R.I.P. Search Volume

TL:DR: Traditional keyword search volume is becoming obsolete thanks to AI like ChatGPT. People use conversational prompts, not just keywords, and the AI handles much of the search journey. Forget optimizing for keywords; focus on influencing the AI's understanding and recommendations. Track 'prompt-driven visibility' – how often your product/service appears in relevant AI responses. This requires manual tracking (for now) and a shift towards context-rich content that addresses specific use cases and compound user intents. The future isn't about keyword counts; it's about being the solution the AI remembers and recommends within the right context.

Welcome to the Era of LLM Demand and the New SEO

We're living through one of the most significant shifts in the digital landscape since the dawn of the internet: the rise of Large Language Models (LLMs) like ChatGPT, Perplexity, and Claude. As technologists and business leaders, we must pay attention, because these changes are fundamentally altering how information is found and consumed online. And that means the old rules of Search Engine Optimization (SEO) are quickly becoming history.

For years, 'search volume' has been the holy grail metric. How many people are searching for "X keyword" per month? It dictated content strategy, ad spend, and perceived market demand. But I'm here to tell you: Search volume, as we know it, is dead. Relying solely on it now is like navigating with a map from the wrong century.

Why "Search Volume" Is Failing Us

Think about how you use tools like ChatGPT. You don't just type "best CRM." You might ask, "What's a good CRM for a small tech consultancy that needs strong project management integration and doesn't cost a fortune?"

See the difference?

1. Queries vs. Prompts: Traditional search volume tracks discrete keyword queries. LLMs operate on conversational prompts that are nuanced, contextual, and often multi-faceted.
2. The AI Middleman: When someone uses Google, they sift through the results. When they use an LLM, the AI synthesizes information, compares options, and often provides a direct answer or recommendation. It's doing 90% of the traditional "search journey," and that journey is largely invisible to current SEO tools like Ahrefs or Semrush.
3. Optimizing for the AI: You're no longer just trying to rank for a human searcher. You need your content and brand presence to be understood, processed, and recalled by the AI's reasoning engine. Your information needs to fit the model's logic and surface naturally within its generated responses and comparisons.

Introducing the New North Star: Prompt-Driven Visibility

If search volume is out, what's in? We need to start thinking about Prompt-Driven Visibility.

Forget obsessing over "Keyword X has 3,200 monthly searches." The crucial questions now are:

• In response to what kinds of user prompts does our product or service get mentioned?
• What specific problems or use cases lead the AI to recommend us?
• In which categories are we consistently surfaced as a viable option?
• Which specific AI models (ChatGPT, Perplexity, Claude, etc.) are organically mentioning our brand?

This is the true measure of demand in an AI-first world. It's not about raw search numbers; it's about relevance and recall within the AI's conversational context.

Tracking the Untrackable (For Now)

The bad news? Your standard SEO dashboards aren't built for this yet. They track keywords, not conversational context within AI responses.

The good news? You can start tracking this manually. It requires effort, but the insights are invaluable. Here’s how we're starting to approach it at Mercury:

1. Create an "LLM Visibility Tracker": A simple spreadsheet (yes, Google Sheets works fine!) is your starting point.
2. Log Prompts: Whenever you see your product, service, or even competitors mentioned in an LLM response, log the exact prompt that triggered it.
3. Track Across Models: Test relevant prompts across different LLMs (ChatGPT-4, Claude 3, Perplexity, etc.) as they often have different knowledge bases and response patterns.
4. Capture Details: Note the position/prominence of the mention, the specific phrasing used, the date, and grab a screenshot for reference.

This isn't automated (yet!), but it's the groundwork for understanding your actual visibility where it increasingly matters.

Deconstructing LLM Demand: Beyond Keywords

LLM demand isn't about isolated keywords; it's a combination of Topics x Use Cases x Context.

Instead of targeting the keyword "best video hosting," you need to understand the context where that need arises. For example:

• A course creator might prompt: "What's a secure, private Vimeo alternative specifically for hosting online course videos?"
• A SaaS company might ask: "Recommend a secure video streaming solution that integrates easily with educational platforms."

The demand lives within these specific, contextual scenarios. Your content needs to address these nuanced needs directly.

Influencing the AI: The Three Layers

To effectively win in this new landscape, you need to influence LLM demand across three critical layers:

1. Prompt Layer: Understand the language your potential customers use when asking the AI for help. What problems are they trying to solve? What comparisons are they asking for?
2. Response Layer: Aim for your brand/product to be explicitly mentioned in the AI's generated answers, comparisons, and recommendations.
3. Source Layer: Ensure the underlying information the AI draws from (articles, forums, documentation, reviews) mentions your brand positively and accurately in relevant contexts.

Dominating means having a presence and influence across all three.

Strategies for the LLM SEO Era

How do you actually influence these layers?

• Seed Use-Case Content: Create detailed content (blog posts, case studies, tutorials) that addresses specific problems and use cases, providing strong context for why your solution fits.
• Earn Brand Citations: Encourage mentions of your brand on reputable forums (like Reddit, Stack Overflow), in technical documentation, high-quality review sites, and industry discussions. LLMs index and learn from these sources.
• Solve Compound Intent: Develop content that addresses multi-step user journeys or complex questions within a single piece. Think "How-to guides" that incorporate tool recommendations.
• Prioritize Utility: Remember, LLMs are designed to be helpful. They often prioritize practical, useful information that directly solves a user's stated problem over sheer brand authority alone. Make your content incredibly useful.

Finding Gold: Prompt-First Market Research

Want to get ahead? Start actively looking for real-world prompts:

• Monitor Communities: Browse relevant subreddits, Slack channels, Discord servers, and forums where your target audience discusses their challenges. Note the exact language they use.
• Test Prompts: Plug these real-world prompts into ChatGPT, Claude, etc., and see which solutions (yours and competitors') appear.
• Reverse Engineer: Analyze why certain products get recommended. What content are they publishing? Where are they getting citations? This is your roadmap.

The Big Shift: Context is King

Let's simplify the transition:

• Google SEO: Demand lives primarily in keywords.
• LLM SEO: Demand lives primarily in context.

Context – the specific use case, the user's underlying goal, the comparison being made – doesn't show up neatly in a monthly search volume report. It gets mentioned, understood, and remembered by the AI.

Adapting to this requires a fundamental shift in how we think about content, visibility, and measuring success. It's challenging, yes, but also incredibly exciting. The businesses that embrace this change and learn to influence LLM demand effectively will be the leaders of tomorrow.

Let's embrace the future of search, together.