A hard-won lesson from rebuilding attribution from the ground up
When I inherited our HubSpot instance, I thought our attribution was working fine. After all, leads were coming in, MQLs were being scored, and SQLs were flowing to Salesforce. What could be wrong?
Everything, as it turned out.
The Attribution Nightmare Most Companies Are Living
Here’s the brutal reality I discovered: For years, businesses have relied on flawed attribution models that only capture part of the picture. Our setup looked sophisticated on the surface—LinkedIn ads, ABM through 6sense, Google search campaigns, all feeding into HubSpot with proper scoring. MQLs went to Salesforce, sales validated them as SQLs, and life went on.
But something felt fundamentally wrong.
When leadership asked, “What’s driving our best leads?” or “Should we increase our LinkedIn ad spend?” I realized I was essentially guessing. We had touchpoint data, but no real understanding of what was actually working.
The breaking point came when I audited what happened to rejected MQLs. I assumed our nurture sequences were capturing these leads and moving them through a sophisticated funnel.
I was dead wrong.
Those leads—sometimes 30-50% of our monthly volume—were just sitting there. Stagnant. No nurture, no follow-up, no second chances. We were hemorrhaging opportunity and had no idea.
The Fatal Flaw in Traditional Attribution Models
Most attribution models assume linear progression: first touch → middle touches → conversion. But that’s not how B2B buyers actually behave, especially in complex sales cycles.
Here’s what I learned analyzing our data: People downloading educational content (guides, how-tos, industry insights) are in a completely different mindset than those downloading solution sheets. Yet traditional attribution models treat them exactly the same.
The educational content downloaders weren’t “failed leads”—they were early-stage researchers who needed nurturing, not selling. Some took 6+ months to become sales-ready, but when they did, they often had higher close rates because they were more educated about their challenges.
Traditional first-touch, last-touch, and even basic multi-touch models completely miss this nuance. They either over-credit awareness activities or over-credit conversion activities, but they don’t account for the messy, non-linear reality of how modern B2B buyers actually research and buy.
The "If It Ain't Broke, Don't Fix It" Trap
I’ve written about this mindset before, and it will kill your attribution efforts: “If it’s not broke, don’t fix it.“This is the worst approach anyone can take in 2025, especially with AI transforming how buyers research and engage with vendors.
The market is constantly shifting—political influences, demographic changes, firmographic evolution. In the healthcare space, regulatory changes affecting payers and providers happen monthly, value-based care models are evolving, and health services buying patterns are shifting. Your attribution model needs to be sophisticated enough to adapt to these changes, not rigid enough to break under pressure.
When I inherited our HubSpot instance, I initially continued with what appeared to be working. But that’s exactly the problem—you can get left behind when the market changes. Your model shouldn’t change every time there’s a significant shift, but it should be robust enough to support those changes.
Here’s the hard truth: If you can’t confidently answer how each marketing activity contributes to revenue, your attribution isn’t working—it’s just collecting data.
What I Built Instead: A Custom Multi-Touch Attribution Model
Most attribution models assume linear progression: first touch → middle touches → conversion. But that’s not how B2B buyers actually behave, especially in complex sales cycles.
Here’s what I learned analyzing our data: People downloading educational content (guides, how-tos, industry insights) are in a completely different mindset than those downloading solution sheets. Yet traditional attribution models treat them exactly the same.
After months of rebuilding, here’s the framework I developed:
The “Highway Enhancement” Attribution Model
Core Components:
- First Touch Attribution (40%): The campaign that introduced the prospect
- Highway Progression Influence (30%): Content and activities that moved them through our “highway” (nurture track)
- Final Stage Attribution (20%): The campaign that triggered conversion
- ABM/Intent Intelligence (10%): Account-level signals from 6sense and other tools
Why This Works:
Credits the full journey while weighting key moments appropriately
Accounts for educational vs. solution content with different attribution weights
Includes account intelligence that traditional models ignore
Measures “highway lift” – the value of proper nurturing vs. direct conversion attempts
Example Attribution Chain:
- Dr. Johnson (provider decision-maker) discovers us through “Cost Containment Blog” (40% attribution)
- Downloads payer case study, attends value-based care webinar, engages with emails (30% attribution)
- Requests demo after 6sense shows high intent from his health system (20% final touch + 10% intent data)
This gives us a complete picture of what’s actually driving conversions—and more importantly, what activities are worth scaling.
My #1 Recommendation: Start With a Brutal Audit
Before you touch any attribution models, you need to understand where you’re bleeding leads and revenue.