Glossary
2026-08-09
This page explains the terms used on the QuorumLab Finland site.
It does not replace the Method page and should not turn into a technical encyclopedia. Its job is to help a visitor pick up the language of the site quickly.
- The machine layer
- What search engines and AI assistants build an answer about a company from: the site and its markup, records in open databases, external identifiers, public sources and the links between them. An assistant answer almost never comes from one place. The error usually appears in this layer rather than in the model itself.
- AI Visibility Baseline
- The initial measurement of what AI assistants and search engines say about a company on a given date. A baseline exists so that the state before and after changes can be compared.
- Baseline date
- The date the first measurement is fixed to. Every later result is compared against it. Without a date there is no honest way to say whether things got better or worse: assistant answers drift on their own.
- Measurement
- A repeatable check of assistant answers against a fixed set of questions. One answer does not count as a measurement. At QuorumLab each question runs at least five times across different days and in separate sessions.
- Run
- A single request to a system using a question set in advance. Five runs exist to show whether a result holds, rather than to catch one answer by chance.
- Second measurement
- A check on the same method after some time has passed since the changes. It shows what changed against the baseline date.
- Red zone
- A state where the correct answer appears in two cases out of five or fewer. A red zone means the error repeats often enough to need an examination of the sources.
- Yellow zone
- A state where the correct answer appears in three cases out of five. Not a disaster, but not a stable result either: one user gets the right answer and another does not.
- Green zone
- A state where the correct answer appears in four or five cases out of five. It means the system recognises the company and answers consistently.
- Structured data
- Machine readable markup on a website. It helps search engines and assistants understand what is actually on a page: an organisation, a product, a person, an article, a service, an address, a profile or a link to an external record.
- Wikidata
- An open knowledge base that search engines, assistants and other outside systems can read. Records in Wikidata are edited by a community. Any edit can be revisited by other contributors.
- External identifiers
- The IDs and profiles that connect a company, an organisation or a person to outside sources. They can be academic, corporate, registry, media or platform identifiers. QuorumLab only connects identifiers that actually exist.
- Owned surfaces
- Open surfaces controlled by the client or by structures connected to the client. This can be a product site, a professional site for an executive, a project page, a reference resource or a research page.
- Reddit Exposure Baseline
- A measurement of whether Reddit discussions appear in assistant answers and for which queries. It shows whether Reddit threads are being cited, whether there are errors in them and whether an official presence is worth building.
- Wikipedia Eligibility Report
- An assessment of whether the independent sources support a possible article in the English Wikipedia. It is not a promise of publication and not the writing of an article, but an analytical verdict on the odds and the risks.
- Google Knowledge Graph
- The external Google knowledge graph, which can feed panels, links and answers. QuorumLab does not edit the Google Knowledge Graph directly. We work on the sources and links that outside systems can build data from.