Glossary
What is Named Entity Recognition (NER)?
Definition
Named entity recognition is the step where a system reads a piece of text and tags which words are names: people, companies, products, places. Entity optimization is the work you do about the result, while NER is the machine process producing it. Get tagged as the wrong kind of thing and every answer downstream inherits the mistake.
At a glance
- CategoriesTechnical
- Related fieldsTechnical
- DifficultyAdvanced
At a glance
- NER labels spans of text with a type, then resolves each label to a known entity.
- The hard part is not spotting the name, it is deciding which real thing it refers to.
- Brand names that double as ordinary words are the most common failure case.
Tagging the names inside a sentence
Two jobs, run in order. Find the spans that are names and label each one: person, organization, product, place, date. Then link each label to a specific record, so “Apple” in a sentence about phones connects to the company and not the fruit.
The second job is where things break. A tagger works from context. “Parsnipp raised a round” reads as a company. “Parsnipp, fresh from the farm” reads as a root vegetable, and a page that never says which leaves the decision to a guess.
Once resolved, the entity joins everything else known about it, which is how a knowledge graph gets populated. Google’s Knowledge Graph Search API returns the type it settled on for a name, a quick way to see what a tagger thinks you are.
When your brand name is also a vegetable
Pick a name that already means something and you inherit that meaning until you outweigh it.
The symptom is an assistant answering about your company with facts about the other thing, or leaving you out because it could not confirm you exist. No error message arrives either way.
It gets worse in the long tail, where your name appears once inside a sentence about something else and that single mention is all the resolver gets. We hit this often enough at Parsnipp to have written a page on parsnip versus Parsnipp.
Fixable, and slow. You are shifting the balance of evidence until the company reading is the obvious one.
Give a tagger fewer chances to guess
On-page work, no tool required.
- Put the type next to the name. “Saucony, the running shoe brand”, in your first paragraph. Two extra words and resolution is trivial.
- Use the full name on first mention, every page. Not the short internal version. Consistency across pages builds the association.
- Name the things around you. City, investors, integrations, industry body. An entity surrounded by known entities resolves faster than one floating alone, which is the core of entity optimization.
- Add Organization schema with a sameAs list pointing at your LinkedIn, Crunchbase and Wikipedia records. A direct statement of which entity you are.
- Search your own name. A vegetable or a competitor holding those results is what you are working against.
Frequently asked questions
-
Is NER the same as entity optimization?
No, they sit on opposite sides of the process. NER is what a machine does to your text: find the names, label them, resolve them. Entity optimization is what you do to make that job easy.
-
Do I need to do anything about NER directly?
Not as a task of its own, and a vendor selling NER services for marketing is repackaging basic content work. You control the text a tagger reads: say what your company is, keep one name, connect yourself to entities that already exist.
-
What if my brand name is a common word?
Then you carry a permanent tax and plan around it. Pair the name with its type in prose, avoid the bare name in page titles, get third parties describing you the same way. Apple and Square solved this with sheer volume of coverage, slower than it sounds.
-
Does NER still matter now that models read whole sentences?
Yes, less visibly. Modern models handle recognition implicitly rather than as a separate labeled step, so nobody discusses it. The failure mode is unchanged: ambiguous name, thin context, wrong thing retrieved. With an unambiguous name and pages that say what you are, stop thinking about it.