Social media lead scoring transforms LinkedIn engagement into revenue signals your team can act on. At Kainos, integrating Oktopost with Marketo produced a 581% increase in ROI and a 554% increase in reach. Charlie Oakham, who was Digital Strategy Director at Kainos at the time, described the result plainly: “The social engagement data we link using Oktopost and Marketo enables us to track, report, and prioritize leads for our sales teams and provides our sales partners with holistic lead-generation insights.”
That outcome should not be remarkable. In a world where marketing technology is supposed to give revenue teams complete visibility into buyer behaviour, connecting social engagement to a social media lead scoring model sounds like table stakes. And yet most B2B marketing teams cannot do it. Not because the technology is unavailable, but because their data architecture was never designed to accommodate it.
That gap is what this post is about.
The lead scoring models most companies built for 2015
B2B lead scoring was largely codified during a period when the buyer’s research journey was more legible than it is today. Buyers attended webinars hosted by vendors. They downloaded gated PDFs. They visited product pages, filled out contact forms, and registered for demos. Every meaningful interaction left a trackable footprint in the marketing automation platform. Scoring models were built to recognise those footprints and translate them into a number that indicated purchase intent.
The world changed. Buyers didn’t stop researching. They changed where they do it.
According to Gartner, 67% of B2B buyers now prefer a rep-free buying experience, and buyers spend just 17% of their total purchase journey in direct contact with potential vendors. The remaining research happens on social channels, through peer networks, and via employee-shared thought leadership content, none of which generates a form submission or a trackable website visit.
This is what researchers now call dark social: buyer interactions with your brand that happen outside the trackable web. A prospect reads a LinkedIn post from one of your sales engineers and spends ten minutes reviewing their company page. Another prospect sees a reshare from a customer advocate and clicks through to a product page, but without a UTM parameter on the link, that visit arrives in your analytics as direct traffic. A third prospect follows your company page after watching a webinar clip shared by an employee. None of these interactions registers in your marketing automation platform. None of them influences the lead score.
The Edelman-LinkedIn B2B Thought Leadership Impact Report found that 58% of B2B decision-makers spend one or more hours per week consuming thought leadership content, and that well-executed thought leadership is more effective at influencing purchase consideration than traditional product marketing. If the most persuasive content your company produces is being read by active buyers and generating zero scoring signal, the model is not giving you an accurate picture of intent.
Employee advocacy compounds the problem. When a company-approved post is reshared by 50 employees, each share extends the distribution of that content into professional networks that the brand does not directly own. Those networks contain buyers at every stage of the purchase journey. But unless the shared post is UTM-tagged and the clicks are resolved back to known contact records, the engagement is invisible. The lead scoring model treats a prospect who has seen your brand content 15 times through employee shares the same as a prospect who has never encountered it.
This is the attribution blind spot that makes traditional lead scoring increasingly unreliable as a predictor of pipeline. The model works exactly as designed. Social media lead scoring was built to close this gap by bringing those channels into the scoring picture. The problem is that the design predates the channels where a substantial portion of modern B2B research now happens.
“Integrating our martech stack of Oktopost and Marketo has closed the knowledge and insights gap in our organic social efforts. The social engagement data we link using Oktopost and Marketo enables us to track, report, and prioritize leads for our sales teams and provides our sales partners with holistic lead-generation insights.”
Charlie Oakham, former Digital Strategy Director, Kainos
The Kainos result is significant not just because of the numbers, but because of what it represents architecturally. Kainos did not rebuild their social media lead scoring model from scratch. They extended it by giving Marketo access to social engagement data it had previously never seen. The 581% ROI increase reflects what happens when a scoring model starts working with a more complete picture of buyer behaviour. The inputs changed. The output improved accordingly.
Why social signals get left out of lead scoring
The omission is not accidental. There are three structural reasons why social engagement data fails to reach marketing automation platforms, and each one requires a deliberate fix.
The first is UTM coverage. Most social media management tools apply UTM parameters to links in scheduled posts, but employee advocacy shares often bypass that tagging entirely. When an employee copies a link and shares it manually on LinkedIn, no UTM parameter travels with it. When a prospect clicks that link and arrives on your website, the session is attributed to direct. The contact record in Marketo or HubSpot receives no social touch. This is a publishing workflow problem, not a reporting problem. It has to be solved at the point of content creation and distribution.
The second is identity resolution. Even when UTM parameters are present, the click from a social post creates an anonymous session in your web analytics. Connecting that session to a named contact record in your marketing automation platform requires a matching mechanism: an email address, a cookie that ties the anonymous session to a known contact, or an integration layer that handles the resolution automatically. Without that step, the click is counted as traffic but never reaches the lead score.
The third is engagement outside your own web properties. A prospect who comments on a LinkedIn post published by your company has signalled interest. A prospect who reshares that post has signalled something stronger. But those interactions happen on the social platform, not on your website, and they produce no session data, no form fill, and no trackable event in your MAP. Unless your social platform can detect those interactions and link them to contact records in Marketo or HubSpot, they contribute nothing to lead scoring.
All three problems are solvable. Social media lead scoring is the outcome when all three are in place. Solving them requires a social media management platform built to connect with marketing automation at the data layer, not just at the campaign reporting layer.
What social scoring signals are worth tracking
Before configuring anything in Marketo or HubSpot, it is worth establishing which social signals carry enough intent to justify a scoring increment. Not all social engagement is equal. A prospect who clicks through from an employee advocacy post has taken a deliberate action toward your content. A prospect who follows your company page has opted into ongoing exposure to your brand. These are different signals and they should carry different weights.
The following values are a proven starting point, calibrated against the scoring weights commonly used for other mid-funnel marketing automation signals:
- Followed the company’s social page: +5
- Clicked a link from an employee advocacy post: +10
- Commented on an employee advocacy post: +15
- Reshared a social post from the company: +3
A comment on an employee advocacy post scores higher than a click because it represents active, visible engagement. The prospect has identified themselves publicly as interested in your content. A reshare is scored lower than a click because resharing is a lighter action and does not necessarily indicate personal buying intent. Following the company page sits between the two: it is a considered signal, but a passive one.
Review these weights against your existing scoring model. If an email open scores +5 in your current model and a whitepaper download scores +20, calibrate social signals on the same scale. The goal is a scoring architecture where social engagement sits in its appropriate position relative to other signals, not a separate track that operates in isolation from the rest of the model.
Connecting social engagement to Marketo lead scoring
The integration between Oktopost and Marketo creates a data pipeline that resolves social engagement back to the contact record. When a contact clicks a link from an Oktopost-managed post or advocacy share, with Oktopost you can link that social activity to the corresponding record in Marketo. From that point, standard Marketo scoring programs can treat the social touch identically to any other scored activity.
The flow works in five steps. Oktopost applies UTM parameters automatically to all links in published posts and advocacy shares. When a tagged link is clicked, Marketo’s Munchkin script resolves the anonymous session to the contact record where a match exists. Oktopost then links the social engagement event to the contact record in Marketo as a custom activity, visible in the contact’s activity log alongside email opens and form fills. A Marketo Smart Campaign listens for the custom social activity and fires the corresponding score change. When the accumulated score crosses the MQL threshold, Marketo syncs the contact to Salesforce and alerts the sales team, carrying the full social activity history with the record.
Setting up the Smart Campaign requires four components: a trigger filtering for Oktopost social activity types, a score change action using the values above, a suppression filter that prevents contacts already past the MQL threshold from receiving redundant increments, and a decay rule that applies a negative score to contacts whose social engagement has been dormant for 90 days. The decay rule keeps the model current as prospects move in and out of active research phases.
The reason to route social activity through the standard Marketo scoring infrastructure, rather than building a parallel social score, is that it preserves the integrity of the MQL definition. Sales and marketing alignment depends on a shared understanding of what a qualified lead looks like. Adding social signals into the existing model extends that definition rather than creating a competing one.
Connecting social engagement to HubSpot lead scoring
The HubSpot integration follows the same principles with a different configuration approach. With Oktopost you can link social engagement activity to the contact record in HubSpot as a timeline activity and custom property update. A HubSpot workflow then reads that property and adjusts the contact’s lead score accordingly.
The workflow branches by activity type, applying different score increments to clicks, comments, and follows. The trigger fires on contact property change when a social activity is linked to the record. Where the score threshold is met, the workflow triggers the MQL notification to the sales team. HubSpot’s predictive lead scoring, available at higher tiers, incorporates the social activity data automatically once it is present in the contact record.
A practical note on the UTM-only limitation that affects teams using HubSpot and other MAPs: relying entirely on UTM parameters and web session data captures clicks on tracked links but misses every social interaction that happens on-platform. Comments, follows, reshares, and views of posts with no outbound link generate no web session and therefore no contact-level event. For companies running employee advocacy programs, the on-platform engagement volume is often larger than the click-through volume. A UTM-only model undercounts social’s contribution to pipeline by design.
Measuring social-influenced pipeline in Salesforce
Lead scoring captures the moment of qualification. Pipeline attribution answers the broader question: which social engagements influenced deals that eventually closed?
Once social activity is linked to contact records in Marketo or HubSpot and synced to Salesforce, the social-influenced pipeline report uses a 90-day attribution window. Any social engagement logged against a contact in the 90 days prior to opportunity creation counts as a social touch on that opportunity. The report aggregates the pipeline value of all opportunities where at least one contact carries a social touch within the window, using contact roles to connect contact-level activity to opportunity-level pipeline.
The comparison between socially-influenced and non-influenced opportunities is where the revenue argument for social investment is made or lost. If socially-influenced opportunities close at a higher rate or carry a larger average deal size, the data supports increased investment in social content and employee advocacy. If the difference is negligible, the model needs refinement: either the scoring signals are not discriminating enough, or the social content is not reaching the right accounts.
The 90-day window is a reasonable starting point for most B2B sales cycles. Teams with longer average cycle lengths should extend the window proportionally. Enterprise teams with 12-plus month sales cycles may find that a 180-day window captures substantially more of the social influence that occurs during evaluation.
The data architecture argument
I want to be direct about where the real problem sits, because I see it misdiagnosed regularly in conversations with B2B marketing teams.
The problem is not the lead scoring model. Most teams have built sensible models that reflect the signals their marketing automation platform is capable of seeing. The model is doing its job. The problem is that a significant category of buying signal, social engagement, sits outside the data perimeter that the model can access.
Fixing this is a data architecture decision, not a scoring calibration decision. It requires a social media management platform built to link engagement data back to contact records in Marketo, HubSpot, or Salesforce. It requires UTM coverage at the advocacy layer, not just the brand publishing layer. It requires a mechanism for resolving on-platform social interactions to known contacts, so that a comment or a follow generates a scored activity in the MAP rather than just a metric in the social dashboard.
When those data pipes are in place, the social media lead scoring model does not need to change. It simply starts receiving better inputs. And better inputs produce better outputs: more accurate MQL scores, higher conversion rates from MQL to SQL, and a pipeline attribution model that reflects what your buyers are actually doing, not just the subset of their behaviour that happens to leave a form fill.
The Kainos result is a data point from a team that built those pipes. A 581% ROI increase is not a marketing technology story. It is what happens when revenue systems are designed to see the full scope of buyer engagement rather than a filtered version of it.
The question is no longer whether social engagement influences pipeline. The question is whether your revenue systems are capable of recognising those signals.
