The Search Intent Framework

The 2002 model

Marketers still build their strategy on a model from 2002.

In 2002, Andrei Broder published an academic paper titled 'A Taxonomy of Web Search'. His insight was simple but powerful: people don't just search for information. Sometimes they want to navigate to a specific website, or execute a transaction. Broder proposed three categories: informational, navigational and transactional. Google later added a fourth: commercial. Since then, little has changed in terms of intent categorization. The four categories appear in every SEO course, every marketing textbook, and are baked into tools like Semrush and Ahrefs.

We know those four categories today as: Informational, Navigational, Commercial and Transactional. Google reformulated them in the early 2010s as 'Know, Go, Do, Buy micro moments'. Sounds more modern. But they're still the same four buckets.

But the world looks radically different in 2026. People no longer search in one place, and they don't search the same way everywhere. They choose per situation which tool works best: ChatGPT for in-depth advice, YouTube for video instructions, Reddit for independent opinions, TikTok for trends, Pinterest for inspiration. Their search behavior differs per platform because their intent differs. Intent determines the platform. Not the other way around.

Three reasons why the old model falls short

1. It's outdated

The model was built for a world where traditional search engines were the only gateway to information. Websites were what ranked. In that context, it made perfect sense. But your customers no longer search only on Google. They discover on TikTok, compare on YouTube and Reddit, ask questions to ChatGPT, seek inspiration on Pinterest. The search ecosystem has radically changed. The model hasn't.

2. It's incomplete

Search behavior is much broader than searching websites or making online purchases. The current model assumes a linear route: first information, then evaluation, then purchase. But that's not how it works. Users constantly jump back and forth. Emotion, inspiration and creativity play a central role in search behavior, but don't exist in the old model.

3. It's not refined enough

The four categories are too broad. 'Informational' can mean anything: a quick factual question or months of studying. 'Commercial' doesn't distinguish between someone cautiously exploring and someone nearly ready to buy. Generic answers don't win. Nuance and context are exactly what's missing.

Reality is not a funnel

The outdated model poorly reflects how consumers actually search in 2026. They don't search for one goal, they search across many more different places than just Google, and those intents constantly jump through each other.

Google calls this the Messy Middle. Rand Fishkin calls it a 'pinball machine'. Users constantly bounce between sources, (sub)intents and triggers. An ad triggers inspiration. Doubt brings someone back to comparing. A conversation with a friend triggers a new search query.

People search differently across platforms and combine the knowledge they gather along the way before making a choice. Take any purchase decision in 2026. Someone discovers a product via a short video on social media, then looks up honest experiences in an online community, watches an extensive comparison on YouTube, and finally asks an AI tool to summarize the pros and cons. None of those steps feel like traditional search, but every step is intent-driven. That's what modern search behavior looks like.

Search is not dead. The old playbook is.

A persistent narrative circulates in some marketing circles: search is dying. AI takes the answer, nobody clicks through anymore, Google is losing its grip. That narrative is wrong. The data tells a very different story.

More people search now than ever before. Search behavior grows globally by 25% year-over-year. AI search alone already generates 2 billion referral clicks per month, and that number is growing fast. Search is alive and well.

"Search is not dying. The old playbook is."

What has changed is where and how people search. And where answers are found. ChatGPT, Gemini and Perplexity provide faster, higher-quality answers to more and more questions. That shifts behavior, but it doesn't replace search. It expands it.

This is exactly why a model that only measures Google rankings and intents structurally falls short. Not because Google has become less relevant, but because the growth lies in platforms and behavioral patterns that fall outside that model. Those who only track their Google traffic may see a shrinking number. Those who map the complete search landscape see growth.

The reality of zero-click

Traffic is a vanity metric. That's the hard truth the profession hasn't fully accepted, and one that becomes ever more urgent as AI Overviews, featured snippets and direct AI answers take up more space in the search landscape.

Major tech publications lost more than 60% of their monthly website visits in 2024-2025 due to the rise of AI answers in search results. But their brand awareness, influence and authority in the field remained intact, or even grew. How is that possible? Because people read their content via AI summaries, were cited in answers, were referred to in communities. Without a single click on the website occurring. Reach and traffic are not the same thing. Google and AI Search are increasingly filtering informational queries themselves via AI Overviews and featured snippets. What remains and still clicks through are transactional and orientation intents. Fewer tourists, more buyers.

For every visitor who clicks, many more learn about your brand on platforms you don't control.

This also explains a pattern that appears more and more often: organizations that see their traffic decline but their pipeline grow. They don't have a problem. They have proof that their brand means something at the moment it matters. Intent determines whether someone clicks, not the ranking.

The question is therefore not: how do I generate more traffic? The question is: how do I influence decisions, directly and indirectly, regardless of whether someone ever clicks on my website? That's a fundamentally different strategic approach.

Direct: your brand is recommended in an AI answer, cited in a Reddit thread, or shown in a YouTube comparison. Someone makes a decision based on that. Indirect: a journalist writes about your category and mentions your name. A community member shares a positive experience in a Slack group. A podcast casually mentions your brand. No click, no session, no conversion in Google Analytics. But definitely influence on the decision that's made afterward.

In B2B this is most pronounced. B2B buyers compile shortlists via AI tools that rely on community sources, reviews and cited publications. Brands are chosen or written off without a single website visit having taken place. The better KPI is not "how many visitors did I have this month?" but "am I being recommended at the moment my potential customer is making a decision?"

The consequence is sharp: if your brand is only visible when someone types a search query into Google, you're too late. In many cases, the decision was already formed elsewhere: in the community discussion, the YouTube comparison, or the AI answer someone consulted before they even went to Google.

That reality changes how we need to think about marketing strategy. Brands that only optimize for website traffic miss the biggest part of the picture. The real opportunity: being present everywhere your audience has attention, and converting that attention into influence.

SEO was never about traffic. Traffic is a vanity metric. What it's really about: how do you ensure you're recommended at the moment your potential customer is orienting themselves on what you have to offer? Regardless of where that orientation takes place: on Google, in ChatGPT, via a Reddit thread, or via explainer videos on YouTube.

The new framework

Our framework accounts for all relevant platforms where users actively search for something. It's more complete, with intents the old model completely misses. And it's refined, with subcategories that capture nuance and context that make the difference between generic and truly relevant content.

The classic model vs. The Search Intent Framework

The difference between the classic model and this framework is not gradual. It's a different way of thinking about search behavior. Here's the core of that difference:

The nine intents aren't a replacement of the classic model for those who only work with Google. They're an expansion for those who want to understand the full search landscape, and want to be visible at the moment an actual decision is being made.

We don't claim this is perfect. Behavior changes. New platforms emerge and old ones disappear. This framework is therefore never finished. That is not a weakness; it is a characteristic of any living model. This is our best version today.

The intents explained

The framework consists of nine intents. The ninth, Local intent, is a dimension that runs across the other eight.

1. Navigation intent

When someone types your brand name, the decision is almost made. This is the moment of confirmation, not persuasion.

The user knows what they're looking for and wants to get there directly. This is the most direct intent: someone already has a destination in mind and wants to arrive as quickly as possible.

2. Information intent

This is the intent AI eats into the most. Only depth and structure keep you visible here.

The user wants to learn, understand or solve a problem, without a direct purchase thought. From a quick fact-check to months of studying: it's always about knowledge building without commercial intent.

Notable: YouTube is for Problem-solving not only the second platform for users, but also the second most cited social source in AI answers, after Reddit. Research by OtterlyAI (2026) across more than 100 million AI citations shows that 94% of cited YouTube videos are long videos (10-20 minutes), and that popularity metrics like views and likes have virtually no correlation with how often a video is cited by AI. Structure and content matter, not reach.

3. Orientation intent

This is where the shortlist is formed. If you are not visible in this phase, you do not exist at the moment of decision.

The user is in the process of making a decision but hasn't made it yet. They explore options, weigh alternatives and seek confirmation. The distinctive difference from information intent: here a purchase or action is in sight.

Dark Social plays a special role in Validation: a large part of real purchase validation takes place in private conversations. "Do you have experience with X?" in a WhatsApp group with colleagues or friends is functionally identical to a Reddit thread, only invisible to brands and analytics.

4. Transaction intent

This is where the money is made. But if you are only present here, you leave the decision to others.

The user has decided what they want. Now it's about arranging it: no longer which product, but where, how and from whom. The orientation phase is over; this is the moment of action.

5. Emotion intent

This is the intent the classic model completely ignores, while it accounts for the largest share of search traffic on TikTok and YouTube.

The user seeks a specific emotional experience, directly and now. Not gathering information, not making a purchase. Just feeling something: entertainment, relaxation, curiosity. This intent is completely absent from the classic model, while it's responsible for an enormous share of all searches on platforms like TikTok and YouTube.

Giphy and Tenor are the purest proof that search is also about self-expression. When someone searches for a GIF of a dancing cat or a head-shaking person, they're not searching for information. They're searching for a visual translation of their emotion to share. It's expressive search: the intent is not to find something, but to feel something and communicate that feeling.

6. Inspiration intent

Most purchase decisions start here. Months before anyone types a search term.

The user gathers ideas, dreams, or builds a vision, without direct purchase intent or time pressure. Where emotion intent is about a direct experience now, inspiration intent is about future-oriented collecting. The difference is subtle but crucial: inspiration is a save-for-later behavior.

7. Generation intent (the make-intent)

This is the fastest growing intent and the one that most directly replaces classic search behavior for templates, examples and checklists.

This is the odd one out, and deliberately so. Generation intent is not a search intent in the traditional sense. The user doesn't want to find something. They want to make, generate or have something devised. This is a fundamentally new intent born from the rise of AI tools.

Previously, someone searched for 'job application email template' and found an example. Now that same person asks ChatGPT to write one. The search for examples, templates and checklists is increasingly being replaced by direct generation.

Yet Generation intent is in this framework. For two reasons.

First: the platform choice is intentional search behavior. Someone who wants an email written goes to ChatGPT deliberately, not to Google. That navigation, from need to platform, is precisely what this framework maps.

Second: the boundary between searching and generating practically no longer exists in 2026. Perplexity searches and synthesizes. ChatGPT searches too now. Google generates answers via AI Overviews. A framework that ignores generative intent describes reality incompletely.

The numbers are telling. Large-scale research by SimilarWeb (2025) shows that 32% of ChatGPT usage falls under writing and creative ideation. By comparison: purely informational usage accounts for only 19% at ChatGPT, while at Google it's 64%. Two tools, two fundamentally different usage patterns. ChatGPT prompts average 60 words; Google searches average 3.4 words. These aren't variations of the same behavior. They are different behaviors.

8. Community intent

This is the most underestimated intent for B2B brands. And the place where most shortlists are formed without you knowing.

The user isn't looking for an answer, they're looking for a conversation. Connection with people who recognize the same situation, have been through the same problem, or share the same opinion. This is an intent completely absent from the classic model, while the behavior occurs enormously often and has an ever-growing influence on brand perception and purchase decisions.

The psychological starting point is fundamentally different from Information or Orientation: you're not looking for the right answer from an authority. You're seeking recognition and human perspective. This also explains why Reddit is now one of the most cited sources in AI answers, precisely because it contains real human experiences that can't be found anywhere else.

But the most invisible, and therefore most underestimated, layer of community intent is Dark Social. In WhatsApp groups, Slack channels and Discord servers, an enormous amount of human validation takes place daily. "Does anyone have experience with agency X?" in a WhatsApp group of twenty marketers, or "which CRM do you use?" in a Slack community. These are community searches that fall completely outside the sight of brands and analytics. Dark Social is the validation layer that never appears in a report, but does influence the purchase decision.

This is also the most underestimated intent in B2B. Research by LinkedIn Marketing (2025) shows that B2B buyers increasingly use AI tools to summarize markets, compare suppliers and compile shortlists; those tools rely heavily on community platforms for validation. Brands can be chosen or written off without a single website visit having occurred.

9. Local intent

When someone searches locally, action usually follows within minutes. No intent leads to a visit, a booking or a purchase as quickly.

Local intent is a separate dimension that cuts across other intents, but is strong enough to be treated as its own category. The distinguishing feature: the user has a geographic anchor.

The differences between intents explained

The nine intents are deliberately sharply delineated. In practice, some searches lie close to multiple categories. Below are the most important borderline cases, and how to tell them apart.

About Active Search

This framework exclusively covers Active Search: search behavior where the user takes the initiative. Algorithmically driven content, such as social feeds, recommendation algorithms or Google Discover, deliberately falls outside this model. Not because those channels are irrelevant, but because they follow a fundamentally different mechanism: the user isn't searching, but is being served. The boundary is sharp: if a user actively expresses an intent, it falls within this framework. If an algorithm decides what someone sees, it falls outside of it.

The impact on your SEO strategy

The framework is not a theoretical exercise. It has direct consequences for how you build content, which platforms you prioritize, and how you measure success.

1. Forget traffic, measure influence

For every visitor who clicks, many more learn about your brand on platforms you don't control. Website traffic is therefore by definition an underestimate of your actual reach and influence.

The data confirms this. Several major tech publications lost more than 60% of their monthly website visits in 2024-2025 due to the rise of AI Overviews, but their brand awareness and influence in the field remained intact. Traffic is not the same as influence.

In B2B this is even sharper. B2B buyers increasingly use AI tools to summarize markets and compile shortlists; brands are chosen or written off without a click having occurred. The better question: am I being recommended at the moment my potential customer is orienting themselves? On Google, in ChatGPT, via Reddit, via YouTube, via a WhatsApp group of colleagues, regardless of whether they then click on your website.

2. Community intent is the blind spot of most brands

Reddit threads, LinkedIn discussions and Facebook groups are completely ignored by most marketing teams, because they're not directly measurable and hard to influence. But that's exactly the place where potential customers form the most honest image of your brand. And it's the place where AI tools increasingly get their answers from.

On top of that: Dark Social is the validation layer that's even more invisible. WhatsApp groups, Slack communities and Discord servers don't count in any analytics dashboard, but they influence purchase decisions day in, day out. In a world where AI-generated answers sound the same everywhere, community-driven trust is the only thing that truly distinguishes brands.

The question is not whether your brand is being discussed in communities and groups. The question is whether you're part of it.

3. Stop thinking in funnels

The customer journey is not a funnel. It's a spaghetti mess of searches across different platforms, in random order. A potential customer might start on TikTok (Inspiration), go to YouTube (Comparison), ask ChatGPT for advice (Generation), read Reddit threads (Validation), discuss it in a WhatsApp group (Dark Social), and finally buy via Google Shopping (Transaction). Optimize for the behavior, not for the step in a funnel that never existed.

4. Be present in all relevant intents, on the right platforms

A large retail brand selling kitchens wants to be present for Inspiration searches on Pinterest and Instagram, for Comparison on YouTube and Reddit, for Validation via reviews on Google, and for Transaction on Google Shopping. Four platforms, four content formats, four KPIs; all feeding into the same purchase decision.

A real-world example: Descript, a 200-person company, successfully competes with Adobe in AI search results. Not through the biggest marketing budget, but by consistently being present everywhere AI looks: software lists on Zapier, user experiences on Medium, reviews on G2 and Gartner. When AI systems cite those trusted sources, Descript automatically comes along. Being present in AI answers is the new backlink.

This framework is a starting point

We present this framework with conviction and with openness. We believe this is a fundamental step forward compared to the 2002 model that has dominated the profession for too long. But we don't claim it's complete or definitive. Behavior changes. New platforms will emerge, categories will shift, and the boundaries between intents will sometimes remain diffuse. That's not a flaw in the model; it is the nature of human behavior.

Put the user at the center, not the platform. That's the first step toward an SEO strategy that works in 2026 and beyond.

This framework is free to use, share and supplement, provided you credit Recommendable as the source. We invite everyone to help build this framework. Does a boundary not hold up? Are you missing an intent? Do you see a platform we haven't assessed correctly? Contact us or send us a message on LinkedIn.