Frequently asked questions

A yield just flips a valuation multiple upside-down and expresses it as a percentage of the price:

• Earnings yield (E/P) = net income ÷ market cap — the inverse of P/E.
• Sales yield (S/P) = revenue ÷ market cap — the inverse of P/S.
• Revenue/EV and EBITDA/EV yields — the inverses of EV/Revenue and EV/EBITDA.

The rule is simple: higher yield = cheaper. A 5% earnings yield is the same statement as a 20× P/E, just written so it behaves well.

P/E puts earnings in the denominator. As a company’s profit shrinks toward zero, its P/E shoots off toward infinity, and the moment earnings turn negative the ratio flips to a meaningless negative number. So P/E is non-linear and discontinuous exactly where the most interesting IPOs live — near break-even.

That breaks comparison. Is a 15× P/E “half as expensive” as a 30×? Not linearly — the gap between 15× and 30× is not the same as between 100× and 115×. Averaging P/E ratios, fitting a trendline through them, or bucketing them all inherit that distortion, and a single near-break-even name can dominate the result.

Yields fix this because the denominator is price (or enterprise value), which is always positive and never near zero. Earnings yield moves smoothly from positive, through zero, into negative, so the whole cross-section sits on a straight, linear axis you can average, plot, and rank on.

Take two IPOs — one profitable, one loss-making — and read each first as a multiple, then as a yield:

“Higher = cheaper” holds cleanly among profitable names. Once an earnings yield goes negative, it stops being a valuation statement and becomes a profitability one: a −30% earnings yield means the company is losing a lot relative to its price — not that it is “expensive.”

So read a negative earnings yield as how deep the losses run, and for pre-profit names lean on the sales yield and the growth-adjusted (GARP) yield rather than earnings alone.

No — a yield is exactly reversible. P/E = 100 ÷ earnings-yield %, so a 5% earnings yield is a 20× P/E and a 2% yield is 50×. The underlying number is identical; we just show the form that adds, averages, and ranks without fooling you.

Newly public companies are disproportionately young, fast-growing, and not yet profitable — and a meaningful share are effectively pre-revenue (biotech, deep-tech, early platforms). On a P/E basis, most of them are simply undefined: their P/E is blank, so they fall out of every P/E average, bucket and scatter. You end up drawing conclusions from only the minority of IPOs that happen to be profitable.

With earnings yield, an unprofitable issuer keeps a real, negative value and stays in the analysis. A pre-revenue company still has a defined sales yield near zero, which is itself informative. Nothing gets silently dropped, so the picture reflects the whole IPO market — not just the profitable slice.

The traditional growth-adjusted multiple is PEG = P/E ÷ growth, which is doubly fragile: it needs a positive P/E and a positive growth rate, and it divides one shaky number by another. Our linear analogue simply adds the two things you care about:

GARP yield = earnings yield % + revenue-growth % (and, on the sales side, S/P yield % + growth %).

A higher number means more combined cheapness-plus-growth. Because it is a sum rather than a ratio, it is defined even for loss-makers and hyper-growth names where PEG is blank — which is most of the IPO universe.

For a genuinely pre-revenue company — a clinical-stage biotech, say — even sales yield is essentially zero, so no valuation yield is meaningful. That is a feature, not a bug: rather than invent a multiple, we lean on pipeline, cash runway and burn rate. The yields are honest about when they do not apply.

Compare within a sector, not across it. A “cheap” earnings yield for a bank is not the same as for a biotech — different growth, risk and capital intensity. That is why the analytics benchmark each IPO against same-sector, same-vintage peers rather than one global number.

Revenue/EV and EBITDA/EV use enterprise value (market cap + debt − cash) instead of market cap, so two companies with very different debt loads still compare fairly. EBITDA/EV is the capital-structure-neutral cousin of earnings yield.

On the Factor Performance analytics tab, the valuation factors and their scatter plots are all yields (Earnings Yield, Sales Yield, Revenue/EV, EBITDA/EV, and the GARP yields). On each IPO’s detail page, the peer-comparison table and the target-price valuation rows are shown as yields too, so the issuer and its comparables line up on the same linear axis.

The underlying data is unchanged — a 5% earnings yield and a 20× P/E are the same fact. We just show the form that adds, averages, and ranks without fooling you.

On each IPO’s Valuation section we pick one primary method, not an average of two models. The pipeline is:

1. Structural gates — hard financial facts (revenue scale, CapEx intensity, asset turnover, financial-institution markers).
2. Phase 0 — a short business read: maturity, economic focus, archetype, and a concrete framework name (e.g. AFFO yield + EV/MW, P/TBV, rNPV).
3. Method router — chooses a default stack or promotes that named framework as a custom method on a cash or opportunity engine family.

Default stacks are: Cash-flow and multiples, Business opportunity scenarios, and Holding company (sum of parts). When Phase 0 names a sector-specific recipe, the UI shows that custom name as primary and the engine family underneath. A dynamic input resolver then fetches the method’s named inputs — including forward-looking disclosures like expansion capacity or contracted backlog — from the prospectus and the web. Nothing is invented: if a core input for the method genuinely can’t be found, we say no valuation possible instead of publishing a made-up number.

The report memo doesn’t just state a number — it shows the math applied, with the real figures substituted in. The valuation legs (the business as it operates today, plus the growth the filing quantifies) produce a low / mid / high scenario range, and the memo ends with one final chain: range mid × market-adjustment multiplier = the target price.

Every figure is then audited in code: each number in the memo must trace back to an approved filing or market fact, and each written formula is recomputed to verify the arithmetic. A memo that fails the audit is repaired or rejected — it never ships with unverifiable numbers.

Because the target is an independent estimate of value per share — it is never pulled toward the offer price. An offer price is set by the issuer and its underwriters to sell a deal; especially for small, speculative IPOs it can sit far above (or occasionally below) any defensible value.

Earlier research anchored targets near the offer with capped nudges, which made every deal look roughly fairly priced. Current research values the company bottom-up from its own economics and peer evidence, so a target well below the offer is the model being honest, not broken. The offer price still appears in reports — as deal context, never as an input to the target.

Yes. A company that discloses quantified expansion — potential capacity, under-construction units, contracted backlog, guided growth — is two assets: the business it runs and the business it is building. The report values both. The growth leg is priced as units × (stabilized value per unit − cost to build per unit), phased, probability-weighted, and discounted to present, and the applied formula appears in the memo like any other.

Uncertainty lives inside the model, not outside it: a thin pipeline prices near zero through low probability and long phasing rather than being ignored. And the share count is scenario-dependent — dilution from options, incentive plans, or a downside financing shows up in the denominator of the very scenario that causes it.

Peers come out of ecosystem research on the filing and the web, then get resolved against the SEC registrant map — only full Nasdaq/NYSE listings qualify (no OTC or pink-sheet names), and each must pass an active-quote and minimum-size check.

Their multiples are computed by us from primary sources: SEC XBRL filings for trailing-twelve-month fundamentals and the latest market close for price — with the convention stated in the data (EBITDA = operating income + D&A; EV = market cap + debt − cash). Vendor ratio feeds disagree with each other precisely because their formulas are hidden; computing our own makes every peer number auditable back to a filing.

Because we would rather tell you nothing than tell you a made-up number. The pipeline resolves the selected method’s inputs from the filing, the web, and derived data — but when a core input genuinely isn’t knowable, the report states that plainly and publishes no target. There is no fallback model and no offer-anchored guess behind the curtain; the rest of the report (business, market, risks) still stands.

Cash-flow and multiples anchors on owner-economic free cash flow, cash per share, runway, and related multiples when the filing supports them.

Business opportunity scenarios builds an explicit downside / base / upside table with probabilities — used when the investment case is stage-gated optionality (contracts, commissioning, clinical or infrastructure milestones) rather than steady free cash flow today.

These are engines behind the primary method, not two competing headline products. Soft cash/opportunity weights appear only on a rare hybrid path when the router explicitly allows it.

The Business stage row is a compact Phase 0 label — usually maturity plus one primary economic focus — not a full taxonomy dump.

Maturity: pre-revenue, early-growth, growth, mature, pe-exit, carve-out.

Economic focus (one primary): cash-flow, unit-economic, binary-option, float-spread, capital-returner, reflexive-growth.

Separately, the method block may show a financial gate (e.g. venture infrastructure, mature business, financial institution) and an archetype (Venture Infrastructure, Asset Light Growth, Mature Conglomerate, Commodity Resources, Financial, Holding Hybrid). Those explain why a stack was chosen; they are not extra competing price targets.

IPOSignal is an IPO research and analytics platform — an IPO calendar with pricing, lock-up and quiet-period dates, per-company research, and factor analytics on what has historically driven post-IPO returns. More on the approach is on the about page.

The research is built from public SEC filings and market data, with sources shown on each IPO page. The analytics are computed from that same underlying data across the whole IPO universe, so a factor you read about is measured the same way for every company.

No. IPOSignal is for information and education only — not investment advice, and not a recommendation to buy or sell any security. Always do your own research and consider your own circumstances.

The IPO calendar and much of the research are free. Deeper research sections and some analytics are part of IPOSignal Premium — see pricing for what each plan unlocks.