How to research a stock before you buy

A screen can produce candidates. It can't tell you which ones to actually buy. This is the workflow for closing that gap on a single stock.

The whole sequence takes 15 minutes per stock once you know what you're looking at. The point isn't to find a reason to buy — it's to surface enough information to know whether to buy.

The 15-minute sequence

1. 1

   Open the stock page and read the strip

   Type the symbol into the command palette (cmd+k) and hit enter. You land on the stock page.

   The conviction & checklist strip directly under the hero is the 30-second read. Note three things:

   • The conviction band and stance (e.g. Strong·Bullish)
   • The checklist score (out of 100)
   • Today's 1D return

   If conviction is Weak/Neutral and the checklist score is below 40, you're already in "this needs a strong reason to buy" territory. If both are healthy, you've earned the right to spend the next 14 minutes.

2. 2

   Expand the conviction breakdown

   On the Overview tab, expand the Conviction Breakdown section if it isn't already open.

   You're checking what's driving the verdict. If the stance is bullish but driven only by a single factor (e.g. a one-time bulk deal), that's a weaker signal than a stance built on three or four converging factors. Look at the breakdown for:

   • Which factors are firing
   • Which factors are silent
   • Which factors are opposing the verdict (these create the engine's "conflict penalty")
3. 3

   Skim the checklist

   Below the conviction section, the investment checklist scores the stock across four fundamentals factors: valuation, growth, returns & margins, financial health.

   Pay attention to the fails. A stock with a 70 overall score but a returns & margins fail is telling you the price is up but earnings quality isn't there: that's a different bet than one where everything is in green.

4. 4

   Check valuation against peers

   Switch to the Peers tab. The default sort is market cap descending.

   You're comparing PE, PB, ROE, and ROCE. Two questions:

   • Is the stock trading at a premium or discount to its peers?
   • If at a premium, is that justified by better profitability (higher ROCE)?

   A premium on the same profitability is expensive. A discount on the same profitability is interesting.

5. 5

   Read the last 5 years of financials

   Switch to the Financials tab. Income statement is the default sub-tab.

   You're looking for the shape of the business, not exact numbers:

   • Is revenue growing? At what rate?
   • Are operating margins stable, expanding, or contracting?
   • Is PAT growth tracking revenue, or are margins doing the work?

   For long-term investors, this is the most important tab. For short-term traders, it's a sanity check — never buy a falling-knife stock without confirming the business isn't actually deteriorating.

6. 6

   Read recent filings

   Switch to the Filings tab. Read the last 3 months of announcements.

   What you're looking for:

   • Earnings filings — read the AI summary; if growth/margins are recent and surprising, the price action you're seeing makes sense
   • Material updates — M&A, capex, fundraising, regulatory orders
   • Management commentary — promoter pledge changes, board reshuffles, MD/CEO statements

   If there's recent material news you didn't know about, that often explains the conviction reading.

7. 7

   Ask the AI chat the question you can't answer

   Press cmd+j. The chat has the current stock's context pre-loaded.

   Ask whatever you still don't know. Practical examples:

   • "How has this stock performed compared to its sector index over the last year?"
   • "What are the main risks an investor should be aware of for this company?"
   • "Compare this company's profitability to its closest peer."

   The chat can use the same data you're already looking at, so it tends to be useful for synthesis questions rather than fact lookups.

What you've earned at the end of 15 minutes

You should now be able to write down, in two sentences:

1. The bull case — why this is interesting at the current price
2. The bear case — what would have to be true for this to be a bad buy

If you can write both, you can decide. If you can't write either, you don't know enough yet — keep researching, or skip the stock.