In the spring of 2019, a B2B software company hired a consultancy to audit why its content marketing programme had produced almost no measurable pipeline in eighteen months. The company had done everything right by the conventional playbook: a blog updated twice a week, gated whitepapers for lead capture, a nurture sequence triggered by download, a scored lead handoff to sales at a defined threshold. The content team was producing work that the industry would recognise as competent. The revenue team could point to almost none of it as a source of closed deals.
When the consultancy interviewed recent customers about how they had first heard of the product, a pattern emerged quickly. Almost nobody had come through the funnel. They had heard about the company from a colleague at a conference. They had seen a founder's tweet thread shared in a private Slack community. They had read a detailed Reddit thread where the product was mentioned alongside two competitors. They had watched a YouTube explainer made by an independent creator. By the time they visited the company's website and downloaded a whitepaper, they had already formed a strong prior view about whether the product was worth their time. The whitepaper was confirmation, not discovery. The nurture sequence arrived into an inbox where the decision had already been substantially made.
This is not an unusual story. It is increasingly the typical one. And it reveals a structural problem with the marketing funnel model that the industry has been slow to fully acknowledge.
What the Funnel Model Actually Claims
The marketing funnel — awareness, interest, consideration, intent, evaluation, purchase — is a model of progressive qualification. It describes a population of potential buyers narrowing through stages, with marketing's job being to move people down through each stage as efficiently as possible. The model has roots in Elias St. Elmo Lewis's AIDA framework from 1898, was formalised for modern marketing practice in the 1950s and 1960s, and received its contemporary top-of-funnel/middle-of-funnel/bottom-of-funnel formulation in the early days of digital marketing, when the internet appeared to offer something previous channels had never delivered: measurable attribution at each stage.
The funnel's appeal is its tidiness. It gives every piece of content a job — awareness content attracts strangers, consideration content educates prospects, decision content converts buyers. It gives every campaign a measurable position in the sequence. It gives marketing and sales a shared vocabulary for handoffs and a shared set of metrics — impression share, click-through rate, lead volume, conversion rate — for evaluating performance at each stage. For a generation of marketers trained in this framework, it is less a model than an assumption about how reality works.
The problem is that it describes a buying process that almost no actual buyer follows.
Why the Funnel Was Always a Simplification
To be fair to the model, it was never meant to be a literal description of individual buyer behaviour. It was always a statistical approximation — a way of thinking about populations moving through a pipeline, not a claim that any specific person moves linearly from awareness to purchase without doubling back, stalling, exiting, or arriving via a completely different route than the model predicts.
The simplification was tolerable when the information environment was relatively controlled. In the broadcast media era, a brand could reasonably assume that a potential buyer's awareness of its product was shaped primarily by its own advertising and, to a lesser extent, by editorial coverage and word of mouth. The funnel's implicit assumption — that marketing controls the buyer's information environment and therefore controls movement through the stages — was close enough to reality to be useful.
The internet broke that assumption, and social media shattered it. Today, a potential buyer's awareness and consideration of a product is shaped by dozens of sources that exist entirely outside the brand's funnel: peer recommendations in private channels, independent reviews, creator content, community discussions, competitive comparison threads, employee posts on LinkedIn, disgruntled customer tweets. The brand's own content is one voice among many, and often not the most trusted one. The funnel describes the brand's intended journey for buyers. It has very little to say about the journey buyers actually take.
The Fragmentation of the Modern Buyer Journey
Research into B2B buying behaviour from Gartner has tracked the actual journey of enterprise software purchasers for several years. Their consistent finding is that the B2B buying journey is neither linear nor sequential. Buyers move between information-gathering modes — learning, problem framing, supplier exploration, solution validation, consensus building — multiple times, in no fixed order, revisiting earlier stages as new information arrives or internal stakeholders with new questions enter the process.
Gartner's 2022 research found that B2B buyers spend only 17 percent of their buying journey in direct contact with potential suppliers. The remaining 83 percent is spent on independent research, internal discussion, and evaluation activity that is invisible to the vendors being considered. This means that by the time a potential customer first makes contact with a sales team or fills out a form, the majority of their opinion-formation has already occurred in channels the vendor cannot observe and could not have influenced through funnel-stage content.
For B2C products, the fragmentation is equally pronounced but plays out differently. The growth of social commerce, creator-driven discovery, peer recommendation networks, and AI-powered search has distributed the consideration process across platforms in ways that make the traditional funnel almost impossible to map. A consumer who buys a skincare product may have first encountered it in a TikTok video, been reminded of it by a friend's Instagram story, read a Reddit thread comparing it to alternatives, looked up the ingredient list on a third-party review site, and then searched Google for a discount code — touching the brand's own properties only at the final moment of purchase. The brand's funnel captures the last step. It misses the entire journey.
Dark Social and the Attribution Illusion
The technical term for this missing journey is dark social — a phrase coined by Alexis Madrigal in a 2012 Atlantic article and since adopted broadly in marketing analytics. Dark social refers to sharing and discovery that occurs through private or untrackable channels: direct messages, email forwards, private group chats, Slack communities, WhatsApp groups, copied links pasted without UTM parameters. It is social in the sense that it is driven by peer-to-peer recommendation, and it is dark in the sense that it is invisible to standard analytics tools.
Research firm RadiumOne estimated in 2014 that dark social accounted for 69 percent of all online sharing. More recent estimates from ShareThis and Parse.ly suggest the proportion has grown as messaging apps have supplanted public social platforms as the primary venue for personal communication. When someone reads an article and sends it to three colleagues via Slack with a note saying "this is exactly what we should be doing," that sharing event registers as direct traffic in the analytics dashboard. The content's influence is real. Its attribution is invisible.
The consequence for funnel-based measurement is severe. If a substantial portion of the buying journey occurs in channels that analytics cannot observe, then the attribution data that the funnel model relies on is structurally incomplete. Last-click attribution — which assigns conversion credit to the final measurable touchpoint before purchase — is not merely imprecise. It is systematically misleading, consistently over-crediting the channels that happen to appear at the end of the visible portion of a journey and under-crediting everything that shaped the buyer's decision before they became visible to the brand.
This has real resource allocation consequences. Brands that manage their marketing spend by funnel stage and last-click attribution tend to over-invest in bottom-of-funnel tactics — paid search, retargeting, conversion rate optimisation — because these are the tactics that appear in the data at conversion time. They under-invest in the top-of-funnel and off-platform work — brand, community, creator relationships, thought leadership — that actually drives the awareness and trust that makes the bottom-of-funnel work possible. The result is a marketing programme that optimises efficiently for the visible portion of the funnel while neglecting the invisible portion that does most of the actual work.
The Content Treadmill and Why Most of It Goes Nowhere
The funnel model has also produced a particular failure mode in content marketing: the production of large volumes of content positioned at specific funnel stages without a clear theory of how that content will actually reach and influence the intended audience.
The logic, as typically applied, runs something like this: people at the awareness stage search for broad educational topics, so we should publish blog posts on those topics to capture organic search traffic. People at the consideration stage are evaluating options, so we should publish comparison guides and case studies. People at the decision stage are ready to buy, so we should offer demos and trials. Each piece of content has a funnel stage. Each stage has a conversion goal. The programme, in theory, moves people sequentially toward purchase.
In practice, this approach has two problems. The first is discovery. Organic search traffic to educational blog content has declined significantly as Google has prioritised its own featured snippets, knowledge panels, and AI-generated overviews for informational queries. The blog post that once attracted 5,000 organic visitors a month from a "what is X" query may now attract a fraction of that as Google answers the query directly in the search results. The content exists; the audience no longer arrives through the channel the content was built to capture.
The second problem is influence. Even when content is found and read, its capacity to move a reader through a funnel stage is far more limited than the model implies. Reading a blog post does not create awareness in the sense of emotional salience. It creates information exposure, which is a much weaker effect. The content that actually creates awareness — the kind where a potential buyer thinks of a brand unprompted when a relevant problem arises — tends to be content that is shared by trusted peers, encountered in communities, and encountered repeatedly across different contexts. A 1,200-word SEO-optimised blog post, however well written, rarely achieves any of those things.
What Actually Drives Purchase Decisions
Research on consumer and business purchase behaviour consistently identifies a small set of factors that account for most of the variance in buying decisions. The first is category entry points — the mental associations that cause buyers to think of a brand when a relevant need arises. These are built slowly, through repeated exposure across multiple contexts, and they are the primary mechanism by which brand awareness translates into purchase consideration. They are not built by a single piece of funnel-stage content.
The second is social proof at scale. Not the social proof widget on a landing page showing 1,247 five-star reviews, but the ambient awareness that people in the buyer's professional or social network have used and endorsed the product. This is fundamentally a function of market penetration — the more customers a brand has, the more likely any given prospect is to have a connection who can vouch for it — but it is also influenced by how effectively a brand activates its existing customer base as advocates. Customer stories shared by customers in their own networks carry far more persuasive weight than customer stories packaged as case studies on a vendor's website.
The third is distinctiveness. Brands that are easily recalled and easily distinguished from competitors in a category convert at higher rates than functionally equivalent brands that lack distinctiveness, because recall is the first step in consideration. This is a consistent finding across marketing effectiveness research from the Ehrenberg-Bass Institute and from IPA effectiveness database analysis. Distinctiveness is built through brand identity investment — distinctive assets, consistent visual and verbal presence, a recognisable point of view — none of which maps neatly onto funnel-stage content strategy.
What Replaces the Funnel
Several alternative models have been proposed. Google's messy middle describes the real-world buying process as an expansion-and-compression loop between exploration and evaluation, with buyers repeatedly broadening their consideration set and then narrowing it before committing. It is a better description of actual behaviour than the funnel but is less useful as a planning framework.
The flywheel model, popularised by HubSpot, replaces the funnel with a circular structure where existing customers generate new customers through advocacy. It correctly identifies the value of retention and referral, but risks becoming a platitude in contexts where the product's natural advocacy rate is low or the advocacy channels are informal and unmanageable.
The most useful reframing may be less a new model than a new set of priorities. Rather than asking "what funnel stage is this buyer in and what content should we serve them," the more productive questions are: How well known are we in the communities where our buyers spend time? How strong is our distinctiveness relative to competitors in the category? How actively are our existing customers recommending us in informal channels? How much of our marketing budget is going to the channels where our buyers actually form opinions, rather than the channels our analytics tools can most easily measure?
These questions do not produce a tidy attribution report. They require a mix of quantitative survey research — brand tracking, share of voice measurement, customer advocacy surveys — and qualitative insight about where influence actually flows in a given market. They are harder to answer than "what was our lead volume this quarter." They are also more likely to produce decisions that actually improve long-term growth.
Practical Implications for Marketing Teams
For a marketing team operating inside a funnel framework, the transition to a more accurate model of buyer behaviour does not require abandoning everything. Bottom-of-funnel tactics — search capture, retargeting, conversion optimisation — remain useful precisely because buyers who have already formed a strong prior view do use search and do click on retargeting ads. The problem is not that these tactics are ineffective. It is that they are credited with influence they did not create and are funded at levels that crowd out the investment in upstream influence that would make them more effective.
A more balanced approach allocates meaningful investment to brand-building activity — sponsorships, creator relationships, community participation, executive thought leadership, customer advocacy programmes — and accepts that this investment will not produce clean attribution data. It treats the absence of attribution as a property of the channel, not a signal of the channel's ineffectiveness. It measures brand health through surveys and share-of-voice tracking rather than lead volume alone. And it builds the internal case for this approach by connecting it to revenue outcomes over longer time horizons, where brand investment has been consistently shown to deliver disproportionate returns.
This is not a comfortable message for marketing teams whose budgets are justified through last-click conversion data. It requires having a more sophisticated conversation with finance and leadership about how buying decisions actually form. But the alternative — continuing to fund a model that everyone quietly knows is incomplete — is not a sustainable position either. The funnel's data shows what we can measure. What we cannot measure is doing most of the work. Any marketing strategy that ignores that reality is not a strategy. It is an accounting exercise.
The Longer View
The marketing funnel will not disappear as a concept. It is too embedded in the language of the profession, too useful as a rough taxonomy of buyer readiness, too familiar to the executives who sign marketing budgets. What will change, and is already changing in the teams that are producing the best results, is the degree of confidence placed in it as a description of how influence actually works.
The teams winning in the current environment are building audiences before they need them, investing in brand distinctiveness in ways that compound over time, and treating community and peer influence as primary channels rather than nice-to-haves. They are measuring what matters — recall, consideration share, customer advocacy rates, revenue attributed through surveys rather than last-click models — even when what matters is harder to measure than what the analytics dashboard offers.
The funnel was a useful simplification for a simpler information environment. The environment has changed. The model needs to change with it — not because simplification is wrong, but because the particular simplification the funnel represents now points marketing teams in the wrong direction more often than it points them in the right one. That is the cost of holding onto a model past its useful life, and in marketing, it shows up directly in the revenue line.