---
id: "concept-strategic-text-sequence"
type: "concept"
source_timestamps: ["§ [ Stage 3 ] Make Other AI Agents Choose Your Brand"]
tags: ["llm-optimization", "algorithmic-marketing", "growth-hacking"]
related: ["action-implement-sts", "concept-share-of-model", "contrarian-nonsensical-optimization", "prereq-seo-mechanics"]
definition: "Algorithmically generated, often nonsensical text embedded in product pages to manipulate LLMs into ranking the product as a top recommendation."
sources: ["agentic"]
sourceVaultSlug: "hbr-seg-agentic"
originDay: 6
articleStem: "hbr-ext-18-preparing-brand-agentic-ai"
sourceUrl: "https://hbr.org/2026/03/preparing-your-brand-for-agentic-ai"
sourceTitle: "Preparing Your Brand for Agentic AI"
---
# Strategic Text Sequence (STS)

A **Strategic Text Sequence (STS)** is an algorithmically generated sequence of text — often entirely nonsensical to human readers — embedded into a product's information page to manipulate LLM attention mechanisms and increase the likelihood of the product being recommended.

Harvard Business School research demonstrated its efficacy with fictional coffee brands: 'ColdBrew Master' was initially *excluded* by LLMs due to its high price, but became the **top recommendation** after an STS was inserted into its page. This highlights a divergence from traditional SEO (which still needs human-readable content — see [[prereq-seo-mechanics-d6]]): LLM optimization can be 'hacked' using adversarial or strategic token sequences. Operationalized via [[action-implement-sts]]; feeds [[concept-share-of-model]]; the counterintuitive core is captured in [[contrarian-nonsensical-optimization]].

**Enrichment / verification.** The provided sources do not directly validate this specific HBS-backed technique. It sits in a credible adjacent literature on **prompt injection / adversarial text** (models are known to be manipulable by hidden or non-human-readable inputs). Governance caveat: widespread use would likely trigger model-side defenses, search-quality penalties, platform policy enforcement, or legal scrutiny — making STS an unstable long-term strategy versus clean structured-data approaches like [[concept-llms-txt]].