Guide · The ontology

The ontology

In a nutshell

The ontology is the lab's vocabulary for earnings: the kinds of thing a release can state, such as a reported value, guidance or a risk, and how they connect. Every node on the graph is typed by one of its classes, which is why a node's panel names its class.

What it is

An ontology is a shared, written-down vocabulary. This one says what an earnings release can contain and how the pieces relate: a company reports values for metrics and periods, gives guidance, names drivers and risks, announces corporate events and quotes its executives. Each kind of thing is a class, and each class has a short name with the prefix ern:, such as ern:ReportedValue.

The lab extracts facts from filings by asking a language model to fill in this vocabulary and nothing else, so a fact either fits a class or is not recorded.

The classes you meet most

Class What it records Article
ern:ReportedValue A number the company reported Reported value
ern:Guidance The company's own forecast Guidance
ern:Driver, ern:Risk Why a metric moved, and what could hold it back Drivers and risks
ern:CorporateEvent A buyback, dividend, acquisition and the like Corporate events
ern:Quote Words an executive said, verbatim Management quotes
ern:AnalystQuestion A question asked on an earnings call Analyst questions
ern:Passage The piece of a filing a fact came from Passages
ern:ExtractionRun The run that produced the facts Extraction runs

Words resolve to meaning

Companies say the same thing in different words: sales, net sales, top line. The ontology's metric concepts carry those wordings as synonyms and, where one exists, the matching XBRL concept the SEC uses for the figure. That is how a question about "sales" finds facts recorded as revenue.

On the page

A node's panel has a section called Class in the ontology. It names the class, gives its definition, and links to the Ontology tab, which lays out every class and how often the run used it.

All articles