The Observer Capital

Investing from India in the world’s best businesses.

🇮🇳 INDIA · 🌍 GLOBAL · ⇄ TRANSMISSION

IPO

The method is the product.

These are the twelve questions this desk asks of every issue. All of them, in full, versioned — never quietly edited. If the checklist changes, the version changes with it and the old one stays readable.

  1. Q1
    Why is this company selling shares now?
    Primary raise for growth, or an exit dressed as one.
  2. Q2
    Who is selling, and how much of what they hold?
    An offer-for-sale is information, not an insult — but it is information.
  3. Q3
    What does the money actually do?
    Objects of the issue, read against the capex the business needs.
  4. Q4
    Is the growth in the numbers the growth in the business?
    Revenue recognition, related-party sales, channel stuffing before the file.
  5. Q5
    Does cash follow the profit?
    Three years of CFO against PAT. Divergence is the tell.
  6. Q6
    What did the last three years of margin do, and why?
    Operating leverage, input cycle, or accounting.
  7. Q7
    How much of the moat is contractual and how much is habit?
    Switching costs you can point at, or a story about brand.
  8. Q8
    What does the balance sheet look like AFTER the raise?
    Pro-forma, not trailing.
  9. Q9
    Who competes, and what happened to their margins?
    The comparable set the RHP does not choose.
  10. Q10
    What is the promoter's history with minority holders?
    Past listings, past dilutions, past promises.
  11. Q11
    What has to be true at the offer band for this to work?
    Reverse the price into the assumptions it needs.
  12. Q12
    What would make this wrong, and by when?
    The falsifier, stated before the listing, not after.

Checklist V3 · current. Earlier versions remain published.

The Vintage backtest

The record of this method applied to past issues, graded — including the ones it got wrong. A backtest that shows only its hits is marketing in a table.

A MISS, IN FULL

The beauty e-commerce issue. The price said no; the business kept compounding. Q11 — “what has to be true at the offer band” — was answered correctly and weighted too heavily against a business whose unit economics were still improving. The override is versioned into V3, which is the only reason the miss was worth having.

The full cohort publishes with its denominator — every issue in the window, not the ones that flatter the method.

Live issues

NO LIVE ISSUE

Nothing is open right now. When an issue is live, its card appears here with the twelve questions answered in public and the score slot locked — the count and the universe visible in both states, because locking hides names, never counts.