Reviewed July 16, 2026. This is the checklist we use when turning a ticker into a research brief. It is designed to produce a decision you can explain later—not a price prediction.

Research standard

Start with primary evidence, write down what would prove you wrong, and separate facts from interpretation. A score or AI summary can organize the work; it cannot replace the filings.

Most stock research fails before the first ratio is calculated. The investor begins with a conclusion—“this company will dominate AI”—and then collects facts that support it. A stronger process begins with a question: what would have to be true for this investment to work, and what evidence would falsify that view?

The one-page output

Before opening ten tabs, create a one-page decision log with eight fields:

Business in one sentence: how the company makes money and who pays it.
Thesis: the two or three developments that could make the business more valuable.
Variant view: what you believe that the current market price may not fully reflect.
Evidence: primary documents or data supporting each thesis point.
Risks: events that would damage the business, not merely move the share price.
Valuation: the expectations embedded in today's price.
Watch items: the next earnings date, filing, product milestone, or regulatory event.
Invalidation: the observable facts that would make you reduce or exit.

If you cannot complete those fields in plain English, you do not yet understand the idea well enough to size it.

1. Understand the business before the stock

Read the company's latest annual report before relying on a summary. For US issuers, begin with the SEC's company filing search. Focus on the business description, segment data, customers, competition, risk factors, and management's discussion of results.

Write answers to five questions:

Which product or service produces the revenue?
Is demand recurring, cyclical, project-based, or dependent on a few customers?
What prevents a competitor from offering the same thing?
Which input costs, regulations, or distribution partners matter most?
What does the company have to reinvest to keep growing?

A business can report fast revenue growth while becoming less valuable if that growth requires heavier discounts, rising stock compensation, or uneconomic capital spending. The income statement alone does not settle the question.

2. Trace the economics through three statements

Use several years, not one quarter. The SEC's company facts data and filed 10-K/10-Q statements are the primary record.

Question	Evidence to inspect	Common trap
Is the core business growing?	Revenue by segment, units, customers, organic growth	Growth caused only by an acquisition
Is growth becoming more profitable?	Gross and operating margin over time	“Adjusted” profit that excludes recurring costs
Does profit become cash?	Operating cash flow, capex, free cash flow	Cash flow lifted by stretching payables
Can the balance sheet absorb a shock?	Cash, debt maturity, interest expense, covenants	Comparing debt without considering lease or pension obligations
Are owners being diluted?	Diluted share count and stock compensation	Celebrating buybacks while shares outstanding still rise

Do not reduce this step to a universal threshold. A 20% margin means something different for a software company and a grocery chain. Compare the company with its own history and a sensible peer group.

3. Check what management said, did, and changed

Read the last two earnings releases and calls. Compare guidance with subsequent results. Then review the proxy statement for incentives, related-party transactions, and executive pay. Insider transactions are filed through Forms 3, 4, and 5, but a sale is not automatically bearish: taxes, diversification, and scheduled plans matter.

The useful question is not “are insiders selling?” It is “does capital allocation match the story management tells?” Watch acquisitions, buybacks, debt issuance, and the per-share result of those choices.

4. Treat valuation as an expectations test

A low multiple is not proof of value, and a high multiple is not proof of overvaluation. Translate the price into operating expectations:

How quickly must revenue grow?
What long-run margin does the price appear to require?
How much reinvestment is needed to reach that outcome?
What happens if growth or margins are merely average?

Use at least three cases—bear, base, and bull—and change the operating assumptions, not just the target multiple. The goal is not a precise target price. It is to learn which assumption carries most of the valuation risk.

5. Search for disconfirming evidence

Spend a separate block of time trying to break the thesis. Search the risk factors for new language. Read a capable competitor's filings. Look for customer concentration, falling retention, working-capital stress, regulatory investigations, product delays, or a shrinking addressable market.

Keep two columns in the decision log: evidence for and evidence against. If the second column is empty, the research process is incomplete.

6. Turn research into a monitoring plan

Research goes stale. Define the handful of events that could change the thesis, then monitor those instead of refreshing the price:

next reporting date and guidance range;
one or two operating metrics management must deliver;
debt refinancing or regulatory milestones;
your valuation or position-size limit;
the explicit invalidation condition.

Use Earnings for dates, News for ticker-specific developments, and Radar to monitor calls and risk flags. These tools reduce the scanning burden; they do not change the underlying evidence standard.

Where AI helps—and where it should stop

AI is useful for triage: grouping news, extracting stated risks, comparing two periods, and suggesting questions you may have missed. It can also hallucinate figures, confuse fiscal periods, or repeat an unsourced claim. Verify every material number against a filing or other primary source.

A safe division of labor is:

AI: organize, summarize, flag changes, and maintain the watch list.
Primary sources: establish the facts.
You: decide assumptions, position size, and whether the risk fits your situation.

The final pre-buy check

Can I explain the business and thesis without jargon?
Did I read the latest 10-K, 10-Q, earnings release, and relevant proxy sections?
Do reported earnings convert into cash over time?
Which assumption matters most to valuation?
What evidence would prove me wrong?
What percentage of the portfolio can I lose without changing my life?
Which event will make me review this decision?

If one of those answers is missing, the next action is more research—not a trade.

Turn the checklist into a live watch list

Use Should I Buy for a structured first pass, then monitor the names that survive your research on Radar. Free for three stocks.

Run a structured check →