How the scores are calculated

01 · Signed yields, trailing fundamentals

Earnings yield uses trailing EPS divided by price, EBITDA yield uses EBITDA divided by enterprise value, and book yield uses book value divided by price. A signed inverse multiple is the fallback; zero multiples are missing. Negative earnings are shown as reported. Free cash flow adds the latest four consecutive quarterly cash flows and divides by market capitalization. We never annualize an incomplete set of quarters or substitute an unverified annual value.

When a company reports in a different currency from the one its shares trade in (an ADR, for example), its reported EBITDA, cash flow, debt and cash are converted at the current exchange rate before any yield or enterprise value is calculated; without a rate, those yields are left out. Debt/equity comes from SEC filings where they report it: debt from the latest balance sheet that shows long-term debt, equity from the same date, lease obligations excluded. Missing debt is unknown, never zero; only then is Yahoo’s figure used (a percentage, divided by 100). Negative debt/equity is treated as missing because it can indicate negative equity.

02 · Compare like with like

Within each sector, each metric is capped at the 5th and 95th percentiles. Its robust Z-score is (capped value − sector median) / median absolute deviation (MAD). A missing value is left out, never filled with the sector median: the category’s remaining weights are rescaled over the inputs that are reported. A category needs at least two of its weighted inputs (all of them when it has fewer), and each report card shows how many were available: 3/3 is high confidence, 2/3 is partial data.

Valuation combines 50% EBITDA yield, 30% earnings yield, and 20% free cash flow yield; book yield is shown for context. Growth gives equal weight to revenue growth over the last year, earnings-per-share growth over the last year, and annual revenue growth over three years. Profitability gives equal weight to return on invested capital (after-tax operating income over equity plus debt minus cash), operating margin and return on assets; return on equity and gross margin are shown for context. For US companies, growth and profitability come from SEC filings (trailing twelve months, per-share figures adjusted for splits). Elsewhere they come from Yahoo Finance’s annual income statements: the latest complete fiscal year against the one before, and operating margin for that fiscal year. Both bases cover twelve months, so companies are compared over the same span; a single quarter is never compared with a year. Each figure on a report names its period.

Some inputs have a special state instead of a number. Negative shareholders’ equity counts as the highest debt/equity in the peer group and is flagged, rather than disappearing. When cash exceeds market value plus debt (enterprise value zero or below), EBITDA yield is not meaningful and is left out. A negative earnings, EBITDA or free-cash-flow yield (a loss or cash outflow) counts as zero, whatever the share price: otherwise a rising price would shrink the loss per dollar invested and make a loss-maker look cheaper. The size of the loss is judged under profitability (since 3.1.2). When earnings per share change sign, a percentage is meaningless, so the change is a state: a turnaround from loss to profit is placed at the peer group’s 90th percentile of EPS growth, a narrowing loss at the 60th, a widening loss at the 20th and a slide from profit to loss at the 10th.

03 · Turn the comparison into a score

Each company’s weighted result is then ranked against its sector from 0 to 100. Companies with identical results share a rank, and a sector with a single company, or where every company ties, gets a neutral 50. Category scores are whole numbers.

The risk calculation gives equal weight to lower debt/equity, lower beta, and higher current ratio. A beta at or below 0.1 is treated as missing: it usually reflects thin or unusual trading data rather than a share price that ignores the market. A safety percentile of at least 67 is lower relative risk, 33–66.99 moderate and below 33 higher. Each level holds about a third of a peer group by construction, so “lower relative risk” means safer than most peers, not safe. The StockQuotient Score averages valuation, growth, profitability and the underlying 0–100 safety percentile. It needs at least three categories; with exactly three it is labelled a partial overall score, and with fewer there is no overall score. Ranks and leaderboards count fully scored companies only; a partial score shows where it would fall among them.

The StockQuotient Score is a summary, not a signal: it describes a company’s fundamentals today relative to its peers and is not a forecast of its share price.

04 · Banks and insurers are scored differently

For banks, capital-markets firms (brokers and investment banks), insurers, mortgage and thrift lenders, and financial conglomerates, free cash flow and EBITDA mostly reflect trading assets, customer balances and funding flows rather than the health of the business. Morgan Stanley’s free cash flow yield, for example, swung from +18% to −24% within two years without a comparable change in the business. Gross margin and current ratio are equally uninformative for these balance sheets.

Since methodology 2.1 these companies form their own peer group, “Banks & Insurance”, and are ranked only against each other. Valuation combines 60% earnings yield (P/E) and 40% book yield (P/B); growth is earnings-per-share growth; profitability is return on equity and return on assets. Since 3.1 they get no risk score: leverage is their business model, and debt/equity and beta say little about a bank’s real risk (capital ratios, loan quality, funding). Their overall score uses the three categories they have. The other metrics stay visible for context with zero weight. Payment networks, asset managers, exchanges and insurance brokers keep the standard method, because cash flow is meaningful for them. Tangible book value, return on tangible equity and regulatory capital would be better still, but they are not in our data source.

05 · Know the limits

Risk is not a bankruptcy probability. Peer groups are global, so accounting standards and markets mix; sector labels are current, not historical. Industry-specific accounting, small peer groups and reporting lags affect comparability. Each report card shows the raw values, weights, Z-scores and input coverage; see Data sources for where the figures come from and how often they refresh.

Industry valuation context

The valuation score compares a company with its peers today. A company can look cheap next to peers that are all expensive, so US stock pages also show where the company’s industry sits against its own history. This never changes any score and does not predict prices.

How it is measured. Each month since December 2015 we take every US company worth $1B or more in the industry and the median of each valuation yield that fits it: earnings, EBITDA, sales and free cash flow for most industries; earnings and book value for banks and insurers; EBITDA, sales and book value for real estate; sales and book value for biotechnology. The page says where the latest month sits in that history (“higher than in 90% of months”). Medians keep loss-makers and the largest companies from dominating. Figures come from SEC filings as they were public at the time and month-end prices, updated monthly.

Breadth is the share of the industry’s companies whose price-to-sales (banks and insurers: price-to-book) is in the most expensive fifth of their own history, using at least three years of it. Price vs earnings compares the median change in company value with the median change in earnings over 24 months, among companies profitable at both ends. Market premium compares the industry’s median price-to-sales with the whole market’s, now and usually.

Which group. An industry is used when it has at least 15 companies in the latest month and in at least 60 months; otherwise the company’s sector is shown, and the page says so. When 40% or more of a group is losing money in a month, earnings-based measures are left out for that month, because collapsing earnings would make a depressed industry look expensive.

Limits. The history covers one decade, so “expensive” means versus 2016–2026 only. It uses today’s companies and industry labels throughout, so companies that failed or were acquired are missing (survivorship bias). In cyclical industries such as energy and materials, trailing earnings and sales fall after a downturn, which can make valuations look high near a cycle’s low. International companies are not covered.

Data with a visible source

US companies: fundamental inputs (earnings, EBITDA, cash flow, book value, debt, margins, returns, growth and the financial statements) are computed from the company’s own SEC filings, which are public domain. Where a filing doesn’t report an item in standard form (some companies tag debt with their own labels), Yahoo’s figure is used for that input only. Share counts of foreign issuers are matched to their US-listed ADRs. Debt excludes lease obligations. Share prices, and beta derived from them, come from Yahoo Finance.

International companies are shown as a beta: their data comes from Yahoo Finance snapshots and is labelled on every page. Each company is refreshed about once a week; each report shows its price date, and every score input lists its source and period under “Behind the score”. Watchlists are stored in your browser.