Everything that went wrong in our last field note was on the public record before it happened. This note is the reading guide: where each warning sign sits in an IPO filing, what it looks like on the page, and how strong the evidence is that it matters.
One thing up front. None of these signs predicts what a single company will do. They are patterns across thousands of listings, and plenty of companies that carried every one of them went on to do well. What they change is the odds — and what they reward is reading the filing before buying, rather than after.
01 / THE DOCUMENTWhere the filing lives.
A US company going public files a registration statement with the SEC, usually a Form S-1 (F-1 for a foreign company). Revisions follow as S-1/A amendments, and the final prospectus, with the price filled in, is filed as a 424B4. All of it is free on EDGAR, and EDGAR's full-text search covers filings since 2001, so you can search a whole prospectus for a single phrase.
The SEC also publishes its staff's comment letters on a filing, and the company's replies, at least 20 business days after the registration becomes effective. They are filed as UPLOAD and CORRESP, and they are the closest thing to reading the regulator's margin notes.
02 / PROFITSIs it making money yet?
The single best-documented warning sign is also the plainest. In Jay Ritter's data on 9,253 US IPOs from 1980 to 2024, companies that were unprofitable in the twelve months before listing returned 30.7% less than the market over the following three years, measured from the first closing price. Profitable ones trailed the market by 13.0%. Against companies of similar size and valuation, the unprofitable group trailed by 23.4%; the profitable group came out 1.8% ahead.
And unprofitable listings are now the norm rather than the exception. Across 1980–2025, 43% of IPOs had negative earnings; across 2001–2025, 61% did. In 2020 it was 80%, and in 2025, 53%.
Where to look: the summary financial data near the front of the prospectus, and the income statement in the audited financials at the back. The line is net loss per share.
03 / SIZEHow big is the business underneath?
Losses matter most when the company is small. Ritter sorts the same IPOs by sales before listing, in 2025 dollars. Companies with under $10 million in sales returned 52.3% less than the market over three years. Companies with $100 million to $500 million trailed it by 2.6%.
Put the two signs together and the picture sharpens. Unprofitable companies with under $100 million in sales trailed the market by 40.9%. Unprofitable companies with more than $100 million trailed it by 2.7% — almost exactly the same as profitable companies of that size, at 3.4%.
A loss on its own says less than you might think. A loss on a small revenue base is the combination the record punishes.
04 / WHO IS SELLINGNew shares, or old ones changing hands?
An IPO can sell new shares, raising money for the company, or existing shares, raising money for the people who already own them. The prospectus says which in two places: Use of Proceeds, and Principal and Selling Stockholders, which must name each seller, their relationship to the company over the past three years, and what they hold before and after.
The research here is more specific than the folklore. A study of 4,219 IPOs from 1980 to 2001 by Brau, Li and Shi found that selling by existing holders in general was not linked to worse results — but selling by the company's own officers and directors was associated with poorer long-run returns. The people who know the business best choosing to leave at the offering is the version worth noticing.
Where to look: the selling-stockholder table, and whether the names in it also appear in the list of executives and directors. In Use of Proceeds, notice when the money goes mainly to repaying loans — the company must say whose, at what interest rate — or only to "general corporate purposes".
05 / THE PRICE GAPWhat you pay, and what they paid.
Every prospectus for a company without a long earnings record has a section called Dilution. It sets the offer price beside the company's net tangible book value per share after the offering: its assets, minus its liabilities, minus intangibles like goodwill, divided by the shares. The difference is the "immediate dilution" a new buyer absorbs on day one.
The same section usually has a table showing what existing holders paid for their shares in total and per share, next to what the public is paying. We found no study showing that the size of the gap predicts returns, so treat it as context rather than a signal: it tells you how much of the price rests on the future arriving as planned.
06 / GOING CONCERNThe two words auditors do not use lightly.
When an auditor doubts that a company can keep operating for the next year, the audit report must say so in a paragraph containing the words "substantial doubt about its ability to continue as a going concern." Management may say it too, in the risk factors or the notes. Searching a prospectus for "going concern" takes seconds.
It is rarer than you might expect in a listing, and the market does notice it. A University of Connecticut study of about 1,000 IPOs from 2001 to 2013 found that roughly 6% of prospectuses carried management's own going-concern disclosure, and those companies went public at lower prices: Santarus priced at $9 against a planned $12, Stereotaxis at $8 against $15. The warning is priced in at the offering. Whether the company survives is a separate question the research does not settle.
07 / THE BOOKS THEMSELVES"Material weakness."
A material weakness is a flaw in a company's accounting controls serious enough that a significant error could slip through unnoticed. It is common at listing: KPMG counted 54 of 122 US IPOs in 2023 (44%) disclosing one in their registration statements, up from 40% in 2021 and down from 58% in 2022. The most common cause was simply not having enough accounting staff.
Common is not the same as harmless. Audit Analytics found that newly public companies reporting ineffective controls were more likely in later years to file late, write down assets, restate their results, or receive a going-concern opinion. Groupon is the textbook case: five months after its November 2011 IPO, it revised its fourth-quarter results downward — revenue by $14.3 million, net income by $22.6 million — and disclosed a material weakness in how it closed its books.
08 / ADJUSTED NUMBERSWhen the company grades its own homework.
Companies may show their own "adjusted" measures alongside the standard accounting (GAAP) figures, but SEC rules require the GAAP number to appear with equal or greater prominence, and the SEC's staff has said that stripping out normal, recurring cash expenses can be misleading.
That is not an abstract rule. In 2011 the SEC's staff told Groupon that its "Adjusted CSOI", which left out the cost of online marketing — the cost of acquiring its customers — was potentially misleading, and asked it to remove the measure. Groupon did, before listing. In 2019 WeWork's S-1 reported net losses of $0.4 billion, $0.9 billion and $1.9 billion over three years, as revenue grew from $436 million to $1.8 billion, and gave the founder's shares twenty votes each. The IPO was withdrawn. The company listed through a SPAC in 2021 and filed for Chapter 11 in November 2023.
Where to look: the prospectus summary and the "key metrics" in MD&A. Find the GAAP net loss first, then read the reconciliation and ask which costs were left out — and whether they will come back every year.
09 / NOT EVERY FLAGWhat the evidence does not support.
Some features look alarming and are not, on the record.
- Dual-class shares. Founders with super-voting stock raise real governance questions, but in Ritter's data dual-class IPOs from 1980 to 2024 trailed the market by 6.7% over three years — better than single-class IPOs, at 22.0%.
- A few big customers. Any customer above 10% of revenue has to be disclosed. One study of 1,023 IPOs found concentration helped at the offering but weighed on profitability over the following four years. Mixed, not damning.
- "Emerging growth company" status. Smaller issuers may show two years of audited financials instead of three and skip the auditor's check of their controls. That means less information, and studies find more first-day and early price volatility, but no established effect on long-run returns.
- A little-known underwriter. Older research found long-run underperformance was less severe for IPOs led by more prestigious banks. It is a real pattern, but a weak reason on its own.
10 / TEN MINUTESA reading list for any prospectus.
- Summary financial data: revenue and net loss per share.
- Search: "going concern" and "substantial doubt".
- Search: "material weakness".
- Principal and Selling Stockholders: are officers and directors selling?
- Use of Proceeds: new money for the business, or repaying someone?
- Dilution: offer price against tangible book value.
- Search: "Adjusted" — then find the GAAP figure it replaces.
- The lockup expiry date, usually 180 days out.
11 / WHAT THIS ISOdds, not answers.
Every figure above is an average over thousands of companies. Averages are how you set expectations, not how you judge one business, and some of the best-known listings of the last twenty years were unprofitable and small when they went public. What the record does say is that the reading is cheap and the omissions are expensive: the facts that mattered were in the filing, in sections anyone can open, before the price reflected them.
That is the same idea behind the daily screen: not a forecast, but the figures from each company's own filings, put in one place so they are hard to miss.
12 / SOURCESWhere the numbers come from.
IPO returns by profitability and sales, share of unprofitable IPOs, and dual-class returns: Jay R. Ritter, University of Florida, Initial Public Offerings: Updated Long-run Statistics (Tables 16 and 16b, updated February 16, 2026), Underpricing (Table 9, updated December 30, 2025) and Dual Class Structure of IPOs (Table 24). Selling by officers and directors: Brau, Li and Shi, Journal of Banking & Finance (2007). Going-concern disclosures: University of Connecticut (2019), on research by Mark Willenborg and co-authors; PCAOB AS 2415. Material weaknesses: KPMG, IPO Material Weaknesses Study (2025); Audit Analytics (2020). Groupon: SEC correspondence of August 10, 2011 and the company's revised results of March 30, 2012. WeWork: Form S-1 filed August 14, 2019. Customer concentration: Saboo, Kumar and Anand, Journal of Marketing (2017). Emerging growth companies: SEC; Barth, Landsman and Taylor, The Accounting Review (2017). Underwriters: Carter, Dark and Singh, Journal of Finance (1998). Disclosure requirements: Regulation S-K Items 404, 504, 506 and 507; SEC non-GAAP guidance; SEC Investor Bulletin, Investing in an IPO.