# The 3x3 Content Matrix: Nine Types of Content for Brands

> A brand's content divides into three voices across three formats, creating a 9-cell matrix used to configure AI content generation agents.

Source: https://www.starboxai.ai/en/insights/ai-marketing/self-other-customer-voice/ (Starbox AI Insights, updated 2026-10-17)

## The Three Voices

A brand requires three distinct perspectives to speak on its behalf:

| Voice | Speaker | Function |
|---|---|---|
| Brand Voice | The brand, founder, or internal subject-matter expert | Introduces the brand and explains its core methodology |
| Third-Party Voice | Influencers, creators, and external industry experts | Provides objective external judgment and establishes trust |
| Customer Voice | Everyday users and verified clients | Delivers peer reviews and case studies that validate capability |

Brand voice also splits into two operating tracks: **strongly tied** brand persona content, and **loosely tied** educational expert content (which provides evaluation criteria without mentioning the brand name). A single brand can operate both tracks simultaneously.

## The Three Formats

Each voice can be deployed across three distribution formats:

- **Short Graphic Posts**: Visual cards and image-text posts on platforms like Xiaohongshu (RED), WeChat Moments, and Douyin (TikTok's Chinese counterpart).
- **Video Scripts**: Direct-to-camera founder talks, site visits, and client video testimonials.
- **Long-Form Articles**: In-depth analysis on WeChat Official Accounts, Zhihu (China's largest Q&A platform), and digital press releases.

## The 9-Cell Matrix

| | Short Graphic Posts | Video Scripts | Long-Form Articles |
|---|---|---|---|
| **Brand Voice** | Product highlights, social feed updates | Founder explainers, industry primers | Brand manifestos, in-depth technical guides |
| **Third-Party Voice** | Influencer recommendations | Creator product reviews | Expert site-visit features |
| **Customer Voice** | User notes and peer reviews | Customer video testimonials | In-depth customer case studies |

## Applying the Matrix to Generative Engine Optimization

When executing Generative Engine Optimization (GEO), the roles of these three voices become sharply defined:

- **Brand Voice** is published through owned brand channels. Keep the main text informative and place author credentials clearly in the metadata so AI engines can index exactly who the brand is and what it does.
- **Third-Party Voice** should be published via independent third-party channels. The copy should naturally integrate the brand name alongside a single verifiable brand fact. When generative AI models recognize multiple independent sources referencing the same entity, they are far more likely to cite that brand in synthesized answers.

## The Benchmark for Every Asset

The fundamental rule of the New Media Content Factory is **reader utility**: if you strip out every brand and product mention, is this piece of content still genuinely useful to the reader?

Different cells in the matrix meet this standard in different ways. Pure educational guides must pass it completely. Commercial product highlights must first resolve a real reader problem before introducing product capabilities.

To see how an automated system converts this matrix into consistent daily output, explore the [New Media Content Factory](/en/content-factory/).