Large investors cannot hide. Indian exchanges publish delivery percentage daily, bulk and block deals the same evening, and a full shareholding pattern every quarter, which means a fund building a position leaves a trail in public data. The difficulty is not access. It is that each dataset lies on its own.
The single most common mistake is reading one signal in isolation, because every one of these numbers has an innocent explanation that looks identical to an accumulation signal on a spreadsheet.
Signal one: delivery percentage against its own average
Delivery percentage is the share of a day's traded volume that actually settled into someone's demat account instead of being squared off before the close. A high number means buyers took ownership; a low one means the day was mostly intraday churn.

The absolute level tells you almost nothing. What carries information is the jump relative to that stock's own five or twenty day average, because a large private bank routinely delivers 40% to 60% of volume while a high-beta small cap might sit near 15%. Our delivery percentage explainer works through how to read the change rather than the level.
Signal two: bulk and block deals
This is the only signal that names the buyer. A bulk deal is any trade, or set of trades by one party in a day, crossing 0.5% of a company's listed shares in the normal market, and the exchange publishes the buyer, quantity and price the same evening.
The clearest recent example is a large one. SBI Mutual Fund bought roughly Rs 632 crore of Urban Company shares on 18 March 2026, and the stock rose as much as 16% that session as post-IPO lock-in supply hit the market and the fund absorbed it. A quieter September 2026 example shows the other side: on 18 September, HDFC Mutual Fund bought about 13.9 lakh shares of Entero Healthcare from the Prasid Uno Family Trust, a disclosed deal that named both parties the same evening.
A block deal is different and quieter, negotiated in a separate window so it never touches the live order book, which is why a large block can pass with no price move at all. The distinction matters, and our bulk deal vs block deal piece sets out how each one prints.
Signal three: the quarterly shareholding pattern
Slow, audited, and the one that settles arguments. Every listed Indian company files a shareholding pattern with the exchanges within 21 days of each quarter end, splitting holdings between promoters, foreign portfolio investors, mutual funds, insurers and retail.
Read it alongside the promoter line rather than in isolation. A rising institutional stake next to a falling promoter stake can mean a promoter selling into institutional demand, which is a different story from institutions buying from the public, and our promoter holding piece covers how dilution can fake both readings.
Signal four: open interest, for F&O names
For the roughly 200 stocks with derivatives, positioning is visible daily. Rising open interest alongside a rising price indicates fresh long positions being added, while rising price on falling open interest usually means shorts covering, a distinction our open interest and put-call ratio explainer works through.
This one is the least reliable of the four for accumulation, because a futures long is a rented position rather than an owned one. It is useful as confirmation, not as evidence.
Putting them together
A single signal is a coincidence. Three pointing the same way over several weeks is a position being built. Delivery percentage rising against its own average, a bulk deal naming a recognisable fund, and the next shareholding pattern showing that category's stake up is the full sequence, and the quarterly filing arriving last is what makes the daily data worth watching at all.
The wider flow picture sits behind all of it. On 21 September 2026, provisional exchange data showed domestic institutions buying a net Rs 2,797 crore while foreign institutions sold Rs 576 crore, the same split that has run through most of 2026 and that our FII vs DII explainer breaks down. Accumulation in a single stock is easier to spot when you know which side of that split is doing the buying market-wide.
What none of this tells you is whether the buyer is right. Institutions buy for index inclusion, for redemption management and occasionally for reasons that turn out to be wrong, and a fund's holding period can outlast an individual investor's patience by years. This is general information, not investment advice.
The data has been free and public the whole time. The edge, such as it is, comes from checking four sources before believing any one of them.