Every Purchase Begins Before the Customer Knows They’re Shopping

The purchase funnel is one of marketing's most enduring ideas. Awareness leads to consideration, consideration leads to intent, intent leads to purchase. Clean, sequential, satisfying. And mostly wrong — not because the stages are false, but because the model quietly assumes that the journey begins when the customer raises their hand. It does not. Brand preference, trust, and category familiarity are forming long before a person types a single search query or clicks a single ad. And now, AI is compressing and front-loading that pre-intent phase in ways most marketing teams are not built to address. This article works through what the traditional funnel gets right, what it consistently misrepresents, how AI has changed the terrain, and what that demands of marketers today.

Key Takeaways

• Consumer preferences form significantly before any detectable intent signal — the "dark funnel" of private conversations, ambient exposure, and passive content consumption shapes decisions before the shopper knows they are shopping.
• The traditional purchase funnel models a journey that is already well underway by the time brands can measure it.
• According to a Bain and Company survey (December 2024, n=1,117), about 80% of consumers now rely on zero-click AI results in at least 40% of their searches, meaning brands are losing influence at the very moment discovery used to happen.
• AI is not replacing the consideration stage. AI is handling early exploration so efficiently that consumers arrive at search with more refined intent than before, compressing the visible part of the journey.
• The operational response is not better retargeting or smarter keywords. It is brand presence in the ambient, pre-intent space — AI-readable content, topical authority, and consistent visibility across channels that do not convert directly but shape the decisions that do.

The Classic Model (And Why It Stuck)

The traditional purchase funnel has been around in various forms since the late nineteenth century. Awareness, interest, desire, action. Later iterations added consideration, loyalty, advocacy. It was adopted by nearly every marketing organization on earth, and for a long time, it was a reasonable approximation of how buying decisions worked.

The model made sense in a world with limited information channels. Consumers had access to a finite number of stimuli — a television commercial, a magazine ad, a recommendation from a neighbor — and they progressed in a broadly observable sequence from not knowing about a product to buying it. Marketers could intercept that sequence at predictable points.

Google refined this significantly in 2011 with the Zero Moment of Truth, or ZMOT. The concept acknowledged that before consumers reached any store or product page, they were conducting independent online research. Searching. Reading reviews. Watching videos. Comparing specifications without telling anyone. Think with Google data showed ZMOT rated as effective by 84% of consumers in influencing decisions, ahead of the traditional in-store First Moment of Truth at 77% and Stimulus (advertising exposure) at 76%. This was an important update to the funnel model, because it named a stage that brands were systematically failing to show up for.

So the traditional funnel is not useless. It is a map with missing terrain.

What Everyone Keeps Misunderstanding About When Intent Forms

Here is the problem with ZMOT as a starting point for marketing strategy: it still treats the research phase as the beginning. But by the time a consumer types "best project management software for small teams" into a search bar, several things have already happened. They have been exposed to brand mentions in Slack channels. Someone in their network talked about a tool at a dinner they half-attended. They scrolled past a sponsored post without clicking, but they registered the name. A podcast host mentioned a product in passing three months ago. None of these moments appear in any analytics dashboard. None of them trigger a trackable signal.

This is what the B2B marketing world has started calling the dark funnel, a term for all the interactions that shape buyer intent before any direct brand engagement occurs. Private conversations, incognito browsing, word-of-mouth in professional communities, peer recommendations inside buying committees — these are the invisible preconditions of a decision that later looks, from the outside, like it began with a search.

And the data on what it actually takes to move someone from stranger to buyer makes this concrete. According to a 2025 report by Focus Digital, the average number of meaningful contact points required to close a sale is 28.87, though this varies enormously by purchase complexity.

The mistake most marketing teams make is assuming that "first tracked touchpoint" equals "first meaningful interaction." It rarely does. The actual formation of a shortlist — the set of brands a person would even consider when a need becomes active — happens diffusely, over time, through channels that are difficult or impossible to attribute. A brand that is not present in those pre-intent moments does not lose the sale in the consideration phase. The loss happened earlier, quietly, when the shortlist was being built without them.

This is also why last-touch attribution is such a misleading way to allocate marketing budget. It credits the final interaction and ignores the long chain of ambient exposures that made the final interaction possible.

How AI Changed Where the Journey Begins

The trajectory described above — where preferences form gradually and passively before intent becomes explicit — has always been true. What AI has done is accelerate and formalize the early exploration phase in a way that changes how visible the journey is, and where brands need to be.

Bain and Company's December 2024 survey of 1,117 consumers found that about 80% of people now rely on zero-click AI results in at least 40% of their searches. Roughly 60% of searches now end without the user ever clicking through to another site. Organic web traffic has dropped an estimated 15% to 25% as a result. ChatGPT saw a 44% traffic boost in November 2024, and Perplexity reached 15 million monthly users in the same period.

This matters because the traffic that is disappearing is not random. Bain identifies this lost traffic as concentrated in non-branded, high-value discovery searches — the exact moment where new customers used to first encounter a brand.

The IAB and Talk Shoppe study found that 80% of shoppers use AI to research and compare products, with usage highest at the beginning and middle of the shopping journey. BCG research adds that 64% of consumers use AI in their purchase process across three or more product categories, and consumers report that generative AI is decisive in shaping their confidence in a purchase decision. Critically, BCG notes that AI helps consumers "figure out what they want even if they don't initially know." That is the pre-intent phase becoming an AI-mediated conversation rather than a passive accumulation of impressions.

The consumer still forms intent before they know they are forming it. But AI is now the medium through which that formation is happening. The question a person might have asked a friend, or researched via a ten-tab browser session, or absorbed from months of industry content — all of that is now being handled by a single AI conversation that synthesizes information and returns a verdict.

Microsoft Advertising data reinforces the directional shift. Single-word search queries have declined by 6%, while longer, more specific multi-word queries leading to AI answers are up significantly year-over-year. Consumers are arriving at search with more refined intent. They have already done the broad exploration phase, they just did it inside a chatbot rather than across a dozen web pages.

The journey is not getting shorter. The part brands could formerly see and measure is getting shorter. The invisible pre-intent phase is where more is happening, faster.

What This Means for How You Actually Show Up

The operational implications here are genuinely uncomfortable for teams that have organized their marketing around measurable intent signals.

The goal is now to appear in AI answers, not just search results. Bain recommends that brands optimize for AI crawlability — semantic content, long-tail specificity, and accessible formats that AI systems can read and cite. PDFs and gated content are effectively invisible to most generative AI systems. If your best thinking is locked behind a form, it is not contributing to brand visibility in AI-generated answers. That is a direct cost to pre-intent influence.

Topical authority matters more than keyword coverage. A brand that publishes one shallow article on ten adjacent topics will not be cited by a language model the way a brand with ten deeply useful articles on one specific topic will. Depth signals expertise to both search algorithms and AI systems. Shallow coverage signals otherwise.

Ambient presence is not a nice-to-have. The dark funnel is not going away. Private conversations, community discussions, podcast mentions, and social sharing in closed groups continue to shape purchase shortlists. Brands that only show up when someone is actively searching for them are showing up too late. The consideration set was already narrowed.

Metrics need to shift. Bain explicitly recommends moving from click-focused metrics to measuring search impressions, AI reach, and share of voice in non-branded category conversations. This is a different kind of reporting, and it requires different internal conversations about what success looks like. Most companies are not there yet. Most attribution models are not built for it.

The honest version of this is that some of the highest-impact work a marketing team can do — building genuine topical authority, generating community mentions, being the brand whose name comes up when someone asks a trusted colleague — is the hardest work to justify in a quarterly dashboard. That tension is real. It does not resolve cleanly.

The Part Nobody Wants to Hear About Patience

There is a version of this argument that leads to the comforting conclusion: invest more in brand, run some awareness campaigns, make sure you show up on podcasts. That is not wrong, but it understates the specificity required.

Building pre-intent visibility is not about generic awareness. Knowing a brand exists is different from trusting it enough to put it on a shortlist. A consumer who vaguely recalls seeing your name in a banner ad three months ago has not formed any meaningful preference. A consumer who read a genuinely useful piece of content you published, or heard your category framing referenced by a podcaster they trust, or watched one of your tools appear in an AI-generated comparison — that person has. The distinction matters for how resources are allocated.

The consumer decision journey, as the additional context for this piece outlines, now includes an "initial consideration" stage where consumers form a shortlist based on previous experiences and brand perceptions before active research even begins. That initial consideration set is heavily influenced by passive, pre-intent exposure. If a brand is not part of that set before the active evaluation phase begins, the odds of being added during evaluation are not impossible but are considerably worse.

There is also something worth naming about the asymmetry between building this visibility and losing it. Topical authority compounds slowly. A brand that builds a consistent presence in a category's conversations over two years has an advantage that is genuinely difficult for a competitor to replicate quickly. The reverse is also true: a brand that lets that presence erode by shifting to performance-only spend can find itself absent from the initial consideration set faster than it realizes.

The Signal You Are Looking For Is Already Gone

Before you start, it is worth acknowledging what this reframe asks you to give up.

Most marketing organizations are built around detectable signals. Clicks, opens, form fills, session starts, attributed conversions. These are real and they matter. But they are, almost by definition, measures of the visible tip of a journey that started somewhere else, at some earlier point, through some channel that left no trace.

The consumer intent signals that shape purchase decisions build across exploration, comparison, and validation, as Microsoft Advertising frames it, until a brand aligns with a consumer at the right moment. But that alignment is prepared for long before the moment arrives.

The funnel is not wrong. Every purchase does move through something like awareness, consideration, and decision. The misunderstanding is treating the funnel as a description of where and when the work of influence happens. Most of that work happens in the ambient, pre-declared, largely untracked phase that precedes the funnel's opening. And AI has made that phase faster, more decisive, and less visible than it used to be.

The brands that figure out how to show up there will have advantages that are real but maddeningly hard to display on a slide.

Frequently Asked Questions

What is the "dark funnel" in consumer behavior?

The dark funnel refers to all buyer activity that occurs outside traditional tracking and analytics. This includes private conversations, peer recommendations, incognito browsing, community discussions, and word-of-mouth referrals. These interactions shape purchase intent before a consumer ever directly engages with a brand, and they cannot be attributed through standard last-click or multi-touch models. The dark funnel is not a failure of measurement — it is a structural feature of how people actually form opinions and shortlists.

Why does AI-driven search change pre-purchase behavior?

AI search tools, including generative AI platforms and AI summaries on traditional search engines, allow consumers to conduct broad early-stage research in a single conversation rather than across dozens of web pages. According to IAB and Talk Shoppe research, 80% of shoppers use AI to research and compare products, primarily at the start of their purchase journey. Because AI handles this early exploration phase, consumers arrive at traditional search with more specific queries and more refined intent. The discovery phase has not disappeared — it has moved to an environment where brand visibility works differently.

How should marketers think about content strategy given pre-intent purchasing behavior?

Content needs to serve two distinct purposes. The first is directly generating traffic and leads from people who already know what they want. The second is building topical authority and ambient visibility in the pre-intent phase, where brand perceptions and initial consideration sets are forming. The second purpose is often underinvested because it is harder to attribute. Practically, this means investing in genuinely deep content on specific topics, making that content accessible to AI systems by avoiding gating, and pursuing presence in community channels, podcasts, and peer conversations where organic brand mentions form.

What is the difference between brand awareness and pre-intent brand trust?

Awareness is knowing a brand exists. Pre-intent trust is having a favorable disposition toward a brand before you have a declared need. Generic awareness campaigns can generate the former without building the latter. The distinction matters because consumers building an initial consideration set are drawing on prior impressions that carry some weight — not just name recognition. A brand that has built trust through useful content, consistent community presence, and credible third-party mentions has a fundamentally different standing in that consideration set than one whose name was seen in a banner ad.

Should brands stop investing in bottom-of-funnel tactics?

No. Bottom-of-funnel tactics capture intent that already exists and convert it efficiently. They remain important, and for most businesses, they drive significant measurable returns. The argument here is not to stop doing those things. The argument is that exclusively investing in bottom-of-funnel tactics means accepting that your brand's presence in the initial consideration set is being determined by factors you are not actively managing. For some categories and competitive conditions, that is an acceptable position. For many, it is a growing liability.

What metrics can actually capture pre-intent visibility?

This is where most analytics setups fall short. Bain recommends shifting toward measuring search impressions, AI reach, and share of voice in non-branded category queries rather than relying exclusively on click-through and conversion metrics. Qualitative inputs also matter: tracking how often your brand appears in industry conversations, community forums, and AI-generated category comparisons provides signal that traditional dashboards miss. Customer surveys asking how someone first heard of or formed an opinion about your brand frequently reveal channels that attribution models credit to zero.