How to Research a Stock End-to-End: A Process, Not a Prediction

Direct Equity — Arc 2, Article Three

The Shape of the Work

Research is a funnel, not a search

Beginners imagine stock research as a hunt: find the good one, buy it. It is better understood as a funnel designed to kill ideas cheaply, so that only survivors consume real time. Each stage costs more than the one before and eliminates most of what enters it. The discipline is in the killing, not the finding.

Budget honestly before you start: done properly, a first-time study of a single company takes ten to twenty hours. That number is a feature rather than a warning. It is what the process is protecting you with, and it shrinks considerably as sector knowledge compounds.

Stages One and Two

From a long list to a short one

Where ideas legitimately come from. Screeners, covered in a moment. Your own circle of competence — the industries you work in or genuinely understand, which is an underrated edge and the only place an individual routinely knows something a fund manager does not. The annual reports of the suppliers, customers and competitors of companies you already follow, which is how one good idea breeds three. And high-quality investor letters, read for the reasoning rather than the holdings.

Illegitimate sources feed the same funnel: tips, Telegram calls, “five stocks to buy” videos. The objection is not snobbery. It is that legitimate sources arrive with reasons attached, which the later stages can test, while tips arrive with urgency attached, which is engineered to prevent testing. If an idea cannot survive being examined slowly, it was never an idea.

Using a stock screener properly. A screener is a filter, not an oracle. A sensible starter filter simply puts the ratio discipline from the previous article into a set of rules — for illustration: ROCE above 15% for five consecutive years, debt-to-equity below 0.5, sales growth above 10%, positive operating cash flow in each of the last five years, and a market capitalisation above whatever liquidity floor you have set. What emerges from that is a reading list, not a buy list — typically twenty to forty names, most of which the next stage eliminates. Change one criterion and the list changes completely, which is itself a useful reminder of how little a screen actually knows.

The thirty-minute triage. One question governs this stage: can I explain how this company makes money, in two sentences, without using words I would have to look up? If not, it goes to the “too hard” pile without guilt or apology. That pile is a legitimate portfolio decision, and for most beginners it should contain lenders, insurers and sprawling conglomerates. If yes, run the quick screen from the statements article — five-year revenue, margin and profit trends, operating cash against reported profit, the debt trend, receivables against sales, and the auditor’s report scanned for qualifications — plus two governance checks covered below: promoter pledging and the shareholding pattern. Perhaps one name in five survives.

Stage Three

The deep dive, where the real work happens

Three to five annual reports, oldest first. The sequence matters more than the number. Reading oldest-first means you watch management’s promises meet reality in order: the capacity expansion announced in year one, quietly rescheduled in year three, absent from year five. The segment that got reclassified. The narrative that shifted. Chronology exposes what any single report is designed to conceal, and no amount of reading the latest report alone can substitute for it.

Four to eight quarters of earnings call transcripts. Companies file these with the exchanges and post them in their investor sections, free. Calls reveal what polished reports do not: which questions management answers crisply and which they deflect, whether guidance stays consistent from call to call, and how they discuss their own failures. Pay attention to the analysts’ questions as much as the answers — they tell you what informed money is currently worried about, which is intelligence you would otherwise have to generate yourself.

Credit rating rationales. Rating agencies publish their rationales free, and they read the debt documents you never will. A rating rationale is a professional risk memo written by people paid to be pessimistic, and a history of downgrades is a pre-written bear case handed to you at no cost.

Industry structure. Ask the boring questions in plain clothes: who holds the pricing power here — this company, its suppliers, or its customers? What actually stops a well-funded new entrant? Is the industry’s capacity discipline holding, or is everyone building at once? An average company in an industry with good economics frequently beats a heroic company in a brutal one, and this is the stage where that becomes visible.

Peer comparison. Run the full set of ratios side by side against two or three competitors. Where this company is superior on margins, ROCE or working capital, you need an explanation you can articulate. “Better company” is not an explanation. “Owns its distribution across the North and runs twice the dealer network” is. Unexplained superiority is either an edge you have not understood or a number you should not trust, and both need resolving before money moves.

Stages Four and Five

Checking the price, then writing it down

This pillar deliberately treats valuation as a reality check on the price rather than an exercise in precision, and there are three honest tools for it. The first is the full set of ratios: what do today’s multiples imply about future growth and returns, and is that plausible against this company’s own history? The second, and the most useful habit in the whole article, is reverse thinking — instead of estimating what the company is worth, ask what it must deliver for today’s price to produce a decent outcome. If the answer is “20% compounding for a decade, executed flawlessly, with no competitive response,” you have learned something precise: the price is carrying heroic assumptions. The third is historical range context: where does today’s multiple sit against this company’s own five to ten year band, and what has genuinely changed to justify a departure from it?

Detailed discounted cash flow modelling has its place, and it is a skill worth acquiring eventually. For most private investors, though, its false confidence costs more than its rigour earns — a model with fourteen assumptions produces a number that looks authoritative and moves 40% when you change any two of them.

The written thesis: one page, non-negotiable. Before any purchase, in your own words, answer four questions. What does this company do? Why will it be worth meaningfully more in five years — in three bullet points? What are the three biggest risks? And what specific developments would prove me wrong and make me exit?

The value of this page is lopsided: it costs twenty minutes to write, and it is the only defence you have against your own memory quietly rewriting the story later. Two years later, when the stock is down 30% and every instinct is manufacturing new reasons to hold, the page you wrote tells you whether the reasons broke or only the price did. That distinction is the whole difference between investing and reacting, and nobody has ever reliably drawn it from memory.

The Warning Signs

Red flags, compiled from Indian market history

These are drawn from the pattern that repeats in Indian company collapses. Checking accounts for signs of manipulation like this is usually called forensic analysis, and none of it needs an accounting degree. One flag means investigate. Two or three together means the “too hard” pile is calling — and they do cluster, because the underlying cause, which is management you cannot trust, expresses itself everywhere at once.

Red flag	Why it matters
Operating cash persistently far below reported profit	Profits that never become cash — the master flag
Receivables or inventory growing much faster than sales	Revenue manufactured by pushing goods onto dealers (channel-stuffing) or by booking sales that never happened
Promoter pledging — a large share of promoter holding pledged as loan collateral	A leveraged promoter is a forced seller in waiting. Price falls trigger margin calls, which trigger more selling. Disclosed quarterly; check it every time
Auditor resignation, frequent auditor changes, or qualified opinions	The professionals with inside access heading for the door
Heavy related-party transactions	Value quietly transferable to promoter-owned entities — read every line of the disclosure
Fresh equity or debt raised every year despite reported profits	The treadmill: dilution as a business model
Capital work-in-progress that never completes	Capital expenditure as a place where profits go to disappear
Contingent liabilities large relative to net worth	Off-balance-sheet obligations one court order away from becoming real
Promoter shareholding declining steadily, with no institutional holders	The people who know it best reducing; the people paid to look staying away
Sudden diversification into unrelated glamour sectors	Capital allocation by press release
Rising royalty or brand-fee payments to promoter entities	A legal siphon — watch the trend, not the level
Depreciation policies or asset lives out of line with peers	Earnings manufactured through accounting assumptions

The shareholding pattern, filed quarterly with the exchanges, compresses several of these into a single page: promoter stake trend, pledged percentage, institutional presence, and one subtle tell — a sudden surge in the number of small retail holders, which is often what a promotion campaign leaves behind rather than what causes one. Our article on market scams and finfluencers traces how those campaigns are run.

After the Purchase

Monitoring: research does not end at the buy order

Owning converts research into a maintenance schedule, and the schedule should be light enough that you actually keep it. Quarterly: results plus the earnings call transcript, read against your written thesis — under an hour. Annually: the new annual report, a re-read of the thesis, and one honest question: would I buy this today, at this price, knowing what I now know? Event-driven: auditor changes, chief executive or finance director departures, rating actions and jumps in promoter pledging each trigger a same-week review, regardless of what the price is doing.

What monitoring is not is daily price-watching, which generates activity and anxiety and precisely zero information about the business. The stock does not know you own it. The business does not report daily.

Cap the workload honestly. At ten to twenty hours per new name plus a few hours per name per year, a working person can maintain genuine research quality on perhaps eight to fifteen companies — which is not coincidentally the portfolio size the next arc arrives at. And the alternative to doing this work is not doing it badly. It is the index fund, chosen deliberately, with self-respect entirely intact.

Key Takeaways

---

Frequently Asked Questions

Your questions answered

Ten to twenty hours per stock. Seriously?

For a first-time full study of a company you may hold for years, yes — and it shrinks substantially with practice as sector knowledge compounds. If that budget is unrealistic for your life, the honest conclusions are a smaller direct portfolio or the index route, and both are wins. The failure mode is wanting stock-picking’s outcomes on tip-following’s time budget.

Where do I find transcripts and rating rationales without paid tools?

Companies file transcripts and recordings with the exchanges and post them in their investor sections — the NSE corporate filings pages carry them alongside results and shareholding patterns. Rating agencies publish rationales free on their own sites. Screeners aggregate links to most of it. The entire deep-dive kit is public and costs nothing; the barrier was always the sitting down.

What if a company passes everything but looks expensive?

Then the research is not wasted — it goes onto a watchlist with your written price context, and a GTT order can wait for your zone mechanically while you do something else. Quality identified but not purchased is a completely successful research outcome. Forcing a purchase just to “use” the research is the sunk-cost trap (spending more because of what you have already spent), and it is expensive.

How do I research a company with no annual report history?

Through the offer document. A DRHP or RHP is a forced mega-disclosure — financials, risk factors written by lawyers who fear liability, related parties, and the intended use of proceeds. Our IPO article covers that specific reading. The honest headline is that less history means a much wider range of things that could happen (wider error bars, in the usual phrase), which is an argument for smaller position sizes rather than more confidence.

Can I shortcut this with someone else’s research report?

Reports from SEBI-registered research analysts are legitimate inputs — read them for the facts and especially for the bear case, not for the target price. But borrowed conviction fails at exactly the moment you need it. At minus thirty percent, only your own thesis can tell you whether to add, hold or exit. And material shared by unregistered tipsters is not research at all; you can verify anyone’s registration on SEBI’s own site in under two minutes.

Keep Learning