AI – Transforming Holiday Shopping and Retail Marketing Strategies

I. Shift in Holiday Shopping Search Behavior

During this holiday season, a significant trend is emerging among American consumers. Instead of relying solely on Google, an increasing number are expected to turn to large – language models for finding gifts, deals, and sales. A recent shopping report from Adobe indicates that retailers may witness a substantial upsurge of up to 520 percent in traffic from chatbots and AI – based search engines in 2024 compared to the previous year. OpenAI has already taken steps to leverage this trend. Last week, the developer of ChatGPT announced a major partnership with Walmart, enabling users to purchase goods directly within the chat window.

II. The Emergence of Generative Engine Optimization (GEO)

As consumers start depending on chatbots to explore new products, retailers are compelled to reevaluate their online marketing approaches. For decades, companies have been engaged in search – engine – optimization (SEO) strategies to manipulate Google’s search results. Now, to gain visibility among AI bots, brands are embracing “generative engine optimization” (GEO). Market research estimates suggest that this burgeoning industry could be worth nearly $850 million this year.

A. Relationship between GEO and SEO

GEO is, in many respects, an evolution of SEO rather than an entirely new concept. Many GEO consultants have roots in the SEO domain. Since the fundamental objective of both remains the same – anticipating user questions and ensuring that relevant content appears in the answers – some of the traditional SEO strategies may still be applicable. However, there is mounting evidence that chatbots surface different types of information compared to search engines.

B. Divergence in Information Sources

Imri Marcus, the CEO of the GEO firm Brandlight, estimates that the overlap between the top Google links and the sources cited by AI tools has significantly decreased. Previously, it was around 70 percent, but now, according to him, this correlation has dropped below 20 percent.

C. Content Preferences of Chatbots

Search engines often favor verbose content, as seen in the long blog posts above recipes on cooking websites. In contrast, Marcus notes that chatbots tend to prefer information presented in simple, structured formats, such as bulleted lists and FAQ pages. An FAQ page, he argues, can address a hundred different questions, providing more options for AI engines to select from, rather than a single article touting the overall brand.

D. Specificity of User Queries

The queries directed at chatbots are typically highly specific. Marcus points out that no one would ask ChatGPT, “Is General Motors a good company?” Instead, users might inquire about the driving range of the Chevy Silverado or the Chevy Blazer. Hence, creating more specific content can yield better results, given the highly targeted nature of user questions.

III. Impact on Brand Marketing Strategies

These insights are influencing the marketing strategies of Brandlight’s clients, which include prominent brands like LG, Estée Lauder, and Aetna. Brian Franz, the chief technology, data, and analytics officer at Estée Lauder Companies, emphasizes the need to ensure that product information and authoritative sources are tailored to feed the AI models. When asked about the possibility of partnering with OpenAI to enable in – chat – window shopping for Estée Lauder products, Franz responds affirmatively.

A. Focus on Awareness over Immediate Sales

For now, brands are more concerned with consumer awareness rather than directly converting chatbot mentions into sales. The goal is to ensure that when consumers ask questions like “What should I put on my skin after a sunburn?” their products are among the results, even if immediate purchases are unlikely. Franz remarks that in this early, rapidly evolving stage, focusing on the return on investment (ROI) of specific content may not be the best approach.

B. Use of AI in Content Creation

To develop AI – optimized content, companies are turning to AI itself. Marcus dispels the initial speculation that AI engines would not train on AI – generated content, stating that this is not the case.

This is an edition of the Model Behavior newsletter. Read previous newsletters here.