> ## Documentation Index > Fetch the complete documentation index at: https://docs.shareofmodel.ai/llms.txt > Use this file to discover all available pages before exploring further. # Brand Catalog > How brands and competitors are modelled, deduplicated, and linked to analyses. A **Brand** in Share Of Model is a first-class entity, not a free-text label. Every analysis you create — Brand or Search — is anchored on entries from the Brand Catalog, which is what lets the platform compare apples to apples across modules and over time. ## What a brand is A catalogued brand has: * a **canonical name** (the form the platform uses internally), * a **domain** (the brand's primary website), * optional **visual identity** assets — colors, logos, banner — used in the UI. Brand names are **case-insensitive and unique**. "AWS" and "aws" are the same brand. This guarantees that when you reference a brand from one analysis to another, both point at the same entity. ## Variations and mentions Brands are not always referenced by their canonical name. LLMs answer in natural language and mention the same brand as: * the canonical name (e.g. *"Amazon Web Services"*), * a common acronym or alias (e.g. *"AWS"*), * a parent company or product line variation (e.g. *"Amazon Cloud"*). The platform groups these mentions back to the canonical brand whenever a section of the product surfaces "any brand the model named". You see the canonical brand, but the underlying mention can come from any known variation. Variation grouping applies to **discovery** sections like Brand Awareness, where the platform extracts brand names from free-form LLM answers. In sections that target a specific brand by name (Brand Perception, Category Perception, Key Attributes), the brand under analysis is the only one being asked about, so variation handling is not relevant there. ## Competitors Competitors are brands too. When you configure a Brand analysis with three competitors, you are picking three other entries from the catalogue. The same brand can therefore appear: * as the **subject** of one analysis, * as a **competitor** in another analysis, * as a **mention** inside a third analysis run by a different team. Use this consistently — naming a competitor "Nikee" instead of "Nike" creates a separate, isolated entry that won't roll up to the real Nike's data. ## Why the catalogue matters A consistent catalogue is what makes cross-analysis comparison possible. It powers: * **Awareness benchmarks** — share-of-voice for the same set of brands across categories. * **Competitor tracking** — your brand's perception trend versus a stable competitor list. * **Source attribution** — which Reddit threads or news outlets fed an LLM's view of a known brand. * **API consistency** — the same `brand_id` returns matching entities across endpoints. ## When to add a new brand Add a new entry when: * you want to track a competitor not already catalogued, * you start covering a new category and need its anchor brands, * you onboard a new client whose brand isn't yet in the catalogue. Avoid creating duplicates — search the catalogue first. ## What's next The data unit that consumes brands. Use the catalogue to run an analysis. Common questions on brand handling.