Ten ways to analyse a stock

What this is about

There is no single correct way to analyse a company, and anyone who tells you otherwise is selling one. There are methods. Each one asks a different question, each one is blind to something, and the useful skill is knowing which question you are asking and what it cannot see.

This lesson lays out ten methods. For each: what it represents, one sentence on why it is done this way, and what it actually checks. Every one of them has been applied to the same company on the same day so you can read the method and then read the result — ServiceNow, analysed ten ways on 9 September 2026, in the worked example.

Read that page and you will find something more useful than any single method: the ten methods disagree with each other. The earnings method says the company keeps beating expectations. The valuation method says the price stopped caring. Both are correct. That disagreement is not a flaw in the analysis, it is the information — and it is invisible to anyone who only ever runs one method.

Before the ten: the rule that makes any of them worth doing

Write down what would prove you wrong, before you look at the answer.

A method you run until it agrees with you is not analysis, it is decoration. Every panel in the worked example names the evidence that would settle its question — and in that case all ten converge on a single line in a quarterly report. Finding that line is what the ten methods were for.

1 · Full stock analysis

What it represents. The whole company in one view — business model, moat, industry, financial health, risks, valuation and scenarios, in that order.

Why it is done this way. Because a single strong number, whether a growth rate or a cheap multiple, will decide the question by itself if you let it.

What it checks. Whether the parts of the story agree. A company can have excellent revenue growth, a real competitive advantage, and still be a poor investment at a given price — and only a method that looks at all three at once will notice.

In the example: Full analysis. It opens with the line the other nine panels keep returning to: revenue grew 22.19% and net income grew 0.54%.

2 · Financial breakdown

What it represents. Several years of revenue, margins and earnings read as one sequence rather than as separate headlines.

Why it is done this way. Because one year is an anecdote and a trend is evidence, and companies deteriorate slowly enough that any single year can be explained away.

What it checks. Whether the company is getting stronger or weaker. The questions are always the same: is revenue growing, is it growing profitably, and is the profit real.

In the example: Financial breakdown. Watch two things there. The gross margin falls in each of the last two periods, which is the finding. And the 2023 net margin is the highest of the three years while the operating margin is the lowest — a year where the profit came from below the operating line and comparing against it would flatter everything after it. That is the trap Revenue vs profit is about.

What it cannot see. Whether any of it is already in the price.

3 · Competitive advantage — the moat

What it represents. Why this company's profits should still exist in ten years, when anything profitable and unprotected gets competed away.

Why it is done this way. Because the numbers in method 2 describe the past, and only the moat says whether they are repeatable.

What it checks. Five specific sources, one at a time: brand, network effects, switching costs, cost advantage, and proprietary technology. Scoring them together produces a number; scoring them separately produces an argument, which is the part worth having.

In the example: Competitive moat, scored 7 out of 10. The score is at the bottom of the panel on purpose — a moat score is worthless without the reasoning attached, and it is not comparable between two people who each wrote "7".

4 · Valuation

What it represents. The price, treated as a separate question from the company.

Why it is done this way. Because a good business bought at the wrong number is a bad investment, and those two facts have different answers.

What it checks. What the current price already assumes. The version you will see everywhere is the discounted cash flow model, which projects future cash and discounts it back. The version in our worked example runs it backwards — hold the price still and solve for the growth it implies — because that needs fewer invented inputs and shows every step.

In the example: Valuation. The table there says that for the buyer to see the share at a still-generous 30 times earnings in five years, the company must compound revenue at 22.7% a year and hold today's margin. It is currently growing at 22.19%. The growth half of the requirement is being met; the margin half is not.

The warning that belongs to this method more than any other. A valuation model produces a precise number from uncertain inputs, and the precision is what makes it dangerous — you remember the answer and forget that three of the inputs were guesses. Any model whose inputs you cannot see is a number to ignore. What the P/E ratio tells you is the short version of why.

5 · Risk

What it represents. What would have to go wrong, ranked by how much it would cost rather than by how frightening it sounds.

Why it is done this way. Because the size of a loss matters more than its likelihood once the position is large enough to matter.

What it checks. The specific risks — economic, competitive, regulatory, technological, financial — and, crucially, which ones you cannot assess with the data you have. A risk list that omits the category you have no data for is the worst kind, because it reads as complete.

In the example: Risk. The top-ranked risk there is not a disaster scenario; it is the multiple, and the panel points out that it has already fired once inside the data window — the share fell 31.08% from its high while revenue grew 22.19%.

6 · Growth potential

What it represents. Where the next several billion of revenue would come from, and whether there is a plausible place for it to come from.

Why it is done this way. Because growth is the only thing that can justify an expensive share, so it should be examined most carefully exactly when the share is expensive.

What it checks. Whether growth has actually slowed, and which specific mechanisms would produce more — more product into the same customers, price, new customers, new markets. "The market is large" is not one of them.

In the example: Growth potential. Note what the panel refuses to do: it prints no total-addressable-market figure, because published ones are marketing documents as often as they are research. It builds the growth case from the company's own revenue history instead, which is a smaller claim and a checkable one.

7 · The institutional view

What it represents. The share as the people who actually move it see it.

Why it is done this way. Because a private investor is trading against institutions, and it helps to know what they are solving for.

What it checks. Not "is this a good company" but "does this position help the portfolio I have to report on" — which is a different question with different answers. Liquidity, position size, career risk and correlation with everything else held are all inputs, and none of them appear in methods 1 to 6.

In the example: Institutional view, including the uncomfortable one: a manager who owns a well-known name and is wrong has an easier conversation than one who owns something unfamiliar and is wrong. That is a real input into what large funds hold, and pretending otherwise would make the method less useful.

8 · Bull versus bear

What it represents. The two strongest opposing cases, argued from identical figures.

Why it is done this way. Because the case you cannot state is the one that costs you money.

What it checks. Where the disagreement actually is. Two competent analysts looking at the same company almost never disagree about the numbers; they disagree about one interpretation, and finding which one is the entire exercise.

In the example: Bull vs bear. Both sides quote the same figures for six exchanges and it comes down to a single question: is the margin compression investment, or competition? Neither has proof, because the evidence has not been reported yet.

How to use it on your own holdings. Write the bear case for something you own, in full, in writing. If you cannot make it convincing, you do not understand the position well enough to size it.

9 · The earnings report

What it represents. One quarter, read as evidence rather than as news.

Why it is done this way. Because the report is the only regular moment a company is required to show its work, and the price reaction to it tells you what the market was expecting — which is not written down anywhere else.

What it checks. Revenue and profit against expectations, the metrics underneath the headline, guidance, and the market's answer.

In the example: The earnings report, and this is the panel to read if you only read one. ServiceNow beat the earnings estimate in all five reported quarters — and the share was lower two sessions after four of the last four reports: −9.92%, −15.34%, −11.43% and −0.34%.

The lesson inside that. A "beat" is measured against an analyst's published number. The price is set by a market's expectation. Those are different things, and when they diverge it is the second one that moves your money. This is the same point How to read an earnings report makes at greater length, and here is a company where five beats in a row did not help.

Watch the size of the beats as a sequence too: +14.73%, +13.28%, +3.95%, +2.11%, +4.65%. Estimates catch up to a company that keeps clearing them.

10 · The decision

What it represents. The other nine, put together into an action.

Why it is done this way. Because analysis that never reaches a decision is a hobby, and a decision made without the nine is a guess.

What it checks. The short-term outlook, the long-term outlook, the dated catalysts, and the risks — and then, honestly, the questions that are about you rather than about the company.

In the example: The decision, where you will notice something missing. There is no "buy, hold or avoid" verdict, and the panel explains why in its own words: SigniBull is not a broker, an adviser, or licensed anything, and nothing here is a recommendation. But there is a second reason, and it is the better one — the nine panels disagree, and averaging a disagreement into one word destroys exactly the information you came for.

The four questions that panel ends on are worth stealing for anything you own:

1. Which side of the central question do you believe, and what evidence would change your mind? If you cannot name it, you have a preference, not a view.
2. Can you hold this if it falls another 30% while the business improves? For this company that is not hypothetical — it did exactly that between September 2025 and April 2026.
3. Is your horizon longer than the argument? A position with a shorter horizon than its own thesis is a bet on the price, not on the company.
4. How much of what you already own depends on the same question? If the rest of your portfolio is the same sector, this is not a new position — it is more of the one you have.

What ten methods cannot do

They cannot tell you what happens next. Ten thorough analyses of a company that then announces something nobody modelled produce ten wrong answers, quickly.

What they do is narrower and more valuable: they tell you what you are betting on, so that when the share moves you know whether your reason was wrong or merely early. An investor who cannot say what would change their mind will interpret every price move as confirmation, and that is a more expensive habit than any single bad stock.

Where to see this in the app

Chart → Key stats carries the P/E, the market cap and the moving averages that methods 2 and 4 start from — from the same source the worked example quotes, so the numbers on the page and the numbers on your screen agree.

Chart → Financials and Balance sheet are where methods 2 and 3 are done. Read several years at once rather than the latest column; the sequence is the point.

Earnings hub shows expected against reported, with history — which is method 9, and it is the tab where you would have seen the pattern of shrinking beats before the analysis pointed it out.

Stock analysis is where full worked examples are published, including the ten-method analysis of ServiceNow this lesson refers to throughout.

Your paper portfolio is the part that turns any of this into learning. Run the methods, record the position here at the market's own prices in a trade log nobody can edit, and find out over several quarters whether your answer to the central question was right — before it is a decision about money.