ooligo
ENTRY TYPE · definition

Multi-touch revenue attribution

Last updated 2026-05-02 RevOps

Multi-touch revenue attribution is the practice of distributing credit for closed revenue across the marketing and sales touches that contributed to it. In B2B, where 5 to 12 stakeholders interact with 20 to 60 touches over 6 months, single-touch models lie. The right model depends on what decision you are making — different attribution models answer different questions, and pretending one model rules them all is a mistake that costs marketing budget every year.

The major attribution models

ModelHow it creditsBest used for
First-touch100% to first interactionBrand and demand-gen budget
Last-touch100% to final interaction before oppConversion-rate optimization
LinearEqual share to every touchReporting parity, low signal
Time-decayWeight rises closer to closeMid-funnel investment decisions
U-shaped40% first, 40% last, 20% middleBalanced view, default for most B2B
W-shaped30% first, 30% MQL, 30% opp, 10% otherWhen you have clean stage data
Data-driven (algorithmic)ML-derived per-touch weightsMature stacks with clean data

No model is “true.” Attribution is a credit allocation policy, not a measurement of physical reality.

How to choose a model

Different decisions need different models:

  • “Should I spend more on this channel?” → Use a multi-touch model (U-shaped or W-shaped) so you don’t starve top-of-funnel.
  • “Is this campaign creating qualified pipeline?” → Use first-touch or sourced-pipeline.
  • “Is this nurture sequence accelerating deals?” → Use time-decay or stage-based.
  • “Should I cut this paid keyword?” → Use last-touch on opportunity creation.

The best practical setup is to track at least two models in parallel: a first-touch (for sourcing credit) and a U-shaped or W-shaped (for influence credit). Reporting both prevents single-model gaming.

How to instrument it

  1. One identity graph. People-level (not session-level) tracking, with company resolution. Without this, you are guessing.
  2. Touchpoints in one place. Web visits, content downloads, ad clicks, emails opened, demo requests, sales activities, all stitched to the lead and account.
  3. Closed-loop to revenue. CRM opportunities tagged back to the lead and account so closed-won amounts can flow to the touch history.
  4. Time bounding. Touches in the 12 months before opportunity create are typically considered; older touches are noise.
  5. Account-level rollup for ABM. Person-level attribution misses 70 percent of buying-committee activity in enterprise.

Common pitfalls

  • Single-model dogma. Reporting only first-touch starves nurture; only last-touch starves brand. Always run two.
  • Last-click web analytics on B2B. Google Ads or GA last-click tells you nothing about a 6-month, 8-stakeholder enterprise deal.
  • No account-level view. A buying committee with 5 people each having 6 touches looks like 30 disconnected leads without account stitching.
  • Comparing channel ROI across models. Each model produces different ROIs for the same channel. Pick a model per decision and stick with it.