Capital, not capex.
Banks score on return on assets, book-value compounding, and capital cushion — the things that actually matter when revenue is interest income and inventory is loans.
Every actively-traded NSE stock, scored against its true peers. 200 companies. 41 clusters. One number, recomputed weekly — and never edited.
We don't score a bank the way we score a paint company. Each cluster has its own recipe — tuned for what an analyst in that space would actually look at.
Banks score on return on assets, book-value compounding, and capital cushion — the things that actually matter when revenue is interest income and inventory is loans.
EBITDA margin and capacity utilization tell you whether prices are firm and whether new plants are paying off. P/E for cement is mostly noise.
Operating-margin trend and cash conversion separate the compounders from the deal-grabbers. Revenue growth alone is a trap.
The way an analyst would actually think about a business — not a dump of every ratio in the book.
Returns on capital, multi-window CAGR, growth consistency, cash conversion, balance-sheet discipline. The question we're answering is whether the business has built durable economics — or just had a good year.
P/E, P/B, EV/EBITDA, FCF and dividend yield — but always relative to peers in the same cluster, never the whole market. A 25× P/E in pharma means something very different from a 25× P/E in cement.
Price momentum across 3, 6, 12-month horizons relative to the market — blended with latest-quarter earnings momentum. The two together separate hype from fundamentals turning.
Five axes, sorted. The strongest dimensions rise to the top — without you having to read a footnote.
Anyone can ship a stock screener in a weekend. None of these three can be cloned in a sprint.
Every Monday at 09:30 IST we freeze the score for every covered name and write it to an append-only ledger. We don't edit history — if we were wrong, you can see we were wrong. A competitor launching tomorrow can backfill price data, but they cannot backfill our judgment from week one.
Every score ships with a paragraph that explains what changed and why it matters. The model behind it gets tuned on analyst corrections and reader feedback — patient input a generic prompt cannot replicate.
Watchlists, screens, and reads — anonymized into a demand-side dataset that's uniquely ours. We feed it back into scoring and surface what's moving. We don't sell it. Every visit makes the next score a little better.