The Dark Funnel & Dark Social: What B2B Marketers Can Measure

A few weeks later, a company from that circle could have landed on our website as a “direct” visit, and we’d have had no idea why.

That’s the dark funnel in a nutshell. It’s every conversation, recommendation and bit of research that shapes a B2B buying decision but never shows up in your analytics.

As CMO of Lead Forensics, I’ve spent a lot of time trying to shine a light into it. We’ve made real progress. We now trace almost every inbound lead back through months of known touchpoints with that organization. But I’ll be honest: we still can’t see a big chunk of it, especially private social conversations. This post covers what the dark funnel and dark social are, why they matter, what we’ve done about them, and what other B2B teams are doing to close the gap.

What is the “dark funnel”? 

The dark funnel is all the B2B buying activity that happens before a prospect identifies themselves to you, and that your tracking tools can’t see. That includes research on review sites and third-party content, conversations with peers, the internal debates inside a buying committee, and anonymous visits to your website. The term caught on in B2B marketing to describe why form fills and demo requests show up so late in the real buying journey.

What is “dark social”?

Dark social is one part of the dark funnel. It’s the sharing and discussion of brands and content through private or untrackable channels: Slack and WhatsApp groups, DMs, email forwards, private communities, podcasts, and word of mouth.

The simplest way to separate them: all dark social is part of the dark funnel, but the dark funnel is much bigger than dark social.

Why these terms exist: the buying journey marketing software can’t see

Marketing software is built to record clicks, page views and form fills. It’s very good at telling you where someone interacted with you. It’s much worse at telling you where they first heard about you.

Refine Labs describes this gap neatly as the “attribution mirage”. A CMO asks peers in a Slack community for recommendations, gets three vendor names, then Googles them. Each vendor’s analytics credits organic search or a paid ad. The Slack thread that actually created the demand never appears anywhere.

I recognize that story immediately. It’s exactly what we saw with that mums’ group discussing our software. The recommendation happened in a space we couldn’t see, and the visit that followed looked like it came from nowhere.

How much of B2B buying now happens in the dark

Most of it. In a Gartner survey of 646 B2B buyers published in March 2026, 67% said they prefer a buying experience with no sales rep involved, and 45% had used AI during a recent purchase. Buyers are researching, comparing and vetting suppliers on their own terms, long before they’re ready to talk to anyone.

That matters because preferences form early. By the time someone fills out your form, they’ve often already decided who’s on their shortlist, and who’s at the top of it.

Earlier Gartner research shows where that time goes. In a survey of 750 B2B buyers, buyers spent just 17% of their buying time meeting suppliers. If three suppliers are in the running, each gets roughly 5%. The rest is independent research, peer conversations and internal alignment.

Peer communities, Slack groups, dark social shares, and direct/branded search

So where does that time go? From what I see in our own pipeline and in the research, the dark funnel tends to run through four places:

• Peer communities. Private LinkedIn and Facebook groups, Reddit threads, and niche professional communities. Buyers ask people they trust, because vendor content all starts to sound the same.
• Slack and Teams channels. Both industry communities and internal company channels, where a link gets dropped with “has anyone used this?”
• Dark social shares. Links forwarded by email, WhatsApp, text or DM. They arrive with no referrer data, so analytics files them as “direct.”
• Direct and branded search. Once someone hears your name, they type it into Google or straight into the address bar. That visit gets credited to search or direct, not to the conversation that caused it.

The last one is the sneaky part. Branded search and direct traffic often look like your healthiest channels, when they’re really the visible tip of demand created somewhere else.

Why this breaks last-click and even multi-touch attribution models

Last-click attribution gives all the credit to the final touch before conversion. In a dark-funnel world, that’s usually a branded search or a direct visit, the moment demand was captured, not the moment it was created.

Multi-touch attribution is better, but it can only spread credit across touchpoints it can track. If the most influential touch was a recommendation in a private Slack channel, no model will ever give it any credit, because it doesn’t exist in the data.

Refine Labs tested this by adding a mandatory free-text “How did you hear about us?” field to its main conversion form for 12 months. Across 620 conversions, its software attribution credited web search with 78% of conversions, but customers named web search only 12% of the time. Customers cited dark social sources, like social media, podcasts, word of mouth and community, for 85% of conversions (source). That’s one agency’s data, and buyers could name several sources, so I’d treat the exact numbers as directional. But the direction is hard to argue with.

The cost of ignoring the dark funnel

If you only measure what’s easy to see, you end up making decisions based on half the story. In my experience, that shows up in two expensive ways.

Misallocated budget from under-crediting top-of-funnel channels

When attribution only sees the last few clicks, the channels that create awareness look like they’re not working. Organic social, thought leadership, LinkedIn video and brand campaigns rarely get the final click, so they rarely get the credit. Paid search and retargeting look like heroes, because they’re standing at the door when the buyer finally walks in.

Here’s how that plays out in practice. The attribution report says most of last quarter’s best leads came from web search. The finance director looks at it and says to the CMO, “Great, I need you to get more of these.” It’s a completely reasonable ask. The problem is that the CMO may not know how to deliver it.

When a lead genuinely comes from a click you can measure, the answer is easy: you work out what drove it and do more of it. But if those searches were really prompted by a recommendation in a Slack channel or a private community, buying more search won’t create more leads. The search was only the last step. The thing that actually worked is hard to see, and that makes it very hard to repeat.

The natural reaction is to cut the “underperforming” top-of-funnel activity and move the money into capture channels. Then, a few quarters later, branded search volume starts to fall and nobody can work out why.

This matters even more because most of your market isn’t buying right now. Research by Professor John Dawes of the Ehrenberg-Bass Institute for the LinkedIn B2B Institute suggests only about 5% of B2B buyers are in-market at any given time (Marketing Week). The other 95% are forming opinions, often in the dark, that decide who makes the shortlist when they are ready.

Sales missing buying signals until a form is finally filled out

The second cost is timing. If your sales team only hears about an account when someone submits a form, they’re arriving late to a decision that’s mostly made. The buying committee has already done its research, compared options and probably picked a favorite.

Meanwhile, the signals were often there. A few people from the same company engaged with your LinkedIn posts. The company visited your website a few times without filling out a form. Someone clicked on your ads three times in a fortnight. A senior stakeholder opened every email in a nurture sequence. On their own, none of these look like much. Together, at an account level, they tell you a buying group is forming. If nobody connects the dots, that window closes.

How to bring dark-funnel activity into view

You’ll never light up the whole dark funnel. But you can shrink it a lot. Here’s how we’ve approached it at Lead Forensics, and what I’d recommend to any B2B marketing team.

Mapping every touchpoint at organization level

This is where our own process changed the most. A lead coming through to our website used to be something of a mystery. We often had no idea what had brought them to us.

Now, for every lead that comes to us direct, we look back over at least the previous three months, and often further, at every touchpoint we know we’ve had with that organization. Crucially, we look at it at organization level, not contact level. B2B purchases involve a wider decision-making unit, and the person who fills out the form isn’t always the one who’s been engaging with us.

So when a lead lands, we can typically see a trail like this:

1. Someone at the company engaged with an organic LinkedIn post.
2. A colleague clicked on one of our LinkedIn ads a few weeks later.
3. Another stakeholder opened and clicked on one of our emails.
4. The company visited key pages on our website several times, without converting.
5. Finally, someone came to us directly and asked to talk.

On its own, step five looks like a “direct” lead with no source. Mapped at organization level, it’s clearly the result of months of influence across several channels and several people.

This manual attribution work has genuinely changed how we invest. We can see which channels show up again and again in the journeys of accounts that convert, even when they never get the final click. That’s helped us defend and grow spend on activity, like organic social and LinkedIn ads, that last-click reporting would have under-credited.

Third-party intent data can add another layer: signals that an account is researching your category on other sites before it ever reaches yours. Used together, your own engagement history and intent data give you a much fuller picture of which accounts are in-market.

What you can (and can’t) realistically see, and setting the right expectations internally

I want to be honest about the limits here, because I think the industry oversells this.

Everything I’ve just described gets better at revealing what happened on channels we own or pay for. It doesn’t reveal the private conversations that kicked things off. We still can’t see what’s said in Slack channels, closed communities, Reddit threads or a group chat like the one where our software came up. These are often the spaces where the real recommendation happens, and they’re spaces where it’s hard for a brand to take part without being unwelcome.

Right now, we don’t have a framework for measuring social or private interactions. We’re working on it. In the meantime, I’ve found it helps to be really clear with leadership about three categories:

• What we can measure: engagement with our website, ads, emails and social posts.
• What we can infer: rises in branded search, direct traffic and unprompted inbound from a particular segment or region, which often point to word of mouth.
• What we can’t see: private conversations, and we say so plainly rather than pretending a model has it covered.

Setting that expectation early protects marketing from being judged only on what’s trackable, and it makes the case for investing in channels whose impact is real but partly invisible.

Turning dark-funnel visibility into action

Visibility is only useful if it changes what sales and marketing do next. These are the three places it has the biggest impact.

Giving sales the full story behind every lead

The most immediate win is context. Knowing that an organization has engaged with your posts, ads and emails, across several people and over several months, changes how the first conversation goes.

When sales get that context, outreach stops being cold. A rep can open with something relevant to what the company has actually been looking at, rather than a generic pitch. And because buyers do so much research before talking to anyone, getting to an engaged account a few weeks before it raises its hand can be the difference between shaping the shortlist and joining it late.

Building ABM target lists from dark-funnel research behavior

Static ABM lists built on firmographics alone tell you who could buy. Dark-funnel signals help tell you who’s looking.

We use engagement patterns, like repeat engagement, several people from one organization getting involved, and activity across more than one channel, to decide which accounts deserve more focus. That keeps ABM budget pointed at accounts showing real interest, while broader brand activity keeps working on the 95% who aren’t in-market yet.

Reporting dark-funnel-influenced pipeline to leadership

This is where the organization-level touchpoint review really earns its keep. Instead of reporting that a big slice of pipeline came from “direct,” we can show what those direct accounts had engaged with beforehand.

A few reporting habits that have helped us:

• Report influence alongside source. Show which channels appear in the journeys of converting accounts, not just which one got the final click.
• Track the right leading indicators. Branded search volume, direct traffic from target accounts and the number of engaged organizations per month are all useful signals of demand being created.
• Be upfront about the blind spot. Tell leadership which part of the journey is still unmeasured, and what you’re doing about it. It builds more trust than a model that claims to see everything.

Dark funnel vs. your attribution strategy: how they fit together

The dark funnel isn’t a reason to give up on attribution. It’s a reason to stop expecting one model to tell the whole story.

The framing I find most useful splits attribution into two jobs. Software attribution tells you where demand is captured: the clicks, visits and form fills you can track. Self-reported attribution tells you where demand is created: the podcast, the peer recommendation, the community thread. Put them together and you get hybrid attribution.

Our organization-level touchpoint review sits firmly on the capture side, and it’s made that side far more accurate than it was. The gap we still need to close is on the creation side, especially private social. Here’s what other B2B teams are doing about it, and what we’re looking at ourselves:

• Self-reported attribution. Add a required, free-text “How did you hear about us?” field to your highest-intent form, like a demo request. Refine Labs recommends free text over a dropdown, so buyers name the real source in their own words. It’s cheap, and it’s the most direct way to surface Slack groups, podcasts and word of mouth. The trade-off is that people misremember, and often name the most recent touch rather than the first.
• Sales discovery questions. Train reps to ask, early in the first call, where the buyer first came across you and who else they spoke to. Log the answer in the CRM so it can be reported on. A recent example brought this home for me. Someone in our sales team mentioned that a customer who had just bought from us had done extensive research in ChatGPT before getting in touch, and ChatGPT had pointed them to us as the right company for their needs. None of that would ever show up in our attribution. We only know because the buyer told a salesperson, and AI assistants are quickly becoming another dark channel.
• Win/loss and customer interviews. Ask recently closed customers how their buying group found and shortlisted you. These conversations often reveal communities and influencers you didn’t know mattered.
• Community and social listening. Monitor public forums, Reddit and LinkedIn comments for mentions of your brand and category. You won’t see private groups, but public conversations are often a good proxy for what’s being said behind closed doors.
• Showing up where buyers talk, the right way. Encourage employees and customers to contribute genuinely to communities they’re already part of, rather than marketing into them. You can’t track a recommendation in a private group, but you can make it more likely to happen.

None of these replaces the others. Our plan is to layer self-reported data on top of the organization-level touchpoint view we already have, so we can see both what influenced an account and where the conversation started. Dark social is something we, as B2B marketers, collectively need to crack, and I expect the teams that combine these methods will get there first.

FAQs

What is the dark funnel?

The dark funnel is the part of the B2B buying journey that happens out of sight of your tracking tools. It includes anonymous website visits, third-party research, peer recommendations, private community discussions and internal buying-committee conversations. Most of it happens before a buyer ever fills out a form or speaks to sales.

What is dark social?

Dark social is content sharing and brand discussion that happens through private channels analytics can’t track, like Slack, WhatsApp, DMs, email forwards and private communities. Traffic that comes from these shares usually shows up as “direct” in your analytics.

Is dark social the same as the dark funnel?

No. Dark social is one part of the dark funnel. The dark funnel also includes anonymous website visits, research on review sites and third-party content, and offline conversations. All dark social is dark funnel activity, but not all dark funnel activity is dark social.

Can you track dark social?

Not directly. Private conversations will always stay private. But you can measure its effects, through self-reported attribution, sales discovery questions, branded search trends and organization-level analysis of the accounts that come to you “direct.”

How do you measure the dark funnel in B2B?

Combine methods. Organization-level touchpoint analysis shows what each company engaged with across your ads, email, social and website. Self-reported attribution and customer interviews reveal where demand was created in the first place.