Sales Attribution Model Guide

Why attribution model choice matters

Every attribution model tells a different story about which activity "deserves" credit for a closed deal — and that story determines where marketing and sales leadership invest budget next quarter. Pick the wrong model for your motion and you'll systematically over- or under-fund the channel that's actually driving revenue.

The Models, Explained

First-Touch Attribution

100% of credit goes to the very first recorded interaction (e.g., the first ad click, first form fill, first cold call).

• Pros: Simple, shows what generates initial awareness/demand.
• Cons: Ignores everything that happens during the (often long) B2B sales cycle — undervalues sales and nurture effort entirely.
• Best for: Short-cycle, single-touch-heavy motions (e.g., PLG signup flows).

Last-Touch Attribution

100% of credit goes to the final interaction before conversion (e.g., the demo request, the last email before close).

• Pros: Simple, reflects what closed the deal in the moment.
• Cons: Ignores all the demand-generation work that built the pipeline in the first place — overvalues bottom-of-funnel activity.
• Best for: Very short cycles, or as a secondary sanity-check metric alongside a multi-touch model.

Linear Attribution (Multi-Touch)

Credit is split evenly across every touch in the journey. Credit per touch = 100% ÷ number of touches

• Pros: Acknowledges every touch mattered; simple to calculate.
• Cons: Treats a passive email open the same as a live demo — doesn't reflect that some touches matter more than others.
• Best for: Teams that want a fast, defensible multi-touch model without building a weighting scheme.

Time-Decay Attribution

Credit increases the closer a touch is to the conversion, using a decay formula (commonly a 7-day half-life). Credit(touch) ∝ 2^(-days_before_conversion / half_life), normalized so all touches sum to 100%

• Pros: Reflects that recent touches usually have more influence on the final decision.
• Cons: Still undervalues the top-of-funnel touch that originally created the opportunity.
• Best for: Longer cycles where late-stage touches (proposal, negotiation calls) are genuinely more decisive.

U-Shaped / Position-Based Attribution

Fixed weight to the first touch and the lead-conversion touch (the touch that turned a contact into a qualified lead), remainder split across the middle. Common split: 40% first touch / 40% lead-conversion touch / 20% split across everything else

• Pros: Credits both demand creation and qualification — the two moments most orgs care about most.
• Cons: Still undervalues middle-of-funnel nurture and, critically, undervalues everything that happens after the opportunity is created (i.e., sales' own work).
• Best for: Marketing-attribution reporting where MQL→SQL conversion is the key metric being managed.

W-Shaped Attribution

Fixed weight to three moments: first touch, lead-conversion touch, and opportunity-creation touch, remainder split across the rest. Common split: 30% / 30% / 30% across the three key moments / 10% split across everything else

• Pros: The first model on this list that gives explicit credit to the opportunity-creation moment — meaningful for orgs with a real SQL→Opportunity handoff.
• Cons: More complex to calculate and explain; needs clean stage-timestamp data.
• Best for: Enterprise motions with a defined MQL → SQL → Opportunity funnel and multiple influencing touches at each stage.

Full-Path / Custom Attribution

Extends W-shaped further to also weight post-opportunity touches (e.g., a champion-enablement call, a security review call) — often built as a fully custom weighting scheme per company.

• Pros: Most complete picture for complex, multi-stakeholder enterprise sales.
• Cons: Requires mature data infrastructure and ongoing model maintenance.
• Best for: Enterprise/ABM motions with long cycles and many post-opportunity touches from multiple stakeholders.

Algorithmic / Data-Driven Attribution

Uses statistical or ML modeling (e.g., Markov chains, Shapley value) on your own historical conversion data to assign credit based on actual observed influence, rather than a fixed rule.

• Pros: Most accurate in principle — lets the data decide what actually correlates with conversion.
• Cons: Needs a large volume of historical deal data and either in-house data science or a platform that runs it.
• Best for: High-volume, high-data-maturity organizations that have outgrown rule-based models.

Decision Tree — Which Model Fits Your Motion

`` START │ ├─ Is your sales cycle under 30 days and largely self-serve/PLG? │ YES → Use Last-Touch (or First-Touch as a secondary demand-gen metric) │ NO ↓ │ ├─ Do you need to credit marketing's role in creating AND qualifying the lead, │ with a simple, explainable model? │ YES → Use U-Shaped (40/40/20) │ NO ↓ │ ├─ Do you have a clean MQL → SQL → Opportunity handoff with timestamped stages, │ and want to credit the opportunity-creation moment explicitly? │ YES → Use W-Shaped (30/30/30/10) │ NO ↓ │ ├─ Is your motion enterprise/ABM with multiple stakeholders touching the deal │ AFTER the opportunity is created (security review, exec sponsor calls, etc.)? │ YES → Use Full-Path / Custom weighting │ NO ↓ │ ├─ Do you have high deal volume, clean historical data, and either in-house │ data science or a platform that runs data-driven attribution? │ YES → Use Algorithmic / Data-Driven Attribution │ NO → Default to Linear or Time-Decay as the most defensible simple model │ available without heavier infrastructure ``

Worked Example

A deal with 5 recorded touches before close:

1. Paid ad click (Day 0)
2. Webinar registration → becomes an MQL (Day 12)
3. SDR outbound call → becomes an SQL (Day 20)
4. Opportunity created after discovery call (Day 35)
5. Final negotiation call (Day 58, deal closes Day 60)

Notice how differently each model treats Touch 4 (opportunity creation) — 20% under Linear, only 6.7% under U-shaped, but a full 30% under W-shaped. If your leadership team cares about crediting sales' opportunity-creation work specifically, U-shaped will systematically make that work invisible in your reporting.