Business terms let you define a piece of vocabulary once — "Email Address," "Customer ID," "Order Amount" — and have it applied automatically across your data. Ins tead of writing quality checks field by field, you define detection and evaluation logic on the term itself, and every field that matches inherits both the tag and the checks.
Overview
The Business Ontology is where you manage your organization's business terms. Open it from Data Quality › Business Ontology. The manager lists every term with its Term name, Description, and Total Rules, so you can see at a glance how much detection and evaluation logic is attached. Click New Term to open the editor and create one.
Each term bundles two independent pieces: rules that recognize the term in your data during discovery, and rules that check the quality of data once it's tagged. You set these up separately, but they work together — recognition drives enforcement.
How does a term auto-tag data?
Open a term and go to its Detection on Attributes section. This is where you set rules to automatically identify and tag attributes that match the term during data discovery. Click Add a Detection Rule to search for and attach an existing rule, or use Create New Detection Rule to build one inline without leaving the term editor.
Each attached rule shows its Data Type and a Threshold % — the minimum percentage of values in a field that must match for the term to be applied. For example, an 80% threshold means at least 8 out of 10 sampled values need to match before the field gets tagged. If you leave the threshold blank, a default of 50% is used.
During profiling, fields are sampled against rules whose data type matches the field. When the share of matching values clears the threshold, the field is automatically tagged with the term — no manual work required.
How does a term enforce data quality?
Below detection, the Data Quality Evaluation section defines the quality checks that automatically run on every attribute tagged with the term. Attach rules with Add an Evaluation Rule or build one with Create New Evaluation Rule; each rule you attach shows its Dimension so you can see what aspect of quality it measures.
This is where the ontology pays off. Define what a valid email address looks like once, on the "Email Address" term, and every field the system recognizes as an email inherits that check automatically — instead of you configuring the same rule separately on every table that happens to store one.
Where do you see and fix what got tagged?
To review how terms landed on your data, open a Monitoring Suite and go to an object's Fields tab. Each field shows the terms currently linked to it. If detection missed a field or tagged the wrong one, use Edit Terms Linked to attach or remove terms manually — manually linked terms behave exactly like automatically detected ones, including running the same evaluation rules.
Notes
Metadata-based detection rules aren't applied yet — only rules based on the underlying data currently tag fields.
A detection rule tied to more than one attribute, or to none, is skipped, since detection evaluates one field at a time.
A rule built for one data type is never tried against a field of a different type — for example, a string-based rule won't be tested against numeric fields.
If a term isn't tagging data you expect it to, try lowering its Threshold %, especially for noisy source data.
Hover a rule in the term editor to open Rule Logic Details and inspect exactly what it checks before troubleshooting further.