Marketing Attribution Best Practices for B2B Revenue Teams
Most attribution models fail because they answer the wrong question. Here's how to build an attribution practice that actually changes how Marketing and Sales invest.
Marketing attribution gets a bad reputation because most implementations try to answer a question the data can't actually support: 'which touch caused this deal?' B2B deals have 20 to 40 touches across 6 to 18 months and 5 to 9 stakeholders. No model causally proves anything. The right goal isn't proof. It's a shared, defensible view of where revenue comes from, updated continuously, that Marketing and Sales both believe.
The four attribution models that matter
- First-touch: what surfaced the account. Best for top-of-funnel program evaluation.
- Last-touch: what converted the opportunity. Best for late-funnel conversion programs.
- Multi-touch (linear or U-shaped): credit distributed across the journey. Best for portfolio decisions.
- Sales-influenced: surfaces the human effort. Critical for honest pipeline conversations.
You don't pick one. You report all four. Each answers a different question and the gap between them is itself useful signal.
Best practices that actually move the needle
1. Anchor attribution to the Account, not the Lead
B2B revenue is account-level. Tracking attribution at the Lead level produces inflated, duplicate-counted numbers that fall apart the moment Sales asks a follow-up question. Roll touches up to the Account and report on Opportunity outcomes.
2. Define what counts as a 'touch' once
Web visit? Form fill? Webinar attendance? SDR email open? Pick a tiered definition (engaged / qualified / high-intent) and apply it identically across channels. Inconsistent touch definitions are the #1 reason attribution reports get challenged.
3. Make it native to your CRM
Attribution that lives in a separate platform creates a second source of truth that Sales will reject the moment a number disagrees. Native Salesforce or HubSpot attribution (where the model writes directly to Opportunity and Campaign objects) is the only model that survives a QBR.
4. Report against pipeline and revenue, not MQLs
MQL attribution measures Marketing's ability to generate leads. Pipeline and revenue attribution measure Marketing's ability to generate a business. Only one of those is interesting to the CFO.
5. Reconcile monthly with Sales-influenced view
Once a month, walk through the top 20 closed-won deals with Sales leadership and tag the human moments: the SDR outreach, the executive intro, the customer reference. Compare that to your multi-touch model. The gap is the conversation that makes attribution credible.
Case study: when a 'winning' channel was actually losing money
For a real-world example of why MQL- and ROAS-only attribution can mislead an entire executive team, see how MAM exposed the true cost of a Facebook campaign that looked like a top performer until sales effort and conversion data were layered in.
Need a hand putting this into practice?
We help revenue teams turn the ideas in these posts into working systems, usually in weeks, not quarters.