Which marketing channels actually produce customers? Marketing attribution models try to answer that question by distributing credit across the touchpoints a buyer encountered. Choose the wrong model, and you may cut the channels that create demand while overfunding the ones that simply collect it.
Common marketing attribution models
| Model | How credit is assigned | Strength | Weakness |
|---|---|---|---|
| First touch | 100% to the first interaction | Highlights discovery channels | Ignores what closed the deal |
| Last touch | 100% to the final interaction before conversion | Simple, shows closing channels | Overvalues branded search and direct |
| Linear | Equal credit to all touchpoints | Recognises the full journey | Treats minor and major touches equally |
| Time decay | More credit to recent touches | Reflects proximity to decision | Undervalues early discovery |
| Position-based (U-shaped) | Most credit to first and lead-creation touches, rest shared | Balances discovery and conversion | Weights are arbitrary |
| W-shaped | Credit to first touch, lead creation and opportunity creation | Fits B2B pipeline stages | Requires good CRM data |
| Data-driven | Algorithm assigns credit based on observed paths | Adapts to your data | Needs volume; less transparent |
First touch vs last touch
First touch tells you how people discovered you; last touch tells you what they did right before converting. For B2B, both are incomplete. A buyer might discover you through a LinkedIn post, read three articles from search, attend a webinar, and finally book a demo after clicking a branded search ad. Last touch credits the ad; first touch credits LinkedIn. Both played a role.
Data-driven attribution
GA4 and Google Ads use data-driven attribution by default for many reports, distributing credit using machine learning based on conversion paths. It is useful but limited to touchpoints the platform can see, typically digital interactions on your website. It does not see podcasts, private messages, events or sales conversations.
Why B2B attribution is hard
- Long cycles: journeys span months, beyond cookie lifetimes.
- Multiple people: several stakeholders interact separately.
- Offline touches: events, calls and meetings are invisible to web analytics.
- Dark social: recommendations in private channels and communities leave no trail.
- Privacy changes: consent requirements and browser restrictions reduce tracking.
Self-reported attribution
Self-reported attribution asks buyers directly: "How did you hear about us?" Use an open text field (not only a dropdown) on demo and contact forms, and code answers monthly. It often reveals podcasts, LinkedIn posts, peer referrals and AI assistants that software attribution misses. Combine it with software data: software shows clicks and paths; self-reported data shows memorable influences.
A practical B2B attribution approach
- Capture source data on every lead: first-touch and last-touch UTM parameters, click IDs and landing page, stored in your CRM; see our UTM parameters guide.
- Use self-reported attribution on forms.
- Report pipeline by first touch and by lead-creation touch to show both discovery and conversion channels.
- Use platform data-driven reports for in-platform optimisation.
- Run incrementality tests where possible, such as pausing a channel in one region or holdout audiences for retargeting.
- Review trends quarterly, not daily, because B2B data is noisy.
Choosing a model for your business
| Situation | Practical approach |
|---|---|
| Short, simple journeys | Last touch plus channel-level trends |
| Mid-length B2B journeys | Position-based or W-shaped in CRM, plus self-reported |
| Long, complex enterprise deals | Account-level engagement, W-shaped, self-reported and sales input |
| Heavy paid media | Platform data-driven attribution plus incrementality tests |
Account-level attribution for B2B
Because several people from one company interact with your marketing, contact-level attribution can miss the bigger picture. Account-level views group all touches from people at the same company: the CFO who read a pricing guide, the manager who attended a webinar and the user who signed up for a trial. Many CRMs and analytics tools can associate contacts with accounts. Reviewing engagement by account helps sales prioritise and shows marketing which programmes reach buying committees, not just individuals.
Attribution and budget decisions
Attribution informs decisions; it does not make them. Before cutting a channel because it looks weak in last-touch reports, check first-touch, assisted conversions and self-reported data. Demand-creating channels such as content, social and podcasts often look weak in last-touch models. Our SaaS demand generation guide explains why.
Reporting attribution to leadership
Present a simple, consistent view: pipeline by first touch, by lead-creation touch and by self-reported source, with trends over time. Explain limitations openly. A marketing dashboard that shows these views side by side helps leaders make balanced decisions. Our field guide to a marketing measurement plan covers how to decide what to measure before launching campaigns.
If your attribution data is incomplete or inconsistent, our conversion and analytics team can connect website, CRM and ad platform data into reporting you can trust.
