Marketing Attribution Models
Which One Is Right for B2B?

What Is a Marketing Attribution Model?

A marketing attribution model is a framework that assigns credit to the marketing touchpoints along a buyer's journey, helping teams understand which activities contributed to a conversion or closed deal.

In B2B, the buyer journey is rarely linear. Someone might find you through a blog post, attend a webinar three months later, read a case study the week before a demo, and then receive a follow-up sequence from sales before signing. Which of those touchpoints deserves credit for the revenue?

That's the question attribution models attempt to answer.

The full context on how attribution works in practice — including how to connect it to your sales process and CRM — is covered in our guide to B2B marketing attribution. This guide focuses on the models themselves: what they are, how they distribute credit, and when each one makes sense.

Single-Touch Attribution: First Touch and Last Touch

Single-touch attribution is the simplest place to start. You assign 100% of the conversion credit to one touchpoint — either the first interaction a prospect had with your brand, or the last one before they converted.

One touchpoint. All the credit.

For teams just getting into attribution, or those without much data infrastructure, single-touch models are genuinely useful. Easy to implement, easy to report on, minimal technical lift. The trade-off is they only ever tell part of the story.

First Touch Attribution

First touch gives all the credit to whatever channel or campaign brought a prospect in originally. Someone clicked a paid search ad, went through a six-month sales cycle, and eventually signed — that ad gets 100% of the revenue credit. Regardless of everything else that happened in between.

Where this model earns its keep is awareness measurement. If your main question is "where does our pipeline actually come from?", first touch data is directly relevant. It tells you which channels are working at the top of the funnel.

The limitation is obvious once you think about it. B2B buying cycles are long. A prospect might find you through a blog post, then attend a webinar, read a few case studies, have two calls with a sales rep — and first touch credits only that original blog post. Everything that actually moved the deal forward gets ignored entirely.

Last Touch Attribution

Last touch flips it. All the credit goes to the final touchpoint before conversion — whatever the prospect did immediately before becoming a customer or submitting a lead form.

This is common in sales-focused reporting. It highlights what closed the deal, which feels useful. If a demo request was the last action before conversion, last touch says the demo drove the revenue.

The problem? It systematically undervalues everything above that final step. Awareness channels, nurture content, early-stage paid — none of it gets credit, even when it was doing real work moving prospects through the funnel. We see this constantly during technical audits: teams cutting content and social spend because last touch data makes them look like they're not contributing. In reality, they're filling the top of the funnel that bottom-of-funnel tactics depend on entirely.

When a Single-Touch Model Makes Sense

Single-touch models aren't worthless. If you're early in your attribution journey, or your sales cycle is short and relatively straightforward, they give you a usable baseline. Think of them as a diagnostic starting point — not a definitive answer on where to put budget.

Teams running complex, multi-channel campaigns across long buying cycles will start seeing misleading data fairly quickly. A common mistake we see is teams continuing to rely on last touch well past the point where it's creating real budget distortions — cutting the channels that are actually feeding the pipeline because the model simply can't see their contribution.

Multi-Touch Attribution Models: Linear, Time Decay and W-Shaped

Single-touch models force a choice. One moment gets the credit, everything else gets ignored. Multi-touch attribution works differently — it spreads credit across the full buyer journey, which is a far more honest way to look at how B2B deals actually close, particularly when campaigns run across multiple channels over weeks or months.

Three models are worth understanding properly: linear, time decay, and W-shaped. Each distributes credit differently. The right fit depends on how your pipeline actually works.

Linear Attribution

Linear attribution is simple. Divide credit equally across every touchpoint before conversion. Six interactions — paid search, a webinar, two emails, a case study, a sales call — each gets one-sixth.

As a starting point, the logic holds up. No assumptions about which stage matters most, no over-crediting a single channel. For teams new to multi-touch modelling, it gives a clean picture of which channels showed up across the journey.

The problem is that equal weighting is almost never accurate. A brand awareness ad from month one and a product demo request two days before signing are not doing the same job. Linear attribution treats them as if they are.

Time Decay Attribution

Time decay still distributes credit across all touchpoints. But interactions closer to conversion get more weight.

A demo request or pricing call in the final week carries more influence than a blog post read six months earlier. That's the assumption baked into this model, and for a lot of B2B sales cycles, it's a reasonable one. Early awareness still gets credited — it just doesn't get treated as equally responsible for closing the deal.

This model suits teams with long sales cycles where late-stage nurture and direct sales activity genuinely drive the final decision. We see it work particularly well when the sales team is already focused on closing efficiency and wants attribution to reflect that.

W-Shaped Attribution

W-shaped attribution assigns fixed, higher credit to three specific moments: first touch, lead creation, and opportunity creation. Everything in between shares the remaining credit equally. Plot it on a chart and you get the W shape the model is named after.

Most SaaS teams miss how well this aligns with how pipeline conversations already happen internally. It recognises that becoming a lead and becoming a qualified opportunity are genuinely significant milestones — not just arbitrary interactions along the way. First touch, MQL, SQL: each matters for a different reason, and W-shaped attribution treats them accordingly.

If your organisation tracks revenue attribution by pipeline stage, W-shaped is usually the most natural place to start with multi-touch modelling.

Choosing Between These Models

There's no universally correct answer. Linear gives you full-journey visibility. Time decay suits teams focused on closing efficiency. W-shaped works when you need attribution to align with how your CRM already tracks pipeline progression.

The tricky part isn't the decision itself — it's making sure the model you choose actually reflects how your buyers behave, rather than just fitting the reporting structure you already have. Those two things are often further apart than teams expect.

Data-Driven Attribution: When and Why to Move Beyond Rules

Rule-based models — first touch, last touch, linear, time decay — all share the same fundamental problem. The credit weighting is decided before you look at the data. It might be a reasonable assumption about how your buyers behave. But it's still a guess.

Data-driven attribution removes that guess.

Instead of applying a preset formula, it analyses the paths that led to conversions versus the ones that didn't, then works backwards to identify which touchpoints actually changed outcomes. Credit comes from your own buyer behaviour. Not a generic template someone else designed.

How Algorithmic Attribution Works

The most common approach uses Shapley values — borrowed from game theory. The question it asks is simple: if we removed this touchpoint from the path, how much would conversion probability drop? Channels that consistently appear on converting paths, especially at moments where they shift buying behaviour, earn higher credit.

Machine learning attribution takes this further. Models trained on large volumes of path data detect non-linear patterns — a specific combination of touchpoints in a specific sequence outperforming any single channel alone. A rule-based model simply cannot see that.

When Data-Driven Attribution Is the Right Move

The honest answer: not always, and not yet for every team.

These models need volume. Without enough conversion events across multiple channels, the output isn't stable. You're not getting insight — you're getting noise that looks like insight. Before switching, make sure you have consistent conversion data and enough completed paths for the model to find real patterns rather than statistical accidents.

Clean, connected data matters just as much. We see this constantly during audits — a team commits to machine learning attribution, then discovers their CRM, ad platforms, and analytics tools have been tracking conversions inconsistently for months. A sophisticated model fed bad inputs just produces sophisticated-looking bad outputs.

The Practical Trade-Off

More accuracy, less explainability. That's the honest summary.

A linear model takes thirty seconds to explain to a CFO. A Shapley-value model takes preparation, documentation, and a willingness to translate methodology into plain language before you walk into the room. That gap is manageable — but you have to manage it deliberately.

Build internal documentation around how the model works. Stakeholders need to trust the outputs without having to follow the maths. Used correctly, data-driven attribution is the most honest picture of how your marketing actually performs. It stops you over-investing in the channels that happen to be present when deals close, and under-investing in the channels that built the intent that made closing possible in the first place.

Choose the Attribution Model That Fits Your B2B Reality

There's no universal answer. The right model depends on your sales cycle length, how many touchpoints buyers move through, and how tightly marketing and sales actually share data.

Short pipeline, a handful of dominant channels? A simple rule-based model might cover it. But deals running six to twelve months across ten different touchpoints need multi-touch or data-driven attribution. Without that, you're guessing which early-stage activity deserves credit.

The most common mistake we see is teams choosing a model based on what's easiest to set up — not what reflects how their buyers actually behave. That gap between convenience and accuracy is exactly where budget disappears.

Before you pick any model, audit your tracking. A sophisticated attribution setup built on inconsistent data will produce numbers you can't trust. And decisions made on bad numbers? Worse than no decisions at all.

Clean, consistent tracking across every channel isn't optional — it's the foundation everything else sits on. If you want to go deeper on putting attribution to work across your whole B2B demand generation programme, that's where the commercial impact really compounds.