Why attribution is broken in B2B SaaS
B2B SaaS sales cycles are long (3-18 months), multi-stakeholder (4-7 buyers per deal), and increasingly anonymous (dark social, community discussions, internal Slack). Traditional attribution, last-click, first-click, linear, captures less than half of what actually influenced the deal.
Most agencies optimize to the metrics they can attribute: MQLs from paid channels, signups from tracked campaigns. These metrics look good in monthly reports and hide the real influence channels (content read on a colleague's recommendation, podcasts listened to during commutes, newsletter mentions forwarded internally).
The result: agencies over-invest in trackable channels and under-invest in the channels that actually move pipeline. That's how you get rising ad spend with declining pipeline velocity.
Our multi-layer approach
We measure pipeline attribution across three layers:
Layer 1: Last-click + multi-touch. Standard attribution via HubSpot, Salesforce, or purpose-built tools (Dreamdata, Heap, Factors.ai). We capture what's capturable, even knowing it's incomplete.
Layer 2: Self-reported attribution. Every demo form, trial signup, and sales call asks 'How did you hear about us?' with structured options. This captures 20-40% of influence that doesn't show in click-tracking.
Layer 3: Leading indicators. Branded search volume, direct traffic, newsletter subscriber growth, community mentions, share-of-voice in key subreddits. These lag-free signals tell us if demand is building before pipeline converts.
The channels we can't attribute (and why we run them anyway)
Some of our best-performing channels have the worst attribution: newsletter ads, podcast sponsorships, Reddit presence, community participation, organic social. Last-click attribution gives these channels zero credit. Multi-touch gives them a fraction.
We still run them because the leading indicators tell us they work. Branded search volume goes up. Direct traffic climbs. Demo form self-reports increase. These signals precede attributed pipeline by weeks or months, but the relationship is reliable.
Agencies that optimize purely to attribution cut these channels prematurely. That's short-term optimization that compounds into long-term pipeline decline.
How we report
Monthly reports include: attributed pipeline (standard), self-reported attribution (layer 2), and leading indicators (layer 3). Plus commentary on which channels are compounding and which are plateauing.
We don't hide unreported performance. If a campaign didn't drive attributable pipeline, we say so, and explain why we're continuing or stopping it based on leading indicators.
Honest reporting is the baseline. Most agencies hide misses; we surface them. That's how you build a marketing function that actually compounds.