IPO Subscription Data Explained

Two numbers sit behind every IPO subscription figure you see on an exchange website: the quantity of shares bidders have asked for, and the quantity of shares actually on offer to that group of bidders. Divide the first by the second and you get the familiar "3.4 times" or "58 times" headline. Everything else on the page - the category rows, the cumulative totals, the refresh through the day - is a rearrangement of those same two numbers. Reading the table well is therefore mostly a matter of knowing precisely which shares sit in the denominator, which bids are allowed into the numerator, and what neither of them has been checked for.

What the denominator actually contains

The denominator is not the full issue size. Anchor investors are allotted shares before the public bidding window opens, and their portion is carved out of the institutional bucket, so anchor money never appears as live demand in the subscription table. What remains is the net offer to the public, divided into fixed buckets: qualified institutional buyers, non-institutional investors, retail individual investors, and any reserved portion for employees or for existing shareholders of a listed parent company.

Those proportions are not chosen at random and they are not identical across issues. They are printed in the red herring prospectus and depend on which regulatory eligibility route the company used to come to market. That is why two IPOs open in the same week can have very different category weights, and why a subscription number from one issue is not directly comparable with the same number from another. The sibling explainers on QIB, NII and retail categories and on anchor investors cover who belongs in each bucket and how the anchor round is settled.

The practical consequence is that each row has its own denominator. Take an illustrative offer of 1 crore shares in which 30 lakh shares are set aside for retail. If retail bids arrive for 2.1 crore shares, the retail row reads 7 times, because 2.1 crore divided by 30 lakh is seven. Those same bids equal only a little over two times the entire issue. Judging a category multiple against total issue size, rather than against that category's own reserved quantity, is one of the most frequent beginner errors.

What the numerator counts, and what it never checks

The numerator is shares bid for. It is assembled from bids uploaded to the exchange bidding platform by brokers, banks, registrars to an issue and other registered intermediaries. Bids placed at the cut-off option, which retail applicants commonly use, are counted at the upper end of the price band, because that is the highest price the bidder has agreed to accept. Bids entered below the floor price are not accepted by the system at all, so they never enter the count. The price band mechanics behind this are covered separately.

What the numerator does not do is verify anything. A bid appears in the table the moment it is uploaded. Whether the applicant approved the UPI mandate, whether the money is genuinely blocked in the bank account, whether the PAN is unique, whether the demat account details match the applicant name - none of that is tested while bidding is live. Verification happens after the issue closes, during registrar reconciliation, and a meaningful number of bids fall away at that stage. The live figure is best understood as an upper bound on valid demand rather than a confirmed count of it.

Reading the category rows one at a time

The total row is an average weighted by the size of each bucket, which means it is dominated by whichever category is largest. A headline total can be pulled up almost entirely by institutional demand while the row that governs your own outcome barely moves. Read the rows individually and in this order: the one you applied in first, then the others for context.

Qualified institutional buyers

This row shows bids from mutual funds, insurers, banks, pension funds, foreign portfolio investors and similar entities, excluding the anchor allocation already settled. A portion of the institutional bucket is typically reserved for domestic mutual funds. Institutions frequently place or finalise bids very late in the window, so this row can look thin for most of the issue and then move sharply. In a book-built issue the institutional portion carries a minimum subscription condition; if it is not met the offer does not proceed. The exact condition is a regulatory parameter, so confirm the current requirement in the offer document or with SEBI rather than relying on a remembered figure.

Non-institutional investors

The non-institutional bucket covers applications above the retail ceiling, and it is itself split into two sub-buckets by application size, with the split proportions set by regulation. Exchange pages usually show both sub-rows. Because non-institutional money is blocked for the whole period and often funded at a cost, these bidders have a strong incentive to bid as late as possible, which is why this row is the one that most often explodes in the final hours. Rules on how much of a non-institutional application must be funded upfront have changed over the years and materially affect how inflated this row can become, so treat older commentary about it with caution.

Retail individual investors

Retail bids are those up to the application value ceiling defined in regulation. For most individual applicants this is the only row that matters directly, because when the retail portion is oversubscribed the allotment rule is to give the minimum lot to as many applicants as possible and then draw lots. A retail multiple of seven therefore translates roughly, though not exactly, into odds of about one in seven for a single-lot application. The oversubscription explainer works through why that translation is approximate.

Reserved portions

Some issues carve out a small portion for employees, occasionally at a discount to the issue price, and some carve out a portion for shareholders of a listed parent when a subsidiary comes to market. These rows are small in absolute terms and are settled within their own quota. They sometimes remain undersubscribed even when the main categories are heavily bid, but eligibility is checked strictly against record dates and identity details, and a failed eligibility check turns the application into an ordinary rejection.

The day-wise build-up and why its shape matters

Exchange subscription pages refresh several times through each bidding day and the figures are cumulative, not daily. Nothing resets overnight. The information is therefore in the shape of the curve rather than in any single snapshot: an issue at two times on the opening afternoon and an issue at two times an hour before close are describing very different situations, because the second one has almost no time left in which to change.

Late loading is structural, not accidental. Non-institutional bidders pay a funding cost for every day their money sits blocked. Institutions want to see how the book is shaping before committing. Retail applicants, many of whom watch the same table you do, often wait for a signal from the other categories. Each of these incentives pushes demand towards the last session, which is why a slow start says less about eventual demand than beginners assume.

• Which category moved first, and whether the rest followed it or ignored it.
• Whether retail demand accumulated steadily across sessions or arrived in one block after a burst of publicity.
• Whether the total figure is being carried by a single bucket while the others stay flat.
• Whether reserved employee or shareholder portions filled at all, since insiders and existing shareholders see the company from closer range.
• How much of the final-session jump came from categories whose bids can still be revised or withdrawn under the applicable rules, and how much from those whose bids cannot.

That last point deserves a note. The regulations treat categories differently on whether a bid may be revised downward or withdrawn after it is placed, and retail applicants generally have more flexibility than institutional and non-institutional bidders. Because the rules here have been amended over time, check the current position in the offer document rather than assuming.

Applications, shares and rupees are three different units

A subscription table generally reports shares bid for, and often also the rupee value of those bids. It does not report the number of applications. Yet the number of applications is what actually drives retail odds, because the retail lottery is run per application, not per share.

Work through an illustrative case. Suppose the retail portion is 30 lakh shares and the lot size is 60 shares. The portion therefore holds 50,000 lots. If 3.5 lakh retail applications each ask for one lot, retail demand is 2.1 crore shares, the row reads 7 times, and roughly one application in seven can be served with a single lot. If instead 50,000 applications each asked for seven lots, the row would still read 7 times, but every applicant could receive one lot and no draw would be needed for the first lot at all. Identical headline, completely different outcome. The application count only becomes public later, in the basis of allotment document.

Misreadings that cost beginners money

1. Using the total row to estimate personal allotment odds when only the row for your own category is relevant.
2. Treating a high multiple as a verdict on business quality. Subscription measures willingness to bid inside a price band under a deadline, not durable value.
3. Assuming the closing figure is the number the registrar will work with. Reconciliation removes bids with unapproved mandates, duplicate PANs, insufficient blocked funds and mismatched demat details.
4. Comparing multiples across issues of very different sizes. A small offer can post a large multiple on a modest rupee amount, while a large offer at a low multiple may have absorbed far more capital.
5. Reading subscription and grey market chatter as two independent confirmations. They influence one another, and the grey market carries its own separate risks.
6. Bidding for more lots in an oversubscribed retail category in the belief that it improves the chance of getting something. When minimum-lot-to-maximum-applicants applies, additional lots on the same application generally do not improve the odds of being selected.
7. Submitting several applications on the same PAN. Duplicates on one permanent account number are identified at reconciliation and typically all of them are rejected, not just the extras.

What the table can and cannot support

Subscription data is a good record of one specific thing: how much demand a price band attracted within a fixed window, split by investor type. That is genuinely useful for understanding how an allotment is likely to be crowded, and for seeing which class of investor is actually funding the offer. It says nothing about margins, debt, promoter intent, competitive position or whether the price band is reasonable relative to comparable listed companies. Those questions are answered by the prospectus, the financial statements and peer comparison, not by a multiple on a screen.

“A subscription multiple describes the queue, not the destination.”

Why sources sometimes disagree

Exchange pages, broker applications and aggregator sites all publish subscription figures, and at any given moment they frequently disagree by a few decimal places. The reasons are mundane: different refresh intervals, different handling of reserved portions when computing the total, and some sites dividing by the full issue size including anchor while the exchange divides by the net offer to the public. When two figures conflict, the designated stock exchange page for that specific issue is the reference, and the final position is settled by the registrar after close. If you are recording numbers in your own notes, always store the source and the timestamp beside the value, because without those two fields a subscription figure cannot be meaningfully compared with anything.

There is a further definitional wrinkle in what counts as fully subscribed. An offer that has attracted bids exactly equal to the net public offer reads as one time subscribed, but it must still satisfy the minimum subscription condition prescribed in regulation and restated in the offer document. Where that condition is not met, the issue is withdrawn and blocked funds are released instead of allotted. This is uncommon in heavily bid mainboard issues and considerably less uncommon at the smaller end of the market, where the mainboard and SME segments behave quite differently in both demand pattern and disclosure depth.

A short reading routine

1. Open the offer document and note the reserved quantity for each category, so you know every denominator before you look at any multiple.
2. Read your own category row first, and convert it into approximate odds only if the minimum-lot rule applies.
3. Check the other rows to see who is actually funding the book, and whether the total is being carried by one bucket.
4. Note the time of the reading. Compare like with like across sessions rather than mixing a morning figure with a closing one.
5. Treat the closing number as provisional until the registrar publishes the basis of allotment with verified application counts.
6. Keep the subscription reading separate from your view on the business, and record the two conclusions in different lines of your notes.

Anyone who follows that routine for three or four issues will start to notice something useful: the number that gets quoted in headlines is almost never the number that determines what lands in a retail demat account. The category row, the application count and the allotment rule do that work, and only the first of those is visible while bidding is still open.