Orchestrating dark social into attributable pipeline

Your best deals started in a Slack thread you will never see. A VP forwards a LinkedIn post to her ops lead, who asks a Pavilion channel who they use. Three weeks later someone types your URL into a browser and books a demo, and HubSpot logs it as direct. When the board asks what's working, you point at the one channel your software can see.

That's the dark social problem at $20K+ ACV. Most of the buying happens where no pixel fires. You can still measure it and you can still orchestrate it, but not with four vendors each reporting on their own slice of the funnel.

Why HubSpot says "direct" when the buyer says "a friend told me"

The referrer data is often gone before the click lands. Visits from private communities and messaging apps can show up as direct. Forwarded emails strip UTMs. Enterprise buyers on corporate VPNs with ad blockers can drop out of GA4 entirely. Browsers, networks, and platforms often remove referral data even when tagging is correct.

Refine Labs compared HubSpot attribution against a mandatory free-text "How did you hear about us?" field. HubSpot credited web search with 78% of conversions, while buyers credited it with 12%. Dark social received roughly 0% of the software attribution but accounted for 98% of closed-won revenue in self-report.

Tracked attribution records which measurable touchpoints occurred before conversion, but it provides an incomplete account of how demand formed. Direct traffic may be the end of the path rather than the beginning.

Bigger deals push more of the journey into the dark

The higher your ACV, the more people research you before anyone raises a hand. Forrester's buying-network research found 73% of purchases involve three or more departments. Those people compare notes, forward material, ask peers, and build a shortlist before a seller enters the conversation.

That shortlist often forms in places your analytics cannot observe. Buyers ask trusted peers in private groups instead of beginning every evaluation with a category search. They may encounter a founder's post or hear a podcast recommendation. A colleague may also send a screenshot. The eventual website visit reveals little about those earlier interactions.

For technical products the pattern is sharper, and the venues are different. Developers ask in Discord and GitHub issues. They also use Reddit threads alongside LinkedIn. Even in a dark channel, the account is often identifiable. Community activity, repeat website visits, content engagement, and company-level intent can expose enough of the surrounding activity to support a relevant follow-up.

Recognize when demand is forming around an account and coordinate the next useful touchpoint without pretending the CRM captured the whole journey.

The hard part: self-reported attribution is messy

The free-text field is not a clean ledger. People may forget or compress a long journey into one answer. Others write something vague such as "Google." Optional fields attract fewer responses, while dropdowns force complex paths into categories chosen by the marketing team. Even useful answers describe what the buyer remembers, not necessarily what caused the purchase.

Owned channels can disappoint too. Some programs work and never show up in attribution; some do not work. A buyer may remember the podcast but forget the paid post that introduced it. Another may name a community even though repeated exposure to your content made the recommendation credible. A single field will not separate those effects.

Treat self-report as a directional signal that tells you where to look, then triangulate. Compare the answer with ad exposure, website activity, sales notes, account engagement, and pipeline progression. Paid and dark social feed each other; scoring them as rivals is the mistake.

Ask, "Which combination of messages and touchpoints keeps appearing in qualified opportunities?" That framing gives marketing and sales something they can act on without claiming causal precision the data cannot support.

Build the measurement stack

The stack captures what buyers remember and preserves that information through the CRM. It then compares the information with observed behavior.

• Make the field required, open-text, and high-intent. Ask "How did you hear about us?" on the demo or "book a call" form, not on a gated PDF. Skip the dropdown. Open text gives buyers room to name a person, community, event, podcast, post, or combination of sources. Keep the question simple so it does not compete with the main conversion.
• Map it carefully in the CRM. In HubSpot, store the answer as a dedicated single-line text contact property (hear_about_us) and copy it to a deal property with the "Edit records" workflow so it survives into pipeline reporting. In Salesforce, create a custom Lead field and map it to a custom Opportunity field during conversion. Keep it additive. It never overwrites tracked behavioral fields.
• Ask again on the discovery call. A live question can surface the details a short form answer misses: the specific community, person, event, podcast, or post involved. Give AEs a consistent prompt and a clear CRM field for the answer. If the form and call responses differ, preserve both rather than asking the rep to decide which one is correct.
• Report the two columns side by side. Put tracked attribution next to self-reported attribution by channel and quarter. The delta between them is your dark funnel gap. Review that gap by pipeline stage and include lead-volume reporting separately, so vague top-of-funnel responses do not carry the same weight as qualified and closed-won opportunities.
• Go to the board with metrics that don't depend on attribution. Report marketing-sourced and marketing-influenced pipeline separately and never sum them. Include blended CAC, calculated as all sales and marketing spend divided by new customers, and Pipeline MER, calculated as pipeline created divided by marketing spend. Add a qualitative summary of what prospects mentioned unprompted in sales conversations.

Together, these views show what the software observed, what buyers remembered, and whether the overall system produced efficient pipeline. That is more useful than forcing every deal into a single-source attribution model.

Orchestrate the channels that create the demand

Measurement without creation is just a better view of an empty room. The channels that fill the dark social layer are mostly the same ones that resist tracking, so they need a shared message and operating rhythm.

• Founder and exec LinkedIn. Post consistently about industry insight, customer problems, product decisions, and what the team is learning. Keep direct product asks in the mix without making every post a pitch. Consistency matters because buyers may encounter several posts before forwarding one internally or visiting the site directly.
• Thought Leader Ads. Paid amplification of a founder's organic post can extend its reach without making the message look like a conventional company ad. Start with posts that already express a clear point of view. Use paid distribution to reach the right accounts, then compare account engagement and pipeline movement with the self-reported answers entering the CRM.
• Signal-based outbound as the capture layer. When a company resolves in RB2B or Warmly after a founder post, that's dark social surfacing as an account you can name. Route it to an Instantly sequence that references the post, enrich it through Clay workflows, and stack it against signals that already matter: a recent CRO hire, a funding round, or a tech-stack change. The email should continue the topic rather than claim knowledge of an interaction you cannot verify.
• Budget split. Bias spend toward demand creation when the market still needs education. Put more toward capture when buyers already know the category and actively search for it. Pick the balance based on whether your category has a name, how much existing demand is available, and whether the current constraint is awareness or conversion.

The advantage comes from connecting these activities: a post establishes the point of view, paid media distributes it, intent signals identify engaged accounts, outbound continues the message, and CRM data reveals which themes reach pipeline.

Why this falls apart across four vendors

The stack above only works with one-team ownership. Separate specialists can each hit their own channel goals while the buyer receives disconnected messages. The underlying coordination risk is longstanding: an HBR agency warning notes that separate agencies "may succeed individually according to their own performance measures" yet their efforts "can directly undermine each other."

Dark social makes this worse. Your paid vendor reports clicks, your outbound vendor reports replies, your content freelancer reports impressions, and no dashboard captures the founder's LinkedIn post that actually started the deal. Nobody owns the hear_about_us field, and nobody reads the AE notes. As we wrote in our piece on faster go-to-market adjustments, each additional vendor adds coordination surface area faster than it adds pipeline.

The failure is operational and analytical. Without discovery-call insights, paid budget decisions lag. Without content-engagement context, outbound feels generic. Without pipeline feedback, creative keeps optimizing for engagement instead of qualified buyers.

Understory runs LinkedIn ads, Instantly-powered outbound on signal triggers, and on-staff creative for SaaS clients as one team. The person who drafted the founder post is the same person reading the self-reported field the next week and adjusting the outbound copy the week after. Plenty of teams, including Directive and Refine Labs, run paid media on their own, and we're glad to work alongside an existing partner when outbound and creative are the gap. We won't hand you a fifth dashboard.

Make dark social show up in your pipeline report with Understory

If your CRM says 60% direct and your reps say "they heard the podcast," your attribution system is missing the coordination between channels. Book an intro call and we'll walk through your current HubSpot or Salesforce setup, show you how we'd wire the self-reported field into deal reporting, and map which LinkedIn, outbound, and creative plays should sit on one team.