Before you pay for a stock research tool, test it on one company

Every stock research tool looks the same from its pricing page. Charts, scores, a phrase like "institutional-grade data", a green button. You cannot tell them apart from the outside, and the outside is the only thing the tour shows you.

So do not judge it from the outside. Before you pay for any stock research platform, take one company you already own and check five things: whether the tool covers it at all, whether you can trace one number on the page back to the filing it came from, whether the tool tells a fact from a forecast, whether you can repeat the one task you would actually use it for, and what the price buys once you hit the cap behind it. Ten minutes, one company, and the tools stop looking the same.

Our own answers to all five are further down, read from our own plan file and our own tables on 16 September 2026, including the two I would rather not have to give.

Why one company, and why one you own

I paid for the serious platforms. After I lost money on a company called Tattooed Chef, I went and did the work: certifications, an unreasonable amount of reading, subscriptions to most of the tools that get recommended. They taught me two things at once: there is a lot of genuine expertise in them, and almost none of it is arranged in a way that tells you what to do on a Tuesday. Eventually I stopped looking for the tool and built one, but that is not the point here.

The point is that I could have found that out in an afternoon, for free, if I had tested each tool on the one company I already held instead of on the demo it wanted to show me. A demo is a company chosen because the tool looks good on it. Your holding was chosen by you, for reasons you can still remember, and you already know things about it that a page has to get right.

That is what makes it the test. You are not checking whether the tool is impressive. You are checking whether it is right about something you can verify, and whether it would have shown you the thing you missed.

With Tattooed Chef, the red flags were in the filings, in public: working capital shrinking quarter by quarter, the auditor's language changing. I did not know how to read them. A tool that had put those two lines in front of me, with the filing they came from, would have earned its subscription in a single afternoon. That is the bar.

1. Is the company covered, and what does missing mean?

Type the ticker. If it is not there, that is your first finding, and it is a bigger one than it looks. Then ask the second question, which almost no tool answers on the page: what does absence mean here? Not measured, or measured and rejected?

Think of a map that shows only the streets the mapmaker has walked. The blank space is not "no road". It is "nobody went there". A tool that covers 900 companies and a tool that covers 40,000 are not making the same promise, and neither is better by default. The small list is measured consistently. The big list is wide. What you need to know is which one you are looking at.

Our answer. We cover 977 companies across 11 sectors and 24 countries, the large majority listed in the US, and 954 of them carry a score. The other 23 are in the universe without one, because their filing history is too short or a figure the score needs cannot be read, and we leave the score empty rather than fill the gap with a guess. Small caps, very thinly traded names, recent listings and companies with no revenue yet are not there, on purpose, for the reasons on the coverage page. If you hold a $400 million company, we fail this check for you, and you should know that before you pay.

2. Can I trace one number back to its filing?

Pick one ratio on the company's page. Return on equity will do. Now try to get underneath it. Can you see the two numbers it was made from? Can you see the date of the filing they came from? Or is it a total with no lines under it, like a restaurant bill that says 84 and nothing else?

This matters more than it sounds, because a ratio can be perfectly correct and still tell you nothing. Home Depot reported a return on equity of 110% for the year to February 2026. The number is right. It is also close to meaningless, because years of buying back its own shares left it with $12.8 billion of equity to divide by, against $60.9 billion of debt. You can only see that by looking at the inputs. A tool that shows you the 110% and hides the equity line has handed you a fact with the meaning removed.

The test is whether you can get from the number on the screen to a line in a document the company filed with a regulator, with a date on it. If the trail ends at "our proprietary data", it ends.

Our answer. Every input behind a score sits on the company's page with its date, and every fundamental is dated to the day the filing reached the SEC, not the quarter it covers. How those inputs combine into one number is ours, and I would rather say that plainly than pretend there is nothing behind the curtain. So on our site the trail goes all the way down to the filing and stops one step short of the recipe. Know that going in.

3. Is it a fact or a forecast?

Revenue in the last annual report is a fact. A price target, a fair value, an "intrinsic value", next year's growth rate: those are forecasts. Somebody made them, on a date, using assumptions, and they are often wrong for reasons that have nothing to do with the company.

Both kinds of number are useful. The problem is a page that prints them in the same font. Yesterday's rainfall and tomorrow's chance of rain are both worth knowing, but you would not want them in the same column of the same table with no heading, and that is what most company pages do with a filed margin and an estimated one.

So for every number that is a forecast, ask three things. Whose is it? When was it made? What does it assume? If the tool cannot answer any of the three, it is asking you to trust an anonymous prediction, and you already tried that with whoever you followed online.

Our answer. Our score is arithmetic on filed numbers and a market price; it is not a prediction and the page never calls it one. Where we do estimate, the fair value page says what the estimate assumes and where it breaks, and the analyst target on a company page says whose it is (the Wall Street consensus, with the number of analysts behind it) and when it was last updated. The one place we forecast in public is the Monthly Top 10, and that comes with the thesis, the risks and every change since the last edition, written down before you can check it against what happened.

4. Can I repeat the task I actually have?

The tour shows you the tool's task. You need to test your own. Write down the one thing you would do twice a month if you paid for this: compare the company you own with its two closest rivals, see how its margins moved over five years, find out what changed since the last quarter. Then do it. Time it. Do it again a week later and see whether it was faster the second time or whether you had to hunt for everything again.

It is the difference between the dealer's loop and your own commute. A car that is lovely on the test route and miserable on the road you drive every day is a car you will stop driving, and a tool you have to fight to do the one thing you bought it for is a subscription you will forget to cancel.

Our answer. The Free plan gives you one full company analysis a month, a watchlist of five, and the Top 10 archive two editions behind the current one. The screener is not on Free. So you can test checks one to three properly without paying, and check four on one company. If the task you actually have is screening, you cannot test it for free here, and I would rather say so than let the pricing page imply otherwise. The road for that is the first month of Core, which carries a seven-day refund window (the refunds page has the exact terms).

5. What does the price actually buy?

The headline price is never the price. The price is the cap behind it: how many companies you can open a month, how many rows the screener will show, how far behind the archive sits, whether the thing you found last month is still available to you next month. A gym with "unlimited" on the sign and a turnstile that counts is a gym you should read the small print of.

So find the cap. Every tool has one, and the good ones print it on the pricing page instead of in the terms.

Our answer, as read from our plan file on 16 September 2026:

There is no free trial. Free is the evaluation path, which is the whole reason this post exists. Core is about the price of one lunch a month, and somebody with a $20,000 portfolio paying that to see every input behind every number is getting the better side of the deal. That is how I want it to be.

The checklist, to copy

Run this on the tool you are about to pay for, and on us, with the same company:

1. Coverage. Is my company there? What does missing mean on this site?
2. Trace. Can I follow one number to the filing it came from, with a date?
3. Fact or forecast. Is every estimate labelled with whose it is, when it was made and what it assumes?
4. Repeat. Can I do my own task, twice, without paying?
5. Cost. What is the cap behind the price, and what happens when I hit it?

Ten minutes. Write the five answers down, because in a month you will not remember why you picked the one you picked, and that note is the only thing that stops you paying for two.

What the test cannot tell you

One company is a sample of one. A tool can pass all five checks and still be wrong about the company, because being traceable is not the same as being right, and a method laid out in full can still be a bad method. This test finds out whether you can check the work. Whether the work is any good is a longer question, and it is the one the methodology pages exist to let you argue with.

It also cannot tell you what to do about the company. That was true of every subscription I paid for, and it is true of this one.

Run it on us first

The fastest way to run checks one to three on this site is the Stock Deep Dive: pick one stock you already own, and the full analysis arrives as a PDF by email, usually within minutes. No card, one per person. If you would rather see the shape of it before you type an email address, the sample report is the whole thing for Amazon, every page, so you can look for the lines you would want to trace and decide whether they are there.

Educational and proprietary. This explains what our research does, not the exact formula behind it, and it is not personalised investment advice. See the full disclaimer.