Marketing attribution is the practice of assigning credit for revenue or pipeline to the marketing touchpoints that influenced a deal. In B2B, where buying committees touch six to ten channels over months, attribution is how RevOps decides where the next marketing dollar should go.
The models that matter
Five models cover almost every real-world implementation:
- First-touch. All credit to the first known interaction. Useful for top-of-funnel demand-gen evaluation. Blind to what closed the deal.
- Last-touch. All credit to the last interaction before opportunity creation or close. Overweights bottom-funnel channels (branded search, sales-led demos).
- Linear. Equal credit across every touchpoint. Simple, defensible, but treats a casual webinar attendee the same as a hand-raise.
- Time-decay. More credit to touches closer to the close date. Good default when sales cycles are short and the goal is influencing close.
- W-shaped or U-shaped. Heavy credit to first-touch, lead-creation, and opportunity-creation events. The most common multi-touch model in B2B because it maps to a real funnel.
Single-touch versus multi-touch
Single-touch is fast to implement and easy to argue about. Multi-touch is harder to instrument but matches reality: in B2B SaaS, the median deal has more than seven attributable touches. Most mature RevOps teams run two views: a multi-touch model for budget allocation across channels, and a last-touch model for individual rep and campaign performance.
How to instrument it
You need three layers of data: ad and campaign tracking with UTMs that survive redirects, a CDP or warehouse that joins anonymous touches to known leads after form fill, and a CRM where opportunity stages are clean. Without all three, multi-touch attribution is fiction. Tools like HubSpot and Salesforce ship native attribution; 6sense and similar platforms add intent and account-level resolution.
Common pitfalls
- Treating the model as truth. Attribution is a heuristic, not measurement. Use it to prioritize, not to declare winners. Run incrementality tests on big channels.
- Ignoring dark social. Podcast mentions, Slack groups, and word of mouth show up as direct or branded search. Ask “How did you hear about us?” on demo forms.
- Switching models mid-quarter. Different models produce different rankings. Pick one as the source of truth, document the rest as supporting views.
Related
- Pipeline velocity — the downstream metric attribution informs
- ICP — the segmentation lens for attribution
- HubSpot — common system of record for attribution