How to decide which stocks to research

Capture the claim before opening more tabs#

Suppose someone shares a headline about Northstar's cooling module for data centres. Write down the claim, where you heard it and the date. Then find the original company announcement.

The distinction matters. “Northstar announced a module” is a verifiable event. “Northstar will dominate cooling” is a prediction. If the argument depends on unnamed customers, a rumoured contract or an unexplained screenshot, those gaps belong in the record.

For a real company, confirm the legal name, ticker, exchange and security type. A related business or similarly named company is not a substitute for the one mentioned in the announcement.

Explain the business without the headline#

Northstar sells components to data-centre operators and equipment makers. The new module is one part of that business. Read the annual report's business and segment sections to understand the existing revenue base, customers, competition and key dependencies.

If you cannot explain who pays and what they receive, your next task is learning the business. Do not compensate by collecting more valuation ratios. A low ratio is difficult to interpret when the earnings behind it are unclear.

You can also park the company. Research time is limited, and an unfamiliar business need not become an urgent project merely because its shares are moving.

Turn the announcement into a question#

Our first question is: will customers reorder the cooling module, and will that growth produce cash?

That leads to a small set of useful evidence: module sales, repeat customers, margins, receivables and cash flow. It also gives the next company report a purpose.

Compare that with “Is this the next big stock?” The second question invites opinions but gives you little guidance about which document to open. The first can become more precise as you learn.

Run an initial financial check#

Northstar's latest quarter shows revenue of US$462.000m, against US$420.000m in the comparable prior-year quarter. Gross margin rises from 36.0% to 38.0%. Yet free cash flow, defined here as operating cash flow less capital expenditure, falls from US$33.600m to US$28.800m. Receivables rise from US$88.000m to US$121.000m.

These facts justify questions, not an automatic rejection or endorsement. Higher receivables can reflect several causes. Read the notes before calling it a collection problem.

The module generated US$36.000m, about 7.8% of latest revenue. That establishes some sales, but not their durability. The example has not disclosed module margin, repeat orders or customer concentration. Keep those unknowns visible.

Decide what kind of research comes next#

Three labels can help manage your attention:

• Deepen: there is a specific, material question and existing filings can help answer it. Open those documents now.
• Watch: a future disclosure could answer an important missing question. Record the event and what you need from it.
• Park or drop: the premise is false, you cannot identify a useful question, or the company is outside the work you can responsibly undertake.

These are research labels. They do not mean buy, hold or sell.

Northstar earns a closer look because the product has reported sales and the financial statements show a discrepancy worth understanding. After reading the available notes, we may still need to watch for repeat-order evidence. Moving from deepen to watch is normal; the labels describe the current task.

Compare candidates by the work they require#

If you have five interesting companies, prioritise the ones with a clear question, reliable sources and a manageable next step. A complex refinancing case may require skills or time you do not have. A simpler product question may be easier to investigate, without necessarily being a better investment.

Avoid ranking candidates solely by excitement, recent price gains or a social account's confidence. Also avoid giving every company an artificial numerical score. A short explanation of why one deserves attention can be more useful than a precise-looking total built from guesses.

Save enough to return without starting over#

Record the original claim, source, business explanation, financial discrepancy and next question. The free research journal provides a compact place for this.

If the missing evidence requires another report, continue with how long to watch a stock. If the existing documents can answer it, the full company walkthrough shows how to connect the statements, risks and valuation.