Let’s Talk About “Information Foraging Theory” in the Context of SEO/GEO: A Hamsterdam Marketing Lesson

By Ethan Lazuk

Welcome to another edition of Hamsterdam Marketing! 🐹

This is a project where we examine marketing topics to learn about their fundamental concepts and apply them to the context of SEO/GEO.

This time, we’ll be looking at information foraging theory.

This theory is about the tradeoff of gathering information vs. the time spent to learn it. It’s an interesting concept in light of AI search.

In traditional search, a user might examine several results, judge their titles and descriptions, then visit several websites and decide whether or not to keep looking. But with AI search, most of that work can be done for them: an AI assistant might consolidate information from multiple sources into a single response, potentially reducing the user’s need to explore individual websites.

This paradigm shift in user behavior introduces an interesting question regarding information foraging theory: if AI assistants increasingly handle the information-gathering process, how does that change the way consumers discover and evaluate brands?

What Is Information Foraging Theory?

Information foraging theory examines how people seek information while weighing its expected value against the time and effort required to obtain it.

The theory involves several foundational concepts:

• Information scent: The cues people use to predict whether a source will contain valuable information. (Strong information scent doesn’t always mean high information quality.)
• Information patches: The places where useful information is concentrated, such as websites, search results, or collections of documents.
• Information diet: People selectively consume information based on what appears most valuable for their needs, and their choices are influenced by relevance, familiarity, trust, and perceived effort.
• Information-seeking costs: The time and effort involved in finding, evaluating, and using information.

Information foraging theory was developed by Peter Pirolli and Stuart Card in the 1990s. It draws on optimal foraging theory from biology. Just as animals seek to maximize the energy gained from food relative to the effort of finding it, people make similar trade-offs when seeking information while minimizing the cost of obtaining it.

The theory also involves the optimal stopping problem, where people don’t necessarily stop searching when they have the best answer, but rather when the expected benefit of further searching no longer justifies the effort.

An interesting angle for SEO/GEO is diminishing returns within information patches, where the more time a user spends exploring a particular source, the less additional value they may find, which can motivate them to move elsewhere.

In short, information foraging theory is about maximizing the rate of valuable information gained relative to the cost of acquiring it

The SEO/GEO Connection of Information Foraging Theory

Getting back to our introduction, AI search may reduce information-seeking costs, thus altering user behavior compared to traditional search.

Another question is whether information scent shifts from traditional search results to AI-generated answers, because, while SEO/GEO professionals may optimize title tags, meta descriptions, and content to persuade users that a webpage is worth visiting, AI-generated answers now supply much of the information directly.

One detail worth adding is that information scent is subjective. The same information source can have strong scent for one person and weak scent for another, depending on their needs and prior knowledge. In the context of AI search, this might mean that a brand has strong information scent for a broad query but weaker scent for a highly specific use case, or conversely, a relatively unknown brand might appear especially relevant when an AI assistant explains how it addresses a user’s exact needs.

Let’s explore some of the SEO/GEO connections of information foraging theory in depth:

1. AI assistants can lower the costs of discovering, summarizing, and comparing information.

They do this by compiling multiple sources into a single AI-generated answer or conversation.

That said, AI assistants may introduce different costs than traditional search, such as verifying claims, checking citations, or identifying omitted information. This gives us a distinction between information acquisition costs and information verification costs.

As practitioners, we should consider whether AI makes people more willing to research complex questions because the initial effort is lower.

2. Information scent may shift to AI-generated descriptions.

While in traditional search, information scent often came from search snippets, titles, URLs, and recognized brands, in AI search, it might now also come from how an AI assistant characterizes a brand, describes its strengths, or contextualizes its relevance.

In short, the wording surrounding a recommendation in an AI-generated answer could influence whether users investigate a particular brand.

As practitioners, we should ask whether a brand receives stronger information scent when an AI recommends it with a specific explanation rather than merely mentioning its name.

3. AI-generated answers may serve as information patches.

It’s possible that a chatbot conversation might function as an information patch because users can satisfy multiple information needs without leaving the interface, such as by performing follow-up prompts that allow exploration of the same patch, instead of moving to external websites.

Alternatively, AI assistants could also direct users toward new information patches through citations and recommendations related to fanout queries, where the AI assistant conducts related searches to gather information.

As practitioners, we should think about whether traditional websites are becoming destinations for deeper verification rather than initial information discovery.

4. There are implications for brand discovery.

Compared to traditional search results, an AI-generated shortlist could concentrate consumer attention on fewer brands. On the other hand, the lower initial search costs could make it easier for consumers to discover unfamiliar or niche brands.

As practitioners, we should investigate these competing hypotheses based on select target audiences, rather than assuming AI necessarily reduces or expands brand diversity.

In my experience from talking with clients, users do consider fewer brands using AI search than traditional search, but they also get exposed to new or niche brands because the details of their prompts are more exact. For example, rather than searching just for SEO/GEO professionals in NYC, a prompt may also include details like “for my custom WordPress website for my architectural firm.” This makes the case for creating rich persona- or use-case-driven content on your website, both for AI-generated summaries and actual visitors.

For related reading, see my article on consideration sets, which talks about how AI might influence which brands enter the set before consumers visit their websites.

Practical Takeaways

Creating strong information scent is the first goal, as this is the cue users will use to judge if your result is worth their time, so focus on title tags, meta descriptions, and preview images.

If the information patch has shifted from the website to AI-generated answers, you’ll want to be present there, so make sure brand information is useful enough to be included by AI assistants. This means creating content with high information value, like original research or firsthand experience. That said, when it comes to information verification, your own website is more likely to be visited, so be sure to reduce its information-seeking costs, such as by structuring content with descriptive headings and clear navigation.

Lastly, you’ll want to understand user behavior by examining whether AI recommendations change the amount of independent research your target audience conducts. This can involve user studies. For example, you could compare the experience of researching on traditional Google results vs. ChatGPT, document which sources appear, how much effort is required, and which brands ultimately enter the user’s consideration set.

In short, the goal is to identify where information-seeking costs occur throughout the customer journey. For example:

• Where do consumers first encounter information about your brand?
• What questions remain unanswered after reading an AI-generated summary?
• Which questions motivate users to visit your website?
• What information do they need to verify before making a purchase?
• Can your content reduce the effort required to make that decision?

Answering these questions is a direct, practical application of information foraging theory that goes beyond simply optimizing for visibility.

Final takeaway

Information foraging theory says people seek information while trying to minimize the effort needed to obtain it. AI search has introduced an alteration to the way people can find information, reducing the initial information acquisition costs, but it also introduces new information verification costs.

As SEO/GEO professionals, we should conduct research into our target audiences to understand how they’re using AI assistants and what influence that has on their consideration sets for brands.

Do AI assistants reduce the number of brands that users consider by providing a truncated shortlist, or does the initially lower information acquisition cost and effects of query fanout introduce opportunities to explore more, potentially niche brands?

We’ll need to do research to find out, but information foraging theory can be a good structural concept for guiding that research.

See you next time!

I hope you’ve found this discussion around information foraging theory helpful for your SEO/GEO strategy.

Stay tuned for a new Hamsterdam Marketing lesson soon.

And if you want a hand with your SEO/GEO strategy, get in touch with me. I’m an independent consultant based in New York City.

Until next time, enjoy the vibes:

Thanks for reading. Happy optimizing!