How Many Brands Can a Market Hold? Sphere Packing Bounds for Multi-Dimensional Positioning

How many distinguishable brand positions can a market sustain? We formalize brand perception as points in \(\mathbb{R}^8_+\) under the Aitchison metric (Aitchison, 1986; Egozcue et al., 2003), which is isometric to Euclidean distance on a 7-dimensional hyperplane after centered log-ratio transformation. Two brands are distinguishable if their distance exceeds perceptual threshold \(\varepsilon\). The problem maps onto sphere packing. Because the \(E_8\) lattice achieves the unique optimal packing in eight dimensions (Viazovska, 2017), its density (\(\pi^4/384 \approx .2537\)) and kissing number (240) supply structural bounds. We derive five results: (1) volume-ratio capacity is at least \((1/\varepsilon)^8\), equaling \(10^8\) positions at \(\varepsilon = .10\); (2) each position has at most 240 nearest neighbors decomposing into 112 specialist (two-dimensional) and 128 generalist (eight-dimensional) vectors; (3) 10,000 brands occupy less than .01% of the unit 8-ball; (4) category saturation occurs near \((1/\varepsilon)^{d_\text{eff}}\); (5) average inter-dimension correlation \(\rho = .3\) collapses effective dimensionality to approximately 2.6 and capacity by five orders of magnitude. An LLM stability experiment (250 calls, five models) finds null competitive-interference effects, consistent with fixed geometry. The \(E_8\) connection is structural, not literal: it establishes the mathematical ceiling on positioning capacity. We discuss implications for white-space strategy, correlation management, and observer-dependent capacity. Includes zharnikov-2026g-r4.yaml (Paper Spec v0.1.0) — a machine-readable specification of the paper's claims, assumptions, and dependencies. The paper's full machine-first bundle (the SPINE claim/dependency graph and the ONTOLOGY term module) lives in the public repository; see https://github.com/spectralbranding/paper-spec for the standard. This PDF is generated programmatically from that machine-first source under a research-as-repository model.