In the article “How Marketers Can Adapt to LLM-Powered Search” by Stefano Puntoni, Mike Ensing, and Jarvis Bowers, published in Harvard Business Review, the authors explore how large language models (LLMs)—AI models designed to understand and generate natural language—are reshaping online search and marketing strategies. Unlike traditional search engines, LLMs provide direct answers rather than lists of links, creating an interactive experience that challenges brands to position themselves effectively to reach their target audience.
LLMs and the Evolution of the Search Experience
LLMs, such as those integrated into ChatGPT, Google’s Search Generative Experience (SGE), or Bing Copilot, create natural language responses that guide users throughout their buying journey. This transformation offers a more seamless and consultative experience. However, it introduces a challenge for brands: If an LLM doesn’t mention your brand in its answers, users may never discover your offer. This shift requires companies to adopt a new approach to ensure brand visibility.
LLM Optimization: A New Discipline for Marketers
Similar to search engine optimization (SEO), a new discipline of LLM optimization or LLMO is emerging. Unlike traditional algorithms, LLMs do not rely solely on keywords and links. They compile responses using diverse data sources: text content, images, videos, customer reviews, and even user-generated content. Brands must ensure that their content is structured in a way that is easily understood and usable by these new systems.
Retrieval-Augmented Generation (RAG): A Strategic Lever
LLMs employ a technique known as retrieval-augmented generation (RAG), which enhances responses with external information such as company documents or web content. An example cited by the authors shows that adding a “strategic text sequence” to a product page increases the likelihood that the LLM will cite this product in its answers. Research by Harvard demonstrated that this technique could, for example, bring a product to the forefront for queries related to affordability, offering brands a new way to influence outcomes.
The Shift from SEO to LLMO and Its Implications for Careers
The transition from traditional SEO to LLM optimization (LLMO, for Large Language Model Optimization) brings both challenges and opportunities. Unlike SEO, which focuses on optimizing websites to appear in traditional search engine results, LLMO encompasses practices that maximize a brand’s visibility in responses generated directly by LLMs, bypassing the need for link-based listings.
While automation tools reduce the need for manpower in certain SEO tasks, the required skill set is becoming more complex. SEO professionals specializing in LLMO will be essential for managing brand visibility across various platforms and meeting the demands of this rapidly evolving environment.
Practical Strategies for Implementing LLM Optimization
Although the article by Stefano Puntoni, Mike Ensing, and Jarvis Bowers does not provide detailed practical guidelines, here are concrete steps for companies aiming to maximize their visibility through retrieval-augmented generation (RAG).
Identify Key Queries for Your Brand
Begin by identifying common questions that users ask and that your products can answer. For example, a bicycle brand might target queries like “best bikes for beginners” or “affordable mountain bikes.”Create “Strategic Text Sequences”
LLMs easily capture information presented directly and concisely. Include specific product features in descriptions, such as “optimal comfort for beginners” or “wide tires for better balance,” which align with users’ search criteria. Highlighting unique attributes like durability or affordability increases your chances of being mentioned.Optimize Content Structure
Organize your content with well-structured tags and section headings that reflect popular questions, making it easier for LLMs to index the information. Use short paragraphs to make the information easier to extract, and include varied formats (customer reviews, FAQs, guides).Add Additional Data Sources
LLMs draw on various types of content to generate responses, so integrating customer reviews or case studies increases your chances of being cited. Enrich your pages with multimedia, such as clearly tagged images and videos.Continuous Testing and Adjustment
Measure the effectiveness of your strategic text sequences and adjust them to capture more attention in LLM search results. Run A/B tests to validate the most effective phrasing. This will help you understand what works best for getting mentions in LLM responses.
A Changing Search Ecosystem
LLMs pave the way for a diversified search ecosystem. Instead of one dominant player like Google, multiple platforms could emerge with different monetization models, such as subscriptions or ads. This evolution broadens options for consumers and may reduce brands’ dependence on a single platform.
Conclusion: Adapt Your Strategy to Thrive in the LLM Search Era
As LLM-powered search continues to transform digital strategies, it is essential for brands to rethink their approach to visibility. By applying LLMO strategies and integrating retrieval-augmented generation techniques, companies can ensure that their products and services remain well-represented in the responses generated by these new models.
This article, inspired by insights from Stefano Puntoni, Mike Ensing, and Jarvis Bowers, provides additional recommendations to help marketers fully leverage the potential of LLM-powered search. To learn more, read the full article on Harvard Business Review here.
