B2B Marketing Attribution Models Explained (2026)

"We tried LinkedIn ads. They didn't work."

When I hear this, my first question is: how did you measure whether they worked?

The answer is usually one of two things: the LinkedIn Campaign Manager said cost per conversion was too high, or there were no obvious direct-attribution deals from LinkedIn in the CRM.

Both of those measurements are using the wrong model. LinkedIn is primarily a warm-up channel. Its value often shows up in the middle of the buyer journey — not as the first or last touch, and not in LinkedIn's own conversion data. An attribution model that only looks at first or last touch will systematically undervalue LinkedIn and over-attribute credit to retargeting campaigns that caught the buyer at the end of a journey that LinkedIn helped create.

Attribution models are not just accounting conventions. They directly influence which channels you fund.

---

TL;DR

• First-touch over-credits awareness channels and ignores everything that converted the buyer
• Last-touch over-credits conversion campaigns and ignores everything that created the buyer
• Linear is balanced but treats a casual blog view the same as a demo booking
• W-shaped is the best fit for B2B: it weights both the awareness moment and the conversion moment
• CRM-linked attribution is more trustworthy than ad platform attribution for budget decisions
• With Cogny, your AI can compare ad-platform claims with CRM pipeline and measure each budget change you make

---

Why Attribution Models Exist

Before there were multi-touch attribution models, most B2B marketers used "last click" because it was the only thing measurable. The last ad the prospect clicked before filling out the form — that's what got credit.

The problem became apparent quickly: last-click attribution makes your retargeting campaigns look brilliant and your brand campaigns look useless. Of course retargeting converts — it's hitting people who were already close to converting. Brand campaigns do the work of creating awareness that makes retargeting possible. Last-click doesn't see that.

Multi-touch attribution emerged as a way to distribute credit across the full buyer journey. But distributing credit requires a model — a set of rules that determines how much each touchpoint contributes. Those rules have major consequences.

Attribution models compared

First-Touch Attribution

How it works: The first marketing touchpoint that brought a prospect to your brand receives 100% of the deal credit.

What it tells you: Which channels are best at introducing your brand to net-new buyers. If LinkedIn shows up heavily in first-touch attribution, it means LinkedIn is effective at creating initial awareness.

What it hides from you: Everything that happened between awareness and close. The nurture emails. The retargeting that re-engaged them after they went quiet. The blog post a colleague shared that reactivated their interest. The sales call that made the technical case. First-touch ignores all of it.

When to use it:

• Evaluating investment in new awareness channels (podcast sponsorships, new social platforms, influencer campaigns)
• Understanding which channels are introducing you to buyers in your target ICP
• Deciding which top-of-funnel content investments to continue

When not to use it:

• Evaluating the ROI of channels that primarily play in the middle or bottom of funnel
• Making budget decisions that affect conversion-stage campaigns
• Reporting attribution to a CFO who needs to understand return on marketing spend

Practical example:

A B2B company runs awareness ads on LinkedIn (sponsored thought leadership posts) and direct-response ads on Google (search campaigns targeting in-market buyers). First-touch attribution would credit LinkedIn heavily — because many buyers first heard of the company through a LinkedIn post. The Google search campaign, which often captures buyers after they've already decided to evaluate the product, would appear weaker.

If you cut LinkedIn spend based on first-touch attribution, you would be measuring it correctly for what it is (an awareness tool), but making a budget decision that damages the pipeline you don't realize LinkedIn was creating.

Last-Touch Attribution

How it works: The last marketing touchpoint before conversion receives 100% of the deal credit.

What it tells you: Which channels are most effective at triggering the conversion event — the form fill, the demo request, the booking.

What it hides from you: Everything that built the relationship before the final click. The LinkedIn ads that created awareness. The blog posts that established credibility. The webinar that moved the buyer from curious to interested. Last-touch ignores all of it.

When to use it:

• Optimizing bottom-of-funnel campaigns (retargeting, direct-response ads)
• Understanding which landing pages or CTAs convert most effectively
• Measuring the impact of conversion-stage campaigns in isolation

When not to use it:

• Understanding the full channel contribution to revenue
• Evaluating awareness or nurture investments
• Making decisions about top-of-funnel budget based on direct revenue attribution

The retargeting distortion:

Last-touch attribution systematically over-credits retargeting. Retargeting campaigns target people who have already shown interest — they've visited your site, consumed your content, or engaged with your ads. They're closer to buying than a cold prospect. The retargeting ad that appears when they've already made up their mind looks like the cause of the conversion, when it was really a nudge at the end of a journey you can't see in last-touch data.

Linear Multi-Touch Attribution

How it works: Equal credit is distributed across every touchpoint in the buyer journey. If there were 5 touchpoints, each gets 20%.

What it tells you: A balanced view of all the channels that touched the buyer at any point in their journey. No channel is ignored; none is overweighted.

What it hides from you: The difference in impact between touchpoints. A prospect briefly scanning a blog post vs. spending 30 minutes watching a product demo vs. clicking directly from a retargeting ad to your booking page — linear attribution treats these the same.

When to use it:

• Getting a baseline view of channel contribution that doesn't over-index on any single touch
• Presenting attribution data to stakeholders who distrust single-touch models but don't want to debate weighting assumptions
• Evaluating channels that appear in the middle of the funnel (content marketing, email nurture) that last-touch systematically undervalues

The equal-credit problem:

Linear attribution implies that a LinkedIn impression is worth as much as a demo request form fill. That's not how buying works. Some touchpoints have outsized influence on whether a deal closes. The blog post that answered the prospect's core objection mattered more than the newsletter they skimmed. Linear attribution can't distinguish between them.

W-Shaped Attribution

How it works: Credit is distributed with specific weighting based on where in the journey a touchpoint occurred:

• 40% to the first touch (first marketing interaction that created awareness)
• 40% to the conversion touch (marketing interaction that triggered the conversion event — typically the session where they filled out a form or booked a call)
• 20% distributed equally across all middle touches

What it tells you: A weighted view that rewards both awareness (for its role in creating the buyer) and conversion (for triggering the action). Middle touches get meaningful credit but not disproportionate credit.

What it hides from you: The relative impact of different middle touches from each other — they're treated equally within the 20% pool.

Why W-shaped fits B2B best:

B2B deals hinge on two critical moments:

1. When a prospect first encounters your brand and decides it's worth paying attention to
2. When a prospect decides to actually act and book a meeting or request a demo

W-shaped attribution is built around these two moments. It's not perfect, but it's more accurate than any single-touch model and avoids the false equality of linear attribution.

Implementation in common tools:

• HubSpot Attribution Reports (Marketing Hub Professional+): W-shaped attribution is a built-in model option
• Salesforce Pardot/Marketing Cloud: W-shaped is available in the Attribution models menu
• GA4: W-shaped is not natively supported; you'd need to implement it in BigQuery using GA4's exported event data

Data-Driven Attribution

How it works: Machine learning algorithms analyze the customer journey data across your entire account and calculate the probability that each touchpoint contributed to conversion. Credit is distributed based on the statistical influence of each touchpoint, not a fixed formula.

What it tells you (in theory): The most accurate possible credit distribution, because it's based on actual patterns in your data rather than arbitrary percentage splits.

What it requires: Enough data for the algorithm to identify statistically meaningful patterns — typically 1,000+ conversions per month to be reliable. Most B2B companies don't have this volume.

The B2B catch: Data-driven attribution is designed for high-volume, fast-cycle conversions (e-commerce, SaaS trials, lead forms). B2B sales cycles with months-long journeys and relatively few monthly deal closings don't give the algorithm enough signal to learn from. Use data-driven attribution if you have the volume; otherwise, W-shaped is more reliable.

CRM-Linked Attribution vs. Ad Platform Attribution

This distinction matters more than any model choice.

Ad platform attribution (what LinkedIn Campaign Manager, Google Ads, or Meta Ads reports):

• Uses pixel-based tracking to identify when someone who saw or clicked your ad later converted
• Has attribution windows: Google Ads uses 30-day click, 1-day view-through by default; LinkedIn uses 30-day click, 30-day view-through
• Claims conversions based on probabilistic matching, not confirmed CRM records
• Typically inflates conversion counts because multiple platforms claim the same conversion (you might see LinkedIn claim 40 conversions and Google claim 30, while your CRM shows 25 new leads — all three numbers include some of the same people)

CRM-linked attribution (what your HubSpot or Salesforce reports, when UTM data is flowing correctly):

• Credits campaigns based on the UTM data captured when the prospect first converted on your site
• Is deterministic, not probabilistic — a deal either has a UTM campaign value or it doesn't
• Doesn't double-count — each deal is attributed to one campaign (or if multi-touch, to the campaigns that touched that specific contact, confirmed by CRM data)
• Often shows lower conversion counts than ad platform data, which is actually the accurate picture

The gap between them is your tracking leak:

If LinkedIn says 50 conversions last month and your CRM shows 15 deals with LinkedIn UTM attribution, you have a significant tracking gap. Some of that gap is legitimate (LinkedIn counts form fills that didn't become CRM deals, or counts view-through conversions that your UTM system can't capture). But if the gap is large, it usually means UTMs are leaking somewhere in the funnel — redirect chains stripping parameters, booking tools not capturing hidden fields, or UTMs not being passed to deal records.

The CRM number is more trustworthy for budget decisions because it reflects actual sales outcomes. Use the ad platform number as a directional signal, not as ground truth.

Which attribution model fits which sales cycle?

Match the model to the length of the cycle and the number of deals. Short cycles of days or a few weeks, with one or two touches, are served well by last-touch or time-decay. Mid-length cycles of one to three months suit W-shaped, which credits the first touch and the conversion. Long, committee-led cycles need W-shaped or data-driven if volume allows, always checked against CRM-linked pipeline.

Choosing the Right Model for Your Stack

Regardless of which model you use, maintain CRM-linked attribution as your source of truth. The model determines how you interpret the journey; the CRM data determines whether the journey actually happened.

How to run this with Cogny

Connect your ad platforms (LinkedIn Ads, Google Ads, Meta Ads) and your CRM (HubSpot through HubSpot's own MCP server, Pipedrive or Upsales) to Cogny. Then ask your own AI or the Cogny app to compare what each ad platform claims with the pipeline your CRM shows by UTM source and campaign, and to flag campaigns where the two disagree. On Solo, Cogny AI and Cogny Pro, a finished chat can become a Loop that reruns weekly; on Cloud, it can be a scheduled report.

When the comparison leads to a budget move, approve it and Cogny ships it. A budget change made through Cogny becomes a ticket on its own, read three days later and closed as worked, partial, didn't work or inconclusive, so the next attribution review starts from what the last change did.

Frequently asked questions

What is multi-touch attribution in B2B marketing?

Multi-touch attribution in B2B distributes credit for a closed deal across multiple marketing touchpoints rather than giving all credit to a single interaction. This reflects the reality of B2B buying: most deals involve 5-15 touchpoints across weeks or months — ads, blog posts, emails, webinars, sales calls — before a deal closes. Multi-touch models vary in how they distribute credit: linear gives equal weight to each touch, W-shaped weights the first and conversion touchpoints most heavily, and time-decay weights recent touches more.

What is the difference between first-touch and last-touch attribution?

First-touch attribution gives 100% of the deal credit to the marketing touchpoint that first brought the prospect to your brand (the initial awareness moment). Last-touch attribution gives 100% of credit to the last marketing interaction before the prospect converted (the final nudge). Both are single-touch models, which means they ignore all the middle touchpoints. First-touch is useful for measuring awareness channel effectiveness; last-touch is useful for measuring conversion channel effectiveness. Neither gives an accurate picture of the full buyer journey.

What is W-shaped attribution?

W-shaped attribution is a multi-touch model that distributes credit with specific weighting: 40% to the first touch (first marketing interaction that introduced the prospect to your brand), 40% to the conversion touch (the marketing interaction that triggered the conversion event, typically a form fill or demo request), and 20% distributed equally across all middle touches. The 'W' shape refers to the credit curve visualized across a timeline — high at the start, lower in the middle, high at the conversion moment. It's well-suited for B2B because it values both awareness (which matters for long cycles) and the specific campaign that triggered action.

What is CRM-linked attribution and how does it differ from ad platform attribution?

CRM-linked attribution traces deal credit using data in your CRM (HubSpot, Salesforce) — specifically the UTM source captured when the contact first converted, or the campaign influence data if your CRM supports it. Ad platform attribution (LinkedIn, Google Ads) uses pixel-based tracking and claims conversions based on who saw or clicked your ads within attribution windows. The two often disagree significantly: LinkedIn might claim 50 conversions last month while your CRM shows only 20 deals from LinkedIn sources. CRM-linked attribution is more trustworthy for budget decisions because it reflects actual sales outcomes, not probabilistic ad-platform claims.

Which attribution model should I use for B2B?

For most B2B companies, use W-shaped multi-touch attribution as your primary model. It rewards both the channels that create awareness (which matters in B2B) and the channels that drive conversions (which matters for ROI measurement). Supplement with first-touch reports when evaluating new top-of-funnel channels, and last-touch reports when optimizing conversion campaigns. Always cross-reference ad platform attribution data against CRM-linked data — the gap between them reveals where your tracking has leaks.

---

Part of the B2B Attribution Series

• Marketing Attribution for B2B: Which Campaigns Actually Drive Sales Calls — The pillar guide (start here)
• How to Track Which Marketing Channel Books Your Sales Calls — UTM and CRM setup
• Does LinkedIn Ad Warm-Up Increase Call Booking Rates? Data & Framework — LinkedIn warm-up measurement
• Connecting Ad Spend to Closed Deals: A Practical Attribution Playbook — End-to-end implementation

---

Need help choosing the right attribution model for your B2B stack? Book a call with Tom or compare Cogny's plans.