How to Research a Stock Before Buying
Quick answer: Research is a written sequence, not a pile of tabs. Understand the business, argue both sides, decide what the price is assuming, and write what would prove you wrong — before you buy. You can do that with a 10-K, the latest earnings release, and a notebook. Tools only speed the same questions up.
This page is for the optional individual-stock step. If you have not funded an account and bought the default index fund yet, start at How to Start Investing in Stocks in 2026. Most people should stop there. If you still want one company you can explain in two minutes, use the sequence below.
You do not need all eight steps on night one. The beginner kit is steps 1–3 and 7. That is enough to own a name on purpose instead of on a thread. Steps 4–6 and 8 are how the work gets serious.
Printable version of the same list: Stock Research Checklist.
Start with the business, not the stock price
The question: What does this company sell, and who actually pays?
Answer it yourself:
- What is the product or service in one sentence?
- Who is the customer — consumer, business, government?
- Which product or segment is most of the revenue?
- How does it get paid — subscription, one-time, advertising, take-rate?
- If the ticker vanished tomorrow, what would customers miss?
Where to look: the latest 10-K, Item 1 (Business). The investor deck is optional and is a sales document.
You're done when: you can explain the business in two minutes to someone who does not follow the market, without looking at the ticker.
Faster version: Framework 01 (Full Company Breakdown) runs this against live filings and news.
Example — AMZN, notes a person could write in an evening:
- Online store + AWS + ads + Prime.
- Shoppers, advertisers, and companies renting compute.
- AWS and ads now do more of the profit work than the store.
- I still could not tell you, off the page, what share of profit is ads vs AWS this year. That's a flag to keep reading, not a reason to skip the step.
Test both sides of the story
The question: What is the honest case for owning this, and the honest case against — at equal length?
Answer it yourself:
- Why might this business be worth more in five years?
- Why might the current price already assume that?
- What is the competitor's version of the same story?
- What would have to go right that is not already in the headlines?
Where to look: 10-K Item 1A (Risk Factors) and the last earnings call transcript. Force the bear case to use company facts, not vibes.
You're done when: the bear case is as specific as the bull case. "It could go down" is not a bear case.
Faster version: Framework 02 (Bull vs Bear + Moat) is built to stop one-sided write-ups.
Check what the stock is actually worth
The question: What is the price already assuming, and what is expensive / fair / cheap in round numbers?
Answer it yourself:
- What did you pay (or would you pay) today?
- What multiple or growth rate does that price imply?
- What does the number look like if growth is 3–5 points slower than that?
- At what price would you be interested, and at what price would you walk away?
Where to look: current price plus the income statement and cash-flow statement. A simple multiple is enough on night one. If you want the implied-growth arithmetic done for you, use the What's Priced In calculator.
You're done when: you can say "this price only works if X happens" in one sentence.
Faster version: Framework 03 (Fair Value Stress Test) builds the optimistic / base / pessimistic range from live data.
Example — AMZN:
- Price is a claim about AWS growth and ad margins staying high.
- If cloud growth cools and ads flatten, the multiple looks different. I don't have a precise fair value in my head yet — that's the point of writing the range down before buying more.
Verify the earnings are real
The question: Did the cash show up, or is the profit an accounting story?
Answer it yourself:
- Is net income close to cash from operations over the last two years?
- Are receivables growing faster than revenue?
- Are "adjustments" and one-time items doing a lot of the work?
- Would the story change if you only trusted cash?
Where to look: income statement next to the cash-flow statement, two fiscal years, in the 10-K or 10-Q.
You're done when: you know whether profit and cash tell the same story.
Faster version: Framework 04 (Earnings Quality Analyzer) is the named version of this check.
Evaluate the people running the company
The question: Would you trust these people with your money for five years?
Answer it yourself:
- How have they allocated capital — reinvestment, buybacks, acquisitions, dividends?
- Do they talk about misses the same way they talk about beats?
- Is compensation tied to something you care about, or to the stock going up this year?
- Are insiders buying or selling shares with their own cash?
Where to look: 10-K discussion of capital allocation and compensation; recent Form 4 filings for open-market buys and sells.
You're done when: you can say yes, no, or "not enough to size this large" in a sentence.
Faster version: Framework 12 (Management Quality Scorecard) and Insider Activity Analyzer — standalone tool (preview): needs a Form 4 feed.
Map the risks
The question: What actually kills the thesis — not what belongs on a 40-page risk list?
Answer it yourself:
- Which Item 1A risks would change your mind if they showed up next year?
- What is company-specific vs. rates, regulation, or a spending cycle?
- Which risks are already obvious in the price?
- Which two risks are you not willing to live with?
Where to look: 10-K Item 1A, then cross out anything that would not change your decision.
You're done when: the list is short and concrete.
Faster version: Framework 13 (Macro Sensitivity Analysis) if you want the map structured. Skip on night one if steps 1–3 and 7 are not written yet.
Build your thesis — and define what breaks it
The question: Why do you own this, and what would prove you were wrong?
Answer it yourself:
- One paragraph: the business, the reason the price is wrong or fair, the time horizon.
- One sentence that means the thesis is dead.
- What you will do if that sentence becomes true — sell, cut, or re-underwrite. Decide before the price moves.
Where to look: your notes from the steps above. Not a new article.
You're done when: both the thesis and the invalidator exist in writing before the buy. A research journal template is fine if you want the page laid out.
Faster version: Framework 05 (Long-Term Investment Thesis) forces the invalidator onto the page.
Example — AMZN:
- Thesis draft: own a piece of a commerce + cloud + ads bundle that still has room to earn more on each customer, as long as AWS growth does not roll over.
- Invalidator draft: AWS growth goes negative for several quarters and ad margins drop with it, and management keeps spending as if nothing changed.
- That is a draft, not a buy ticket.
Check how it fits everything else you own
The question: If this goes to zero, does the rest of the account still work?
Answer it yourself:
- What percent of invested assets would this be?
- Do you already own the same bet under another ticker or an index fund?
- What size keeps you from checking the price every hour?
- Write the max size before you enter.
Where to look: your brokerage positions. For the arithmetic, Position Size calculator.
You're done when: a maximum size is written down.
Faster version: Framework 16 (Portfolio Risk & Fit).
There is no best first stock. There is a first stock you can defend. If you cannot finish steps 1–3 and 7, do not buy the name yet. Own the index and come back.
Where to go deeper
Printable boxes for the same eight steps: Stock Research Checklist.
The sequence filled in on one company: How to Analyze a Stock.
Which of the other frameworks maps to which question: Which Framework, When?
Still on the beginner path: Learning Paths.
FAQ
Do I need to research individual stocks before I invest?
Most people are better served by a low-cost index fund. If you choose to buy individual stocks, a short written research process reduces the odds of buying on a headline alone.
What should I read first, the price chart or the 10-K?
Start with the business: the 10-K and the latest earnings release. The price chart tells you what happened, not why, or what the business is actually worth.
How is this different from asking ChatGPT about a stock?
A chat answer is a one-off response drawn from training data. A structured framework run pulls current filings, prices, and news and repeats the same process for every ticker.
What would count as enough research to buy?
Enough to state, in writing, what the business does, the strongest case against owning it, what the current price assumes, and what would prove the thesis wrong.
What's the difference between researching a stock and analyzing a stock?
Research gathers the facts — filings, numbers, news. Analysis weighs those facts against the current price to form a view. Both belong in the same written sequence. The analysis version of this page is How to Analyze a Stock.
Do I need AlphaLens to do this?
No. You can run the sequence with a 10-K, the latest earnings release, and a notebook. Frameworks 01–06 and 12 speed the same questions up against live filings and prices. Tools only speed the arithmetic.
Which steps are enough for a first look?
Steps 1–3 and 7: business, both sides, what's priced in, and a written invalidator. That is enough to own a name on purpose instead of on a thread. Steps 4–6 and 8 are how the work gets serious.
Is this a recommendation to buy a stock?
No. This page is a process. Nothing on NorrisAI is investment advice or a recommendation to buy, sell, or hold any security.
Where is the empty checklist?
Stock Research Checklist.