Indian Markets · Primary Market · Field Guide
The previous chapters explained what an IPO is. This one is the practical companion: how to actually judge one before you bid — a fixed 20-minute checklist, the real numbers to read, and two true stories of a debut that soared and one that sank.
An IPO is not a lottery ticket — it is a business you can vet in twenty minutes. The market leaves a public paper trail: the DRHP, the day-by-day subscription data, and the grey-market chatter. Learn to read that trail against a fixed checklist, and the hype falls away — you can tell a real business from a good story before your money is on the line.
Chapters 4–5 come straight from Zerodha Varsity. This chapter goes a step further into practice, so most of the method here — the checklist, the scorecard, the AI prompts — is my framework, marked [+]. The two case studies use real, verified public figures (sources at the foot of the page). Prices are historical facts, not recommendations.
[+] = my addition, not in the Varsity source
| IRCTC (2019) | Paytm (2021) | |
|---|---|---|
| Subscribed | 112× | 1.89× |
| NII (HNI) part | 354× | 0.24× (under!) |
| Profitable? | yes, monopoly | no, cash-burning |
| Listing day | +101% | −27% |
On 18 November 2021, India's largest-ever IPO listed. One97 Communications — the parent of Paytm — had raised ₹18,300 crore at ₹2,150 a share. Retail buyers who wanted a piece of "the future of Indian payments" had queued up. By the closing bell that same day, the stock sat at ₹1,564 — down 27%, roughly ₹38,000 crore of shareholder value gone in a few hours. Two years earlier, a far less glamorous railway-ticketing company had listed and doubled on day one. What did the market see in one and not the other — and could an ordinary investor have seen it too, before bidding? Yes. It was all in the numbers. This chapter is how to read them.
The fastest way to learn what matters is to put a hit and a miss side by side. Both were real, both were hyped, both are now textbook cases. Here is the public record.
| IRCTC — the hit | Paytm (One97) — the miss | |
|---|---|---|
| Listed | 14 Oct 2019 | 18 Nov 2021 |
| Issue price | ₹320 (band 315–320) | ₹2,150 (band 2,080–2,150) |
| Issue size | ₹645 cr (100% OFS, govt selling) | ₹18,300 cr (₹8,300 fresh + ₹10,000 OFS) |
| Business | Monopoly on rail e-ticketing; profitable | Payments/fintech; large accumulated losses |
| Subscribed | 112× | 1.89× |
| QIB | 108.8× | 2.79× |
| NII (HNI) | 354.5× | 0.24× (under-subscribed) |
| Retail | 14.7× | 1.66× |
| Listing day | ₹644, +101% | closed ₹1,564, −27% |
Everyone quotes the overall subscription. The revealing number is the NII (Non-Institutional / HNI) line — wealthy individuals writing large cheques with their own money and skin in the game. For IRCTC they bid 354× the shares on offer; for Paytm they bid 0.24× — the category didn't even fill. When the big private cheques stay home, that is the market quietly telling you the price is wrong. Retail was the last to notice in both cases.
Notice what is not the difference. Both had a big Offer For Sale (existing owners cashing out) — IRCTC was 100% OFS. Both were "hot." The real gap was price against substance: IRCTC was a profitable monopoly offered cheap; Paytm was a loss-making business priced for perfection. The checklist below is built to surface exactly that gap.
Before the method, know your sources. An IPO is one of the most heavily-documented events in finance — the hard part isn't finding data, it's knowing which four numbers to look at.
The offer document. Business, 3-yr financials, use of funds, risks, valuation. Filed with SEBI; linked in every broker app.
NSE / BSE publish the live subscription split by QIB / NII / retail, updated through the bidding days.
Zerodha, Groww & co. pre-digest the crux — band, dates, lot size, a summary, and the RHP link.
The DRHP for any Indian IPO lives on SEBI's Public Issues page; live and upcoming issues are on the NSE IPO and BSE Public Issues pages. Your broker's IPO tab is fine for the summary, but the DRHP is the primary source — when the summary and the DRHP disagree, the DRHP wins.
Nobody reads all 300 pages of a DRHP — and you don't need to. A DRHP is written for lawyers and regulators; 90% of it is boilerplate. What you need is eight specific things, each in a known place. This is the whole method.
| # | What to check | What "good" looks like | Where it lives |
|---|---|---|---|
| 1 | Business & moat | Clear, defensible edge (brand, licence, network) | DRHP "Our Business" |
| 2 | Financial trend | Revenue rising, and profit (not just revenue) | DRHP "Financial Info" / broker summary |
| 3 | Use of proceeds | Fresh money for growth, not just OFS exit | DRHP "Objects of the Issue" |
| 4 | Promoter stake | Meaningful holding kept; little/no pledge | DRHP "Capital Structure" |
| 5 | Valuation | P/E in line with (or below) listed peers | DRHP "Basis for Issue Price" |
| 6 | Top risks | No dependence on one client / lawsuit / person | DRHP "Risk Factors" (read top 3) |
| 7 | Litigation | No large pending cases or contingent liabilities | DRHP "Legal & Other Info" |
| 8 | Demand signals | Deep, broad subscription; NII > 1× | NSE/BSE bidding + GMP |
Points 2 and 5 — is it profitable, and is it cheap — carry most of the weight. IRCTC passed both; Paytm failed both. If you only ever check two things, check those. Everything else on the list is there to catch the traps that hide behind a good headline.
An IPO's money can go to two very different places, and the DRHP tells you the split. In a fresh issue, the company creates new shares and keeps the cash — it funds growth. In an Offer For Sale (OFS), existing owners sell their old shares and pocket the cash themselves — the company gets nothing.
A tempting shortcut says "if it's all OFS, the owners are dumping, run away." But IRCTC was 100% OFS (the government divesting) and it doubled. OFS just means you're buying from an existing owner, not funding the company. It's a flag to ask who is selling and at what price — a government meeting a divestment target is different from insiders exiting at a stretched valuation. Read OFS together with valuation, never on its own.
GMP is the unofficial price at which IPO shares (or applications) trade before listing, in an informal market. A GMP of ₹100 on a ₹500 issue hints the stock might open near ₹600. People treat it as a crystal ball for listing gains.
GMP is the tout outside the stadium quoting a resale price before the gates open. It tells you something about the buzz — but it's unofficial, unregulated, thin, and easily talked up by the very people selling. When the match starts (listing), the tout's number often evaporates.
GMP is not reported by any exchange, can be manipulated by operators, and moves on rumour. Paytm carried a positive GMP (around ₹125–135) early in its bidding — and it had collapsed to near zero by listing, right before the stock fell 27%. Use GMP as a mood-ring for demand, never as your reason to buy. The DRHP and the subscription split are real; GMP is a whisper.
The one genuinely tedious step is digging the eight data points out of a 300-page PDF. This is exactly what a language model is good at — feed it the DRHP (or paste the relevant sections) and ask precise, checkable questions. Treat the output as a first draft to verify against the document, never as gospel — models can misread a table, so spot-check the numbers it returns.
From this DRHP, give me a table of total revenue and
net profit (or loss) for the last 3 financial years.
State plainly: is the company profitable, and is
profit growing, flat, or falling? Quote the page
numbers you used.From the "Objects of the Issue" and "Capital
Structure" sections: what is the split between fresh
issue and offer-for-sale (in rupees and %)? Who are
the selling shareholders? What is the promoter's
holding after the IPO, and are any promoter shares
pledged?From "Basis for Issue Price": what is the P/E at the
upper price band, and how does it compare to the
listed peers the company itself lists? Then summarise
the top 3 risk factors and any pending litigation or
contingent liability large enough to matter.AI turns hours of PDF-hunting into minutes, but the judgement is yours. It can pull "P/E is 49× vs peers at 30×" — deciding that's too rich is the part you can't outsource. Always ask it to cite page numbers so you can verify, and cross-check any number that would change your decision.
Reading the eight points is useless if you don't force a decision. Score each one + (good), ~ (neutral / unclear), or − (bad), then apply one rule. Here's the whole framework, and how it would have flagged Paytm in advance.
CHECK GOOD (+) PAYTM (2021)
1 business/moat defensible ~ broad, unproven
2 financials/profit profitable, growing - loss-making
3 use of proceeds funds real growth ~ huge OFS
4 promoter stake kept, unpledged ~ founder small stake
5 valuation vs peers in line / cheaper - priced for perfection
6 top risks no single-point fail ~ regulation, competition
7 litigation nothing large + nothing decisive
8 demand / NII deep, NII > 1x - NII 0.24x, GMP faded
RULE: two or more hard (-) on rows 2, 5, or 8
-> PASS, or bid only a token amount.
PAYTM: minus on 2, 5, AND 8 -> the framework says pass.This scorecard would not have told you IRCTC would exactly double — nobody can. What it does is keep you out of the deals where the numbers are visibly against you. That's the whole game in the primary market: you don't need to catch every winner, you need to avoid the obvious losers you were about to buy on hype.
The chapter in one breath: an IPO leaves a public trail; read eight specific things off the DRHP and the exchange bidding data, weight profitability, valuation, and the NII signal most heavily, let AI turn the PDF into a table, and force a plus/minus verdict — then buy the business, not the listing pop.
// answer
NIIs are non-institutional high-net-worth individuals bidding large amounts of their own money, often after real diligence — they have skin in the game and no mandate forcing them to participate. When that category fills many times over (IRCTC's 354×) it signals genuine conviction; when it fails to fill at all (Paytm's 0.24×) the smart private money is voting no. The overall number can be propped up by one segment, so the NII line is the cleaner read on conviction.
// answer
Not by itself. OFS just means you're buying shares from existing owners rather than funding the company. IRCTC was 100% OFS and doubled. What matters is who is selling and at what valuation — a government divesting to meet a target is different from insiders exiting at a stretched price. Read OFS alongside valuation and the identity of the sellers, never in isolation.
// answer
No. GMP is an unofficial, unregulated, thin grey-market price driven by rumour and easily manipulated by the operators quoting it. It's a rough mood-ring for demand, useful only as one weak signal among many. Paytm's GMP was positive early in bidding and near zero by listing, just before it fell 27% — the DRHP and the exchange subscription split are the data; GMP is a whisper.
// answer
Profitability trend (is it actually making money, and is that growing) and valuation versus peers (is it cheap or priced for perfection). Together they answer "is this a good business, and am I paying a fair price for it." IRCTC passed both and rose; Paytm failed both and fell. The other six items mostly exist to catch traps hidden behind an otherwise good headline.
// answer
In the DRHP section titled "Basis for Issue Price." Companies are required to state the P/E (and other ratios) implied by their price band and to list comparable listed peers with their ratios. It's the single most useful page for a valuation sanity-check — if the company's multiple towers over its named peers with no clearly superior growth or margins, that's a valuation red flag.
// answer
No — the opposite. Heavy oversubscription means retail shares are rationed or allotted by lottery, so you may receive a fraction of your bid or nothing at all. Strong demand is good for the company and a positive signal, but it lowers your odds of a full allotment. Bidding is not owning.
// answer
Use it to extract — feed it the DRHP and ask precise questions (financial trend, fresh-vs-OFS split, promoter pledge, P/E vs peers, top risks) with page-number citations. That collapses hours of PDF-hunting into minutes. But deciding whether a 49× P/E is too rich, or whether a risk factor is fatal, is the judgement you keep. Verify any number that would change your decision against the document itself.
Vet it like one. The pop is a bonus, never the thesis.
Business, financials, use of funds, promoter, valuation, risks, litigation, demand.
Profitable? Cheap vs peers? IRCTC yes/yes, Paytm no/no.
Big private cheques staying home is a loud warning.
Unofficial, manipulable, faded to zero for Paytm.
Score +/~/−; two hard minuses on profit, price or demand → pass.
The reading proves you recognise the signals; these prove you can run the process. Attempt each before opening the walkthrough.
Here is a broker-app summary for a fictional IPO. Using the checklist, list every red flag and give a one-word verdict (Consider / Pass). State which signals you weighted most.
BrightRetail Ltd - IPO summary
Price band : 540 - 570 (issue price 570)
Issue size : 4,000 cr (fresh 400 cr, OFS 3,600 cr)
FY revenue trend : 2,100 -> 2,600 -> 3,050 cr (rising)
Net profit trend : (120) -> (85) -> (60) cr (still a loss)
P/E at upper band : not meaningful (no profit)
Peer P/E (listed) : 28 - 34x
Promoter pledge : 22% of promoter shares pledged
Subscription (final): overall 3.1x
QIB 5.4x | NII 0.6x | Retail 4.0x
GMP : 12% early in bidding -> 2% at closeExpected: at least four red flags, and a Pass (or token-only) verdict.
Step 1 — profitability (check 2). Revenue is rising, which looks nice, but the company is still loss-making every year (−120 → −85 → −60). Losses are shrinking, but there's no profit. Score: −. This is one of the two heavyweight checks.
Step 2 — valuation (check 5). There's no P/E because there's no profit, yet peers trade at 28–34× earnings they actually have. You're being asked to pay up for losses while profitable peers are cheaper. Score: −. That's the second heavyweight, also failing.
Step 3 — use of proceeds (check 3). Only ₹400 cr of ₹4,000 cr is fresh money — 90% is OFS. Remember the nuance: OFS isn't automatically bad. But paired with losses and a rich price, it means insiders are cashing out heavily while the company itself gets little. Score: − in this context.
Step 4 — promoter pledge (check 4). 22% of promoter shares are pledged — the promoters have borrowed against their own stock, a sign of stretched personal finances. Score: −.
Step 5 — demand & the NII tell (check 8). Overall 3.1× looks healthy, but the NII category is 0.6× — under-subscribed. Just like Paytm, the big private cheques passed. And the GMP faded from 12% to 2% during bidding — buzz evaporating. Score: −.
Root read: hard minuses on the exact three rows the rule cares about — profitability, valuation, and demand/NII — plus a pledge flag. The rule triggers on two; we have three.
Verdict: PASS (or token bid only)
Weighted most: check 2 (no profit),
check 5 (priced above profitable peers),
check 8 (NII 0.6x + fading GMP).Why it matters: this is the Paytm pattern wearing a different name — losses, a stretched price, and NII staying home. The checklist turns "it feels hyped" into a defensible, written no.
Open a broker app or the NSE IPO page, pick any open or upcoming IPO, and fill this template. Where a value isn't in the summary, open the DRHP (or use the AI prompts from §1.6). Finish with a scored verdict.
Company : ____
Price band / issue : ____
1 business + moat : ____________________ (+ / ~ / -)
2 profit trend (3y) : ____________________ (+ / ~ / -)
3 fresh vs OFS : ____________________ (+ / ~ / -)
4 promoter + pledge : ____________________ (+ / ~ / -)
5 P/E vs peers : ____________________ (+ / ~ / -)
6 top 3 risks : ____________________ (+ / ~ / -)
7 litigation : ____________________ (+ / ~ / -)
8 subs + NII + GMP : ____________________ (+ / ~ / -)
VERDICT (rule) : ____________________Expected: a completed sheet where every row is filled from a real source and the verdict follows the rule — not a gut feeling.
Step 1 — pick and pull. We'll run the sheet retrospectively on IRCTC (2019) using its real figures, so you can see a completed, honest example. In practice you'd do this during the bidding window using live data.
Step 2 — fill each row from the record. Business: a legal monopoly on rail e-ticketing plus catering and packaged water — a genuine moat (+). Financials: profitable and growing (+). Structure: 100% OFS (government divesting) — neutral, judged with valuation (~). Promoter: the Government of India, no pledge concern (+). Valuation: the price band was widely seen as modest for a profitable monopoly (+). Risks: regulatory/government-linked, but no single-point failure (~). Litigation: nothing decisive (+). Demand: 112× overall with NII at 354× — extraordinary conviction (+).
Company : IRCTC Ltd (2019)
Price band / issue : 315-320 / 320
1 business + moat : rail e-ticketing monopoly (+)
2 profit trend (3y) : profitable & growing (+)
3 fresh vs OFS : 100% OFS (govt divestment) (~)
4 promoter + pledge : Govt of India, no pledge (+)
5 P/E vs peers : modest for a profitable moat (+)
6 top 3 risks : regulatory / govt-linked (~)
7 litigation : nothing decisive (+)
8 subs + NII + GMP : 112x overall, NII 354x (+)
VERDICT (rule) : no hard (-) on 2/5/8 -> CONSIDERStep 3 — apply the rule. Zero hard minuses on the heavyweight rows (2, 5, 8); the OFS and risk rows are neutral, not negative. The framework says Consider — and the stock then listed at +101%.
Why it matters: the sheet forces you to source every claim and reach a verdict by rule, not vibe. Run it on a live IPO and you've converted this whole chapter from something you read into something you can do.