Explain a Metric
Choose the metric, type the figure as it is reported, and add any peer or historical figures you want compared against it.
Why Are There No Good or Bad Thresholds?
Almost every published explanation of a financial ratio ends with a range. A price-to-earnings ratio under 15 is called cheap, a current ratio above 2 is called comfortable, a debt-to-equity ratio above 1 is called aggressive. None of those numbers comes from anywhere. They are conventions that survive because they are memorable, and they break the moment they meet a specific company.
A software business with almost no assets and a utility with a balance sheet full of them cannot share a debt-to-equity threshold. A grocery chain that collects cash at the till and pays suppliers sixty days later runs a current ratio below 1 as a permanent feature of a working model, while a manufacturer at the same reading may be about to run out of room. A price-to-earnings ratio of 40 on a company whose earnings triple over five years turns out to have been a lower multiple than 12 on a company whose earnings halve.
So this tool reports a different category of statement. Every line it produces about a specific value is either arithmetic or a definition. A negative price-to-earnings ratio means earnings per share was negative: that is not a judgement, it is what the sign of the denominator implies. A payout ratio above 100% means dividends paid exceeded net income for the period, so the difference came from somewhere other than that period's profit. A current ratio below 1 means current liabilities exceed current assets. Each of those is true regardless of the industry, the cycle or the reader's assumptions, which is exactly why it can be stated without knowing them.
How Do You Use the Output?
The four sections of a result do different jobs, and reading them in order is what turns a number into an understanding of a company.
- Definition, formula and inputs. Start here even for a metric you think you know. Most disagreements about a ratio turn out to be disagreements about what went into it: trailing or forward earnings, GAAP or adjusted, gross or net debt, whether leases count.
- What follows from this value. These lines are the mechanical consequences of the specific figure. They are most useful at the edges, where a metric quietly stops meaning what the reader assumes: at zero, below zero, above 100%, or below 1.
- What to establish first. These are the questions that decide whether the number is informative at all. They are deliberately questions rather than answers, because the answers live in the filing rather than on this page.
- How it misleads. Every ratio has a characteristic failure. Knowing the failure mode of a metric is usually worth more than knowing its typical range, because the failure is what produces a confidently wrong conclusion.
The comparison box adds a fifth job. Paste a peer median and your own multi-year average alongside the current figure, and the difference and ratio give the number a frame that a universal threshold cannot. A margin at 22% against a five-year average of 25% and a peer median of 16% is a specific situation. The same 22% with no frame is a digit.
Worked Example: A Price-to-Earnings Ratio of 31.7
Enter 31.7 against the price-to-earnings ratio and the tool returns the definition, the formula, and one observation: the reciprocal is the earnings yield, so 3.15% of the current price was earned per share over the period used. That single restatement is often the most useful line, because it puts the multiple on the same footing as any other yield the reader might be weighing.
Then come the questions. Is the denominator trailing twelve months or a forward estimate? Is it GAAP diluted earnings per share or an adjusted figure? Was the period distorted by a one-off item? What growth rate, sustained for how long, would justify paying 31.7 times on the reader's own assumptions? None of those is answerable from the number, and all of them change what it means.
Then the failure modes. Earnings are an accounting output rather than cash. The ratio breaks completely at or below zero earnings and becomes arbitrarily large as earnings approach zero from above. Comparing it across industries compares capital intensity and accounting convention rather than valuation.
Add a comparison line reading Peer median,18.4 and the tool reports a difference of 13.3 and a ratio of 1.72x. It does not say the company is expensive. What it says is that on the figures supplied, this multiple is 1.72 times the peer figure, and the questions above are now the ones that matter.
Which Metrics Does It Cover?
Twenty-six metrics, grouped into five families. The list is deliberately bounded: each entry carries a written definition, a formula, an input note, at least three context questions, at least three limitations and at least one primary source, and a metric that cannot carry all of those does not belong in the registry.
- Valuation: price-to-earnings, forward price-to-earnings, PEG, price-to-book, price-to-sales, EV/EBITDA, earnings yield, free cash flow yield.
- Profitability and returns: gross margin, operating margin, net profit margin, return on equity, return on assets, return on invested capital.
- Balance sheet and liquidity: debt-to-equity, net debt to EBITDA, interest coverage, current ratio, quick ratio.
- Growth and distributions: revenue growth, earnings per share growth, share count change, dividend yield, dividend payout ratio.
- Fund metrics: expense ratio, tracking difference.
Every metric also names the ones that should be read next to it. A return on equity figure points at return on assets and return on invested capital, because the gap between them is where leverage shows up. An earnings per share growth rate points at share count change, because buybacks move the first without moving earnings at all.
What This Tool Does Not Do
- It does not look anything up. There is no ticker field and no market data. You supply the figure; the tool explains it.
- It does not check your number. If the value came from a source that computed it differently, the explanation still describes the standard construction rather than that source's.
- It does not value anything. Nothing here produces a target price, a fair multiple or a conclusion about a security.
- It does not know the industry. Where the industry decides the reading, the tool says so and asks the question rather than guessing at an answer.
- It does not store anything. Nothing typed into the form leaves the browser or survives a reload.
Frequently Asked Questions
What does Explain This Number do?
It takes one financial metric and one reported value and returns the definition, the formula, where each input comes from in a filing, the statements that follow definitionally from that specific value, the questions to answer before reading anything into it, the ways the metric misleads, the related metrics worth reading alongside it, and primary sources. It covers 26 metrics across valuation, profitability, balance sheet, growth and fund families.
Why does it not tell me whether a number is good?
Because no defensible universal threshold exists for any of these metrics. A debt-to-equity ratio that is unremarkable for a regulated utility would be extreme for a software company; a current ratio below 1 is a permanent feature of businesses that collect cash before paying suppliers. Published ranges are conventions rather than findings. Instead of inventing one, the tool reports statements that are true regardless of industry or cycle, and names the questions whose answers would settle the judgement.
What is a definitional observation?
A statement that follows necessarily from the value itself rather than from any assumption about the company. "A payout ratio above 100% means dividends paid exceeded net income for the period" is definitional: it is what the arithmetic of the ratio implies, and it holds for every company in every industry. "A payout ratio above 100% is unsustainable" is not definitional, because sustainability depends on cash flow, the balance sheet and the company's intentions, none of which the ratio contains.
How should I enter a percentage?
As the percentage, not the fraction: 12.5 for 12.5%, never 0.125. Every metric declares its own unit and the form shows a hint for the one selected. Values outside a metric's definitional bounds are rejected with an explanation rather than accepted: a gross margin of 250 is refused, because gross margin cannot exceed 100% when cost of revenue is positive, and the most likely cause is a fraction entered where a percentage was expected.
Can I compare a figure against peers?
Yes. Enter one line per comparison as a label and a value, for example a label of "Peer median" followed by a comma and 18.4. The tool reports the arithmetic difference and the ratio between your figure and each comparison. It does not say which is preferable, because for several of these metrics the answer genuinely depends on the business: a lower price-to-earnings ratio can reflect a cheaper valuation or a deteriorating one, and the ratio alone does not distinguish them.
Why does the ratio show as not defined for one of my comparisons?
Because the comparison value was zero, and dividing by zero has no result. The tool reports the difference for that row and marks the ratio as not defined, rather than showing an infinity or an arbitrarily large figure that would look like a real reading. The same principle applies throughout: where a calculation has no answer, the output says so.
Does it use live market data?
No. Nothing is fetched, no ticker is resolved, and no price or filing is retrieved. The tool explains the number you supply. That is a deliberate boundary rather than a missing feature: an explanation of what a metric measures and how it misleads does not depend on which company reported it, and pretending otherwise would require licensed data the explanation does not need.
Where do the primary sources come from?
Each metric names the SEC or Investor.gov documents that define its inputs, such as the Form 10-K instructions for the statements a figure is drawn from, or the SEC's non-GAAP interpretations where a metric is commonly reported on an adjusted basis. Those links are the authority for what a line item is, not for what a value means, which is a distinction the tool keeps deliberately visible.
Is my input stored or sent anywhere?
No. The calculation runs entirely in the browser, nothing is written to storage, and nothing is transmitted. Reloading the page clears the form. The same function that produces the on-screen result is what an AI browser agent receives if it calls the tool through the page, so there is no second code path and no second answer.
Why does one metric show no observations for my value?
Because nothing definitional follows from that particular reading. Return on assets at 7% carries no consequence that is true independent of the industry, so the tool prints nothing rather than filling the space with a band. The definition, formula, context questions and limitations are still returned in full, and those are where the useful content sits for a value in an unremarkable range.
Reading a Metric Instead of Scoring It
The habit this tool is built to replace is the one where a reader sees a ratio, recalls a threshold, and reaches a conclusion in a second and a half. That habit is fast and it is frequently wrong, because the threshold was never derived from anything and the conclusion skips every question that would have made the number informative.
The slower habit is not much slower. Establish what went into the ratio, since most disputes about a figure are disputes about its construction. Note what the value implies mechanically, which matters most at the edges where a metric stops behaving: a multiple over negative earnings, a payout over 100%, a coverage ratio under 1. Frame it against the same company's own history and against a peer set you can defend, which the comparison box does arithmetically. Then read the failure mode, because knowing how a metric lies is worth more than knowing its typical range.
Done in that order, a single number turns into a short list of things to check in the filing. That is the intended output. The tool is not a substitute for reading the 10-K, and the metrics it explains are not a substitute for understanding what the business sells and to whom. It is a way of making sure the number you are about to act on means what you think it means, which is the failure that quietly precedes most of the others.
The same registry backs the tool call an AI browser agent makes on this page, so an agent summarising a company's metrics gets the identical definitions, caveats and refusal to grade. That was the point of building the explanation as data rather than as prose: one source of truth, whether the reader is a person, a page or a model.