B2B marketing attribution models: which one actually works for your sales cycle

Your marketing team spent $40K last quarter. LinkedIn, content, webinars, cold outreach. Then the pipeline review happens. Sales says half the deals came from referrals and the AEs worked those themselves. Marketing's dashboard says LinkedIn drove 68% of closed-won. Your CRM says "direct" on three of the five largest deals.

Who's right? Probably none of them. And the reason is almost always the same: the wrong attribution model is attached to the wrong sales cycle.

I've sat in enough quarterly business reviews to know what "attribution model" usually means at a B2B company: last-touch, set during HubSpot onboarding, never changed. The result is that the last thing a contact clicked before converting gets 100% of the credit. For a deal that took nine months and involved seven people, that's not measurement. It's a random number generator with a LinkedIn icon on it.

Here's the scope of the problem. A 2025 survey by Revsure found that nearly 90% of B2B marketing teams still rely on last-touch or basic multi-touch attribution. Only 21% of B2B marketers say they're confident in their attribution data. The average B2B buyer journey now runs 272 days, involves 88 separate touchpoints, and includes ten stakeholders. And 86% of teams struggle to even connect those stakeholders to a single opportunity in their CRM.

So the problem is not that attribution is hard. The problem is that most teams picked a model designed for a two-click e-commerce checkout and applied it to a nine-month enterprise sale.

The attribution models, actually explained

HubSpot offers nine attribution models. Most guides just list them. Here's what they're actually for.

First touch assigns 100% credit to the first interaction. Good for understanding what brings people into the funnel initially. Useless for understanding what closed the deal.

Last touch assigns 100% credit to the final interaction before conversion. Fast to set up, easy to understand, and deeply misleading for anything with a sales cycle longer than a few weeks.

Linear splits credit equally across every touchpoint. A contact with 20 interactions and a contact with 3 interactions each get proportional slices. This sounds fair but it can be too flat: a webinar that closed the deal gets the same credit as a Twitter ad they scrolled past six months earlier.

Time decay gives more weight to touchpoints closer to conversion. That logic works well for shorter cycles. For an 18-month enterprise deal, it ends up over-crediting the last few touches and ignoring the content that built awareness a year ago.

U-shaped (also called position-based) assigns 40% to the first touch, 40% to the lead creation moment (usually a form submission), and splits the remaining 20% across everything in between. This is a significant upgrade over last-touch for most B2B teams. It at least acknowledges that creating a lead matters, not just closing one.

W-shaped expands U-shaped by adding a third milestone: opportunity creation. Each of the three milestones gets 30% credit, and the remaining 10% is distributed across other touchpoints. This is the model I reach for most often with pipeline-focused teams. It matches how B2B sales actually work: something created awareness, something converted a lead, something created a real sales opportunity.

Full path (Enterprise tier in HubSpot) adds a fourth milestone for deal close, assigning 22.5% to each of the four major milestones. Best for teams with clean, well-defined CRM stages and enough closed deals to make the weights meaningful.

J-shape and Inverse J-shape are weighted toward the last or first touch respectively, with some distribution in the middle. Niche models. I rarely see them used well in B2B contexts.

Picking the right model for your sales cycle

The single most practical way to pick a model is to look at your average sales cycle length.

Most of Ziel Lab's clients are B2B companies with sales cycles between 60 and 180 days. W-shaped is our default recommendation for them. It measures three things that every B2B revenue team cares about: how someone heard about you, when they became a lead, and when they became a real opportunity. Those three moments are trackable, consistent, and meaningful.

If you're on HubSpot Professional, W-shaped is available out of the box. Enterprise tier adds full-path and data-driven models.

Why last-touch keeps surviving (and why it shouldn't)

Last-touch attribution is still the default at most companies not because it's accurate, but because it's easy to explain. "This channel drove that deal" is a clean story. CFOs like clean stories.

The problem is the story is almost always wrong.

I worked with a SaaS company last year that was seriously considering cutting their content program. Last-touch attribution showed content driving about 8% of pipeline. After switching to W-shaped, content was visible in the first-touch position on 43% of deals. It was doing the awareness work. The problem wasn't content: it was the model.

Setting up attribution in HubSpot

If you're on HubSpot Pro or Enterprise, here's the practical setup path.

First, decide what you're measuring: contact creation, deal creation, or revenue. Most B2B teams care about deal attribution (what marketing activities influenced deals), not just contact creation.

Second, audit your UTM parameters before you build anything. HubSpot's attribution depends on cookies linking sessions to contacts, UTM parameters labeling traffic sources, and interaction records. If your UTMs are inconsistent (sometimes "LinkedIn" sometimes "linkedin" sometimes "linkedin-paid"), your attribution data will be garbage regardless of which model you pick.

Third, go to Reports in HubSpot, select "Attribution reports", and create a custom deal attribution report. Choose W-shaped as your model. Add dimensions for channel, campaign, and asset type so you can see which specific campaigns are showing up at which milestones.

Fourth, think about offline touchpoints. Trade show conversations, referral introductions, executive briefings: these drive significant pipeline but don't appear in digital attribution automatically. HubSpot lets you log activities manually against contacts. If your team is doing this consistently, those activities will show up. If they're not, your model will systematically undercount offline channels.

See how we set up HubSpot for RevOps clients for what a clean attribution architecture looks like in practice.

The mistake that kills attribution before it starts

The most common error I see is companies choosing a model before fixing their data.

A W-shaped model on a CRM where 40% of contacts have no source data and 60% of UTMs are broken will produce W-shaped garbage. The model can't fix underlying data problems.

Before you configure attribution, run these checks:

1. What percentage of your contacts have a known original source? If it's below 70%, stop and fix that first.
2. Are your UTM parameters consistent across campaigns? Audit the last six months of campaigns.
3. Are deal stages in your CRM actually being used? W-shaped needs deal creation logged correctly to give that 30% credit to the right moment.
4. Are your reps logging activities in HubSpot? If not, offline touchpoints are invisible.

One more thing on this: data-driven attribution (available on HubSpot Enterprise) sounds like the most sophisticated option, but it requires statistically significant conversion volume to produce reliable weights. Most B2B teams with high-ACV, low-volume pipelines don't meet that bar. At 40 closed deals per quarter, the ML weights are guesses. W-shaped with clean data will outperform data-driven with messy data, every time.

What attribution still can't tell you

Even a well-configured W-shaped model in HubSpot has blind spots.

86% of B2B teams can't connect multiple stakeholders within the same account to a single opportunity. HubSpot tracks contacts, not buying committees. If the CFO read three of your case studies anonymously before the VP of Sales submitted the demo request, that reading behavior is invisible. The VP gets the first-touch credit because they're the identified contact.

This is why account-based attribution matters for teams selling to larger organizations. Tools like Marketo Measure (formerly Bizible) or Terminus are built for account-level attribution where multiple contacts across the same company need to be linked to one opportunity.

For most SMBs and Series A/B companies, HubSpot's native attribution is enough if the data is clean. For teams selling to enterprise accounts with five or more stakeholders per deal, you'll hit the ceiling quickly.

The other gap is offline influence. A warm introduction from a mutual connection, a conference conversation, a customer who vouched for you: these often drive the most important deals, and they're the hardest to attribute. Building in a structured process for reps to log referral sources and relationship origins in HubSpot is the closest thing to a fix.

For attribution that connects to actual pipeline health, see our work on RevOps and CRM setup and go-to-market strategy.

What to actually do this week

If you're still on last-touch attribution, here's the shortest path to something better:

Check your current settings (Reports > Attribution in HubSpot). If you see "Last Interaction," you're on the default. Switch to U-shaped as a first step: it takes two minutes and immediately gives you more signal. Then run an audit of your UTM consistency before doing anything more complex. If your UTM data is clean and your deal stages are being used correctly, switch to W-shaped and run it for one full quarter before drawing conclusions.

One quarter isn't enough to make budget decisions, but it's enough to see whether the story changes from what last-touch was telling you. In my experience, it usually does.

FAQ

What is the best attribution model for B2B?

For most B2B companies with sales cycles between 60 and 180 days, W-shaped attribution is the best starting point. It assigns 30% credit to first touch, 30% to lead creation, and 30% to opportunity creation, with the remaining 10% distributed across other touchpoints. This matches how B2B deals actually progress and gives credit to awareness-stage activities that last-touch models ignore entirely.

Which attribution models does HubSpot support?

HubSpot offers nine models: first interaction, last interaction, linear, time decay, U-shaped, W-shaped, J-shaped, inverse J-shaped, and full path. Full path and data-driven models are available on Enterprise tier only. W-shaped is available on Professional and above.

Why is last-touch attribution bad for B2B?

Last-touch attribution gives 100% of the deal credit to the last marketing interaction before a contact converts. For a deal that took nine months and involved ten stakeholders, this ignores everything that happened before the final click: the content that built awareness, the webinar that turned a cold contact warm, the case study that got the deal onto the shortlist. Budget decisions based on last-touch data consistently defund top-of-funnel programs that are actually driving pipeline.

How do I set up W-shaped attribution in HubSpot?

Go to Reports in HubSpot and create a custom attribution report. Select "Deal creation" as your attribution object. Choose the W-shaped model. Add dimensions for channel and campaign so you can see which channels show up at which milestones. Before building reports, audit your UTM parameters and make sure deal stages are being logged correctly in your CRM. Attribution models only work if the underlying data is clean.

What's the difference between U-shaped and W-shaped attribution?

U-shaped assigns 40% to the first touch, 40% to lead creation, and 20% distributed across everything in between. W-shaped adds a third major milestone: opportunity creation. Each of the three milestones gets 30%, with 10% distributed across other touches. If your team has a defined "opportunity created" stage in your CRM and uses it consistently, W-shaped is the better model because it measures one more thing that actually matters in a B2B sales process.