Annual figures curated from annual reports / exchange & MCA filings (consolidated unless noted): listed brokers (some at listed-parent / group level); bank-owned brokers (HDFC, Kotak, Axis, SBI Securities), disclosed in the parent bank's annual report; and large privately-held brokers (Groww — listed Nov 2025; Zerodha, Upstox — self-disclosed / MCA). Smaller private brokers' financials sit behind MCA / paid sources and aren't shown.

Latest-FY revenue / net profit ÷ the latest half-yearly headcount the member reports to the exchange. Approximate — financials may be group/consolidated while headcount is the broking entity's filing; implausibly-high ratios (entity-scope mismatch) are dropped.

Latest-FY revenue / net profit ÷ the broker's current NSE active-client base — the business-model lens. Discount brokers earn a little from each of a great many clients; full-service & wealth houses earn far more from fewer. Approximate: financials are company-wide (may be group/consolidated) while the client count is NSE-only, so it flatters multi-venue and non-broking-heavy houses. Never summed across venues.

Broker-disclosed client assets, from each broker's homepage / investor-relations / results / DRHP. Not in any CDSL/NSDL report — client demat holdings sit at the depository, off the broker's balance sheet, so most brokers never publish it. Definitions vary (equity demat · equity + MF · total incl. cash), shown per row with a confidence flag — ordered by size but not strictly comparable. As-of dates differ too — each row shows its own (from live homepage snapshots to older annual reports), so the vintages aren't aligned. A ~ marks a low/medium-confidence or undefined figure. Never summed with active clients or demat-account counts.

Long-term grade (with outlook) + short-term grade, from the latest ICRA / CRISIL / CARE / Infomerics rationale for each broker. Long-term scale (high→low): AAA · AA+ · AA · AA− · A+ · A · A− · BBB+ · BBB… ; short-term: A1+ · A1 · A2+ · A2 · A3+ · A3 · A4. The rating is on the rated legal entity's debt / bank facilities (usually the broking subsidiary; group rating where noted) — an independent read on financial strength, not a guarantee of any broker service. Unrated brokers (most) simply aren't shown. Gearing = total debt ÷ tangible net worth.

Change in NSE active client codes from the same month a year ago. Net adds & share shift are absolute moves (large brokers dominate); the growth-% view is limited to brokers with ≥50k clients a year ago so a tiny base can't top the chart. Share shift is in percentage points of the all-NSE active base; it is measured against the nearest month with a published exchange-wide total (so its reference month can differ slightly from the others).

Pearson correlation over the monthly era; growth rate (not absolute adds) so broker size doesn't drive it. Needs a ≥50k client base each month and ≥18 paired months. Monthly client-growth is noisy, so correlations are modest — it's the ranking that's the signal.

Activation = NSE active client codes ÷ CDSL demat clients (other than closed), both taken as of the same fiscal year-end (March) — the demat count is a year-end snapshot, so the active-client count is read at that same March rather than the latest month (which would inflate the ratio by comparing newer actives to an older base). Use the size toggle to compare like-with-like — a broker with ~1 lakh CDSL clients and one with ~1 crore sit in different leagues. The board is anchored to the latest year-end (shown above); brokers whose CDSL data hasn't refreshed to it yet are held back until it lands. This is a CDSL-only measure, so it's meaningful for CDSL-primary (mostly discount) brokers; brokers whose demat sits mainly on NSDL show more active traders than CDSL accounts (>100%) and are excluded rather than shown wrong.

Google Play "Downloads" tier (coarse bands), read from each app's store listing. Star ratings are not tracked (too volatile to source reliably).

Cash-equity & F&O client counts are from the NSE half-yearly member filings; commodity is active UCCs on MCX. Currency & debt aren't split per broker in the filings — only segment participation (how many members are active in each), shown in the participation board above.

Largest by headcount

Self-reported headcounts in the half-yearly filings; members with a one-off reporting-basis change (headcount halving/doubling in a single period) are excluded. Traditional and bank-owned brokers trimming as they digitise; troubled/exiting members show the largest cuts.