are generated by rules, not by AI, so they cannot be invented. For the top companies we also publish a short AI-written summary. It is only allowed to use numbers already on the page, it is checked before it is shown, and if it fails the check it is not shown at all.
Signals we show but do not score
- AI lean. A labeled reading of recent headlines about a company: leaning positive, neutral, or negative, with the headlines it rests on. Older headlines fade, and repeated notices count less.
- Buzz. Trading volume and headline count compared with the company’s own normal level. Unusual attention is a reason to look closer, not a verdict.
Neither one changes a score or a rank. They sit beside the score, clearly labeled, so you can weigh them yourself.
Where the data comes from
- Financial statements come from each company’s own filings with the SEC, read directly from EDGAR.
- Prices come from a licensed market-data provider, updated after each close.
- Headlines come from a licensed news feed and link to the original publisher.
Every figure shows when it was last updated. If a feed is late, the site says so rather than showing old numbers as current.
The rules we hold ourselves to
- Identical for everyone. The ranking never looks at who you are or what you own.
- No ads, no sponsored coverage, no paid placement. The only revenue is the subscription.
- Nothing is rewritten after the fact. Each day’s ranking is stored the moment it is published, with the model version that produced it.
- The track record is public, including the misses. Every published pick is measured against the S&P 500 at one, three, six, and twelve months.
- Model changes are announced. When the weights or factors change, the version number changes and the change is written down.
What it is not
It is not personal advice, and it is not a prediction. A high score means the numbers look strong by our method; it does not mean a stock will go up. Past results, including our own track record, do not guarantee future results. Investing involves risk of loss. Read the risk disclosure before relying on anything here.