Before they ever fill in your form. The buyer journey has moved upstream. If your marketing only starts working when someone becomes a lead, you've probably started too late.
For years, B2B marketing has been built around a comforting sequence: see an ad, visit the website, download something, become a lead, speak to sales, buy.
SEE AN AD → VISIT THE WEBSITE → DOWNLOAD SOMETHING → BECOME A LEAD → SPEAK TO SALES → BUY
So we measure what we can see. Clicks. Sessions. Form fills. MQLs. Cost per lead. Pipeline.
The problem is that your buyer doesn't experience the journey in that order. They search Google. They ask ChatGPT. They read Reddit. They watch a product demo. They ask a colleague. They compare three suppliers. They look at reviews. They read a competitor's pricing page. They search your brand name. They disappear for six weeks. Then, eventually, someone fills in your form.
And your CRM tells you: lead created. It doesn't tell you the decision may have been forming for months.
Research from 6sense backs this up. In its latest study, buyers contacted vendors around 61% of the way through their journey, down from 69% previously. More tellingly, 94% said they'd already ranked their preferred vendors before speaking to a seller. The vendor ranked first at that early stage went on to win the deal 77% of the time.
Think about what that means. By the time a prospect appears in your CRM, they may already have defined the problem, researched the category, shortlisted vendors, compared your proposition, formed an opinion of your brand, discussed the purchase internally, and decided who they'd prefer to buy from.
Your beautifully optimised lead-gen campaign may simply be capturing the moment they finally raise a hand. It isn't necessarily creating the preference. And that's a very different job.
The buyer's journey is a mess. A typical B2B buyer might Google “best CRM for a 200-person sales team”, see your brand three times on LinkedIn without clicking, ask an AI assistant to compare you with two competitors, check a rival's pricing page, ask a colleague if anyone's used you before, search your brand name, read a case study, check what people say about you on a review site, ask AI for one more comparison, and only then reach sales.
GOOGLE → LINKEDIN (NO CLICK) → AI SEARCH → COMPETITOR SITE → INTERNAL CHAT → BRANDED SEARCH → YOUR WEBSITE → REVIEW SITE → AI SEARCH → SALES
There’s no neat linear path. No reliable “first touch → last touch” story. And a frightening amount of it happens outside the places your marketing dashboard can comfortably see.
That's the dark funnel: AI assistants, private Slack and Teams threads, WhatsApp groups, Reddit and communities, peer recommendations, internal procurement chats, untracked links, word of mouth, branded search. None of it produces a clean conversion event. All of it can shape the one that eventually gets recorded.
This matters more now that LLMs sit inside the research process. 6sense found that 94% of B2B buyers used an LLM during their buying journey in 2025, with usage strongest in the middle: comparing vendors, synthesising information. The question isn't “who should I buy from?” It's “what's the difference between these three?”, “what should I look for?”, “is this company credible?”, “what are the risks?” Those questions happen long before the lead form.
If the job isn't simply to generate leads, what is it? To create preference before the buyer is ready to talk. That's four jobs.
Be present when the buyer first realises they have a problem, whether that’s Google, LinkedIn, SEO or an AI-generated answer. The channel matters less than the question: were you there when the problem first appeared?
Being known isn't enough. Buyers build a shortlist, and that's where the real competition happens. Your website, case studies and brand all need to answer one question: why you? Not with corporate waffle. With evidence.
Your buyer is almost certainly comparing you with someone else, so make it easy. What do you do differently? Who are you best for? What does it cost? Answer honestly and you don't just look transparent, you make their research easier, and give the AI they're using something real to summarise.
Only now does the form fill get interesting. A lead built on existing preference isn't the same as one that's just discovered you. One says “tell me what you do.” The other says “I think you're the right option, let's talk.” Treat them the same and your reporting stays unhelpful.
If your dashboard says LinkedIn cost £182 per lead, Google £74 and organic £0, you haven't learned very much. You've learned what happened at the point of conversion. Not what created the demand.
Imagine this: a buyer sees your LinkedIn campaign three times without clicking. Two weeks later they search your brand and read a case study. A month later they ask an AI assistant about your company. Three weeks after that, they search your brand again, then submit a form through organic search.
DASHBOARD SAYS: ORGANIC SEARCH → 1 LEAD
ACTUALLY HAPPENED: LINKEDIN + BRAND + CONTENT + AI RESEARCH → 1 LEAD
Channel-level CPA becomes actively misleading when it’s treated as the whole story. The question isn’t “which channel generated the lead?” It’s “which channels and experiences helped create the buyer’s preference?”
This is one of the easiest mistakes to make. You run LinkedIn for three months. Clicks are expensive, direct conversions are low. Someone asks why you're spending £8,000 a month on it. So you turn it off.
Six weeks later, branded search declines. Organic traffic declines. Demo requests decline. Sales says the pipeline feels weaker, and nobody can quite prove why. The original interaction was never the conversion. It was memory. Familiarity. Credibility. It was being one of the companies the buyer remembered when the buying process eventually started.
That doesn't mean every awareness campaign works. Far from it. It means we need to stop demanding that every marketing interaction behave like a bottom-of-funnel conversion campaign.
Not “ignore leads.” Just stop pretending they’re the whole picture. We’d build measurement around four layers.
| Layer | Question | What to track |
|---|---|---|
| Demand | Are more people looking for the problem, category or brand? | Search volume, brand search, website demand, audience growth |
| Preference | Are buyers choosing to engage with you? | Engaged visits, returning visitors, case study engagement, direct traffic |
| Pipeline | Does that preference turn into commercial opportunity? | Qualified leads, opportunities, pipeline value, win rate |
| Revenue | Did it actually make money? | Won revenue, CAC, gross margin, lifetime value, payback period |
That gives a far more useful picture than “Google generated 47 leads this month.”
The old model was: generate the lead. The better model is: influence the shortlist. Because if buyers are forming opinions before they contact you, your marketing has to work before the form appears. That means:
Ask your marketing team this: if a prospect researched us anonymously for three months before filling in our form, would our reporting tell us we’d influenced them?
If the answer is no, you're not necessarily doing bad marketing. You might just be measuring the wrong part of the journey.
B2B buyers haven't stopped talking to sales. They've simply moved more of the decision-making before they do. AI hasn't created this behaviour. It's accelerated it, giving buyers another way to research, compare and validate vendors without booking a meeting. 6sense's latest research suggests buyers are now engaging sellers earlier than before, but they still arrive at that conversation with a heavily formed shortlist.
So don't build your marketing around the moment someone becomes a lead. Build it around the moment they start deciding. Because by the time they fill in your form, you might already have won. Or lost.
This is what giving a damn looks like in practice. Tell us what's not working.