Attribution arguments often sound mathematical but are usually strategic. One team wants to know what creates new demand. Another wants to know what captures buyers who are ready to act. Finance wants a defensible connection to revenue. A single model cannot answer all three questions equally well.
Start With the Decision
If the decision is where to invest in awareness, first-touch is useful. If the decision is which conversion paths close demand, last-touch provides a clear operational view. If you need to understand the full journey, a multi-touch model is more appropriate. Write the decision at the top of the report so readers know what the model is designed to explain.
What the Common Models Reveal
- First-touch: gives all credit to the earliest known interaction. Good for demand creation; blind to nurture and conversion work.
- Last-touch: credits the final interaction before conversion. Easy to explain; tends to overvalue branded search, direct traffic, and retargeting.
- Linear: divides credit evenly. Useful as a neutral baseline; assumes every touch mattered equally.
- Time-decay: gives more weight to recent touches. Helpful in shorter cycles; can understate foundational education.
- U-shaped: emphasizes first touch and lead creation. Useful when acquisition and conversion are the clearest milestones.
- W-shaped: emphasizes first touch, lead creation, and opportunity creation. Better suited to B2B journeys with a meaningful sales handoff.
Match the Model to the Sales Cycle
A short self-serve purchase may be represented adequately by last-touch plus first-touch context. A six-month B2B sale involving several stakeholders needs milestone-based analysis. The longer the cycle and the larger the buying group, the less believable a single-touch answer becomes.
Show first-touch, opportunity-touch, and revenue attribution side by side. When the stories differ, the difference is often the insight.
Check Whether Your Data Can Support the Model
Complexity does not repair missing data. Before using multi-touch attribution, measure the percentage of deals with known contacts, captured sources, reliable campaign values, and complete lifecycle timestamps. If half the journey is invisible, publish a confidence note and use a simpler model until collection improves.
Keep Self-Reported Attribution
Add “How did you hear about us?” to the conversion process and preserve the answer exactly as submitted. Buyers often remember podcasts, communities, referrals, and people that tracking cannot observe. Compare self-reported and software-reported sources instead of forcing one to replace the other.
Choose Explainability Over False Precision
A model should be documented in plain language, recalculable from source data, and stable enough to compare over time. If leadership cannot understand why a channel received credit, the result will not guide a budget decision. Start simple, compare models, and increase complexity only when it changes an action.
Reference: HubSpot documents contact, deal, and revenue attribution as distinct reporting questions.