Marketing attribution is the practice of assigning credit for revenue to the marketing touchpoints that influenced a deal. It sounds straightforward. In practice, it's a source of endless analytical debate that rarely produces decision-relevant insights.
The attribution problem: every B2B deal has 10-20 touchpoints across email, ads, content, events, and direct outreach. Assigning fractional credit to each touchpoint requires assumptions that are arbitrary (first-touch, last-touch, linear, time-decay) and sensitive to attribution window choices. The resulting attribution model tells you more about which model you chose than about which marketing activities drove revenue.
The organizational cost of attribution obsession: months of analytics investment, political fights between marketing and sales over credit, and a false precision that misleads decision-making.
What to track instead:
Pipeline by channel with realistic sample methodology. Run "how did you hear about us" surveys on deals at discovery call. The answers are imperfect but directional, and they capture dark social signals that attribution models miss entirely.
Cohort analysis by acquisition channel. Do customers acquired through content have better retention than customers acquired through paid? If so, invest in content. This is the attribution question that matters for LTV, not just for the sale.
Marketing efficiency ratio by program. How much ARR did a specific campaign, event, or content investment influence? Use a generous attribution window (touches in last 90 days) and conservative methodology, but measure the programs you're actively evaluating, not all programs simultaneously.
Brand search volume trend. Rising branded search volume indicates growing awareness and trust. It's not attributable to a specific campaign, but it reflects the cumulative impact of all marketing activity.
Track what changes decisions. Stop tracking what generates reports.