How to Analyze a Stock Without Taking Management's Word

Analyzing a stock starts with one sentence the company says about itself, and with checking that sentence in three places at once: the financial statements, the earnings-call transcript, and a competitor's annual report. Not with ratios. Ratios are the output of this work rather than its starting point, and research that ends in a verdict instead of a list of conditions is usually a decision you had already made, written down afterwards.

What to Read First When You Analyze a Stock#

The document the company signed. Primary sources have one advantage over commentary: nobody shortened them on the way. Commentary is written from the filing, so it drops the footnote, the qualifier and the conditional clause, and that is usually where the content lives.

For U.S.-listed companies the full set sits in the SEC's EDGAR database, and for most others on the company's own investor-relations page.

The calendar for this work is set by the regulator, not by you. The largest filers have 60 days after the fiscal year end for the annual report and 40 for the quarterly one, which means your sources refresh on a rhythm you know in advance and can put in a diary. Anything the company decided was too important to hold until the quarter goes into a Form 8-K within four business days.

There is one more document almost nobody opens, though it says more about management than every earnings call combined: the proxy statement, filed as a DEF 14A, which carries the table of what management is paid and on which measures.

Now the move that appears on no list of ratios. Before you calculate anything, pick one sentence out of those documents: the one where the company explains why it makes money and why it will keep making money.

It usually sits in the Business section of the annual report or in the first paragraph of the shareholder letter, and the SEC's own guide to reading a 10-K tells you to start there. It reads roughly like this: "we grow faster than the market because customers who connect to our platform do not leave."

That is the story. Everything else is checking whether it is true. If you are still assembling the wider process around this work, the piece on investing in stocks from the ground up shows where it fits.

Three Statements, One Story#

Financial statements are not a test you grade the company on. They are three accounts of the same quarter, written from the same ledger, which is why they are obliged to agree. A witness describing the same evening three times is under the same obligation, and fails the same way, on a detail.

The income statement says what the company earned on paper. What it paid with and what it owes belongs to the balance sheet. How much of that profit reached the bank account is in the cash-flow statement, and that is the one hardest to dress up.

Say revenue grows 12% year over year while receivables grow 40%. Which means the company sold but did not collect, and financed its own growth by lending to its customers. On its own that gap is not a verdict: it also describes a company that just landed a large retail chain with long payment terms. It is a question the story has to answer.

The same gap running the other way is more serious. Say net income comes in at 200 million for the third year running while operating cash flow comes in at 120. Which means 40 cents of every dollar of profit never turned into money, and the company has spent three years reporting a result its own bank account cannot see.

The third point of contact is the most practical and the one analyses drop first: debt maturities set against cash on hand. A company can carry a fine margin and growing revenue and still owe, fourteen months from now, an amount it does not have and will not earn. At that point its future is decided by a bank and a refinancing rate rather than by its management.

The reading rule here is single: comparatively and historically, never at face value. One number from one quarter says nothing, because there is nothing for it to disagree with. Is an 18% margin high? Unanswerable until you know what it was three years ago and what the people selling the same thing report today.

The Question Management Walks Away From#

The filing shows what happened and says nothing about what management thinks of it. That is sometimes the difference between one weak quarter and the beginning of the end of a margin.

An earnings call splits into two parts of wildly different value. The first was written in advance and read off a page. The second is the analyst questions, and nobody wrote that one.

You are looking for one thing in it: the question whose answer does not fit the question. An analyst asks about margin in a specific segment, the chief executive answers about strategy across the whole company. If the same question comes back later in the same call from a second analyst, you have been handed the address of the risk for free, by people who are paid to find it.

I think that single exchange says more about a company than the entire results presentation, because it is the only part of the quarter nobody had time to edit.

It works in the other direction too, and it is one of the few genuinely positive signals a transcript carries. Management that answers a question about a weak quarter with a number, the segment responsible and a date for the fix has just agreed to be checked in three months. People with something to hide do not do that.

The second test is slower and better. Read four consecutive transcripts and see whether management describes the strategy in the same words. The team that holds one measure of success for a year is credible even when that measure is going against them. The team that changes the measure every quarter has just told you which one it does not intend to meet.

How Management Gets Paid#

A transcript tells you what management wants you to think. The proxy statement tells you what management gets paid for, which tends to be closer to the truth.

The document is the DEF 14A, and it carries the compensation table along with the measures the bonus depends on. That table is the shortest route to knowing which number management will defend to the end and which one it will give up without a fight in a weak quarter.

Say half the bonus depends on growth in earnings per share. Which means management has two routes to the same target: improve the business, or reduce the number of shares through a buyback. The second is faster, cheaper in effort and entirely legal, and in the income statement it looks almost identical to the first.

If the bonus depends on revenue growth instead, the tension runs the other way: a debt-funded acquisition looks sensible from the chief executive's chair even when, from a shareholder's chair, it is growth bought on credit.

None of which makes either scheme improper. It means you know where management's interest parts company with yours, and you can check it in the numbers instead of guessing from tone.

The Company Picks Its Own Rivals#

An 18% operating margin means nothing until you know what everyone else earns. The trouble starts with the word "everyone else."

A ready-made peer group from a screener sorts companies by industry code and will cheerfully put two businesses in one table that have never competed for a single customer. The more reliable list sits in the annual report itself, in the Competition section, where the company names the firms taking the same revenue. It is a statement signed under liability, so it tends to be cautious, but it does not tend to be invented.

Then comes the step most analyses skip: read one of those competitors' annual reports. Not to line the two companies up in a table, but to hear the same market described by someone with an interest in describing it differently.

Where the two descriptions come apart is the most interesting question in the whole exercise. One company writes about rising switching costs; the other about a market where a customer moves in a week. They cannot both be right, and both signed their name to it.

While you are there, test the structural advantages your company claims for itself: distribution scale, switching costs, network effects, regulatory protection. The rule is simple and unkind. An advantage only your company mentions is a marketing hypothesis. An advantage its competitor describes in its own filing as a source of its problems is a fact.

The Ratio That Settles Nothing#

Ratio lists are popular because they genuinely work as a sieve. Return on equity, operating margin and price to earnings let you sort five hundred companies in an afternoon and keep thirty. That is what they were built for and they do it well.

Fair enough. Except that a sieve is not analysis, and a ratio is an answer to a question you have not asked yet.

Say two companies both show a return on equity of 20%. One carries no debt at all. Its neighbour carries debt equal to its own equity. That single ratio is describing two different businesses: the first earns its result, the second borrows it and gives it back at the first expensive refinancing. The ratio cannot see this, because all it can see is a quotient.

Price to earnings is worse still, because it is the one popular ratio where half the fraction comes from the market rather than the company. When it falls from 30 to 12, did earnings change or did buyers change their minds? The ratio does not answer that, and everything depends on the answer.

A number only earns its meaning with a comparator and a mechanism: what it is high against, and what specifically drives it. The list of 25 checks every company goes through is an attempt to attach to each number the question that number actually answers.

Where This Method Breaks#

Checking one story in three places works for exactly as long as the three places are independent of each other.

At a company deliberately falsifying its books, they are not. The filing, the presentation and the answers on the call all come off the same desk, and the competitor has no idea. I do not know how to separate a well-executed fraud from a well-run business using only that company's own documents, and I treat that as a risk this method does not remove.

The second limit is gentler. At a bank or an insurer the balance sheet is the business rather than the scaffolding behind it, so comparing profit with operating cash flow says considerably less there than at a manufacturer.

The third applies to recent listings. Two annual reports are the minimum for any historical comparison, and a company one year past its IPO does not have them and will not for another year. What then? You are left with the prospectus and the competitors, which is two sources rather than three, and it is more honest to say so than to pretend the third will turn up.

Three Conditions Instead of a Verdict#

Research ends in a list, not a grade. Once you have been through the statements, the transcripts and the competitor, write down three things:

1. Three sentences that have to stay true for the story to hold. Each with a number, not an adjective: "gross margin does not fall below 34%", not "margin stays healthy".
2. A date against each sentence for checking it. Usually the next results, because that is when the new evidence arrives.
3. The questions you could not answer from the sources. Those go back on the next earnings call.

Where do the candidates for those three sentences come from the first time you look at a company? Partly from the company. Its annual report carries a Risk Factors section, and the SEC says plainly that the risks are generally listed in order of importance. Read as a list, the first three entries are what the company's own lawyers believe could break it fastest, and that ordering is rarely accidental.

The rest is yours. The risk section describes what the company could lose. Your list describes what would stop you believing the story. Those are two different documents, and only the second one changes a decision.

I do not trust research that ends in a verdict, because a verdict cannot be proven wrong. Three sentences with a number and a date can be proven wrong every quarter, and that is the whole of their value.

Where the Method Becomes a Product#

The same route sits behind the Full report: from primary sources, through the statements and the earnings call, to a thesis laid out across 25 points. It costs 100 credits, and the Free plan gets 100 credits a month, which means one complete pass of this method a month costs nothing.

The finished report produces an Investment thesis: the same sentences with measurable conditions, written down for you. Monitoring then watches them, costs no credits and is included in every plan. The Transcript opens from the Earnings brief when you want to check for yourself which question management walked away from. To see what all of it looks like on a finished piece of work, read the sample reports.

Start with one sentence. If after a week of reading you still cannot prove it wrong, you have a thesis. If you prove it wrong in ten minutes, you have saved yourself a year.