Overview

BrandTrackers Docs explains what we track, how those things connect, and what the system must do to support research. It is written for people and machines from the same source.

What this is

BrandTrackers is a library of brand and creative work. People use it to study companies, products, campaigns, complete pieces and useful moments. The same work can be reached through several meaningful relationships without making another copy of it.

These Docs are the shared explanation of that model. They connect approved definitions to the vocabulary, database, pipelines, prompts and research experience. Each kind of evidence has a clear home: an approved meaning, a dated system observation, a proposed choice and a demonstrated capability are different claims.

A definition explains intended behavior. It does not establish that storage, classification or the interface already implements it. The program audit records that comparison and the remaining work.

The connected model

The model has several connected subjects. Their pages provide useful reading paths; they do not require one exclusive folder tree.

• Entities explain the organizations and people we follow, their identities and their roles.
• Products explain product identity, families, bundles and named features. Products have their own explanation without becoming a competing identity system.
• Groupings explain bodies of work such as campaigns, identity programs and projects, and how they connect with events.
• Content explains complete pieces, creative versions, publications, captures and the relationships between an original asset and documentation of its use.
• Moments explain useful beats inside videos, with evidence tied to the part that supports a description.
• Concepts explain the shared vocabulary, names, types, independent descriptions and overlapping browsing entrances.

These arrows name relationships, not automatic inheritance. A company can serve an enterprise audience without every piece linked to it targeting that audience. A campaign can gather a video and a photograph without making them the same work. A short logo animation can make its parent video discoverable without making the whole video a Logo Animation.

Four surfaces, and why they are separate

Surface	What it explains	How to read it
Model	Approved meanings and relationships.	The intended behavior, expressed in plain language.
Decisions	Approved answers and remaining choices.	The source and scope of each decision.
Vocabulary and technical reference	Definitions, stored terms, schema and observations.	Check whether each entry is approved, proposed, legacy or measured.
Working history	Plans, research and earlier evidence.	Dated context; current execution is selected by the maintained plan.

Pipelines explain how information is collected and processed. Their implementation claims need dated evidence, just as database claims do. A local source file, deployed code, a historical run and a fresh verification establish different things.

State of the data and conformance are recorded measurements and checks. They are not continuous proof of the current system. Check their observation date, tested rule and coverage before relying on a result. Open questions holds unsettled meanings; model pages explain approved rules.

How to use this

Start with Concepts for the shared rules, then follow the relevant subjects above. Use the vocabulary reference to inspect individual terms and their review status. Use the program audit to see where approved meaning, current explanation and implementation still differ.

For execution order and completion gates, use the plan. For requirements and acceptance, use the product requirements. Historical reviews preserve the reasoning; they do not ask you to approve settled answers again.

Last reviewed 6 October 2026.