Marketing Attribution Models Explained: First-Touch, Last-Touch, Multi-Touch, and Data-Driven

Most operators running a Monday review have seen this scene. Marketing reports that SEO and paid social drove 38% of qualified pipeline. Sales reports that direct traffic and outbound drove 62% of closed revenue. Both numbers are correct under their respective attribution models. Both numbers are useless for deciding next quarter's budget. This guide explains every marketing attribution model that matters, how each distributes credit across the buyer journey, and the decision framework for choosing the one that fits your sales cycle, conversion volume, and business model.

The problem is not dishonesty. The problem is that teams use different rules to score the same game. Marketing reports first-touch or multi-touch to defend top-of-funnel spend. Sales reports last-touch to credit the closer. Finance reports last-non-direct to strip out organic traffic that would have converted anyway. Until the company agrees on one model, every budget conversation is a negotiation disguised as analysis.

This article treats attribution as a decision, not a default. We cover each model's mechanics, its biases, the business models it fits, and the business models it breaks. Then we provide a decision matrix you can run against your own numbers this week. For a broader look at how attribution fits into the full RevOps operating system, see our complete RevOps guide.

What marketing attribution actually means

Attribution is not measurement. Measurement tells you what happened: 1,000 clicks, 50 leads, 10 deals. Attribution tells you why it happened and which touchpoints deserve the credit. A click is a measurement. Saying that click was worth $4,200 in attributed revenue is an attribution claim. The model is what bridges the gap.

The buyer journey in most B2B companies is not linear. A prospect might discover your company through a LinkedIn post, download a whitepaper two weeks later, attend a webinar a month after that, click a retargeting ad, visit the pricing page directly, and finally respond to a sales outbound sequence. Seven touches. Six channels. One deal. The attribution model decides whether LinkedIn gets 100% of the credit, the outbound email gets 100%, or the credit is split seven ways.

Marketing LTB's 2026 attribution survey found that 91% of marketers say attribution is important to success, but only 31% are very confident in their current model. The gap between importance and confidence is the gap this article addresses. Most teams are running a model they inherited from their CRM default and have never stress-tested against their actual sales cycle.

Before choosing a model, you need two pieces of data: your average sales cycle length and your average monthly conversion volume. A model that works for a D2C brand with a 7-day cycle and 2,000 monthly orders will break for an enterprise B2B company with a 9-month cycle and 15 deals. The decision matrix at the end of this article uses both inputs.

First-touch attribution

First-touch attribution gives 100% of conversion credit to the channel that brought the prospect into your ecosystem. The first click. The first view. The first interaction. Every subsequent touch gets zero credit.

When it works: First-touch is useful for understanding demand generation. It tells you which channels are creating awareness and bringing new prospects into the funnel. If your primary question is "where did our customers first hear about us?" first-touch answers it directly.

When it breaks: First-touch systematically under-credits the channels that close deals. A prospect who discovers you through a blog post, engages with five nurture emails, attends a webinar, and finally converts through a sales call will attribute 100% of revenue to the blog post. The sales team's effort is invisible. The nurture sequence is invisible. The webinar is invisible. Over time, first-touch attribution starves bottom-of-funnel investment and over-funds top-of-funnel content.

Best fit: Early-stage companies with a single dominant acquisition channel, or brand-awareness campaigns where the goal is measuring reach rather than revenue allocation. First-touch is also useful as a secondary model. Run it alongside a multi-touch model to see which channels are creating demand versus which channels are closing it.

First-touch adoption sits at approximately 19% of companies in 2026, down from higher levels in prior years as multi-touch models have become more accessible. Most teams that still run first-touch do so as one of two parallel models, not as their primary allocation rule.

Last-touch attribution

Last-touch attribution gives 100% of conversion credit to the final touch before conversion. The last click. The last ad view. The last email open. Everything that happened before gets zero credit.

When it works: Last-touch is the default in most advertising platforms for a reason: it is simple to implement and easy to understand. For short sales cycles — D2C ecommerce with impulse purchases, low-consideration SaaS signups — last-touch can be a reasonable approximation. If the journey from first touch to conversion happens in a single session, there is only one touch to credit.

When it breaks: Last-touch is the most common attribution failure in B2B. A typical B2B buyer journey involves five to ten touches across multiple channels over several months. Last-touch attributes 100% of revenue to the final click — usually direct traffic, branded search, or a sales outbound email — and gives zero credit to the content, events, and nurture sequences that built the relationship. The result: marketing cuts brand and content budgets because they "don't drive revenue," and sales gets over-credited for closing deals that marketing created.

Best fit: Short-cycle D2C with single-session conversions. Last-touch is also useful for understanding immediate conversion triggers — what caused the prospect to act today, as opposed to what caused them to become a prospect in the first place.

Last-touch remains the most widely used model at 41% adoption, largely because it is the default in Google Analytics, most ad platforms, and many CRMs. That default status is not a recommendation. It is a historical artifact of what was easy to build, not what is accurate to use.

Linear attribution

Linear attribution distributes credit equally across every touchpoint in the buyer journey. Five touches? Each gets 20%. Ten touches? Each gets 10%. The model assumes every interaction contributed equally to the final outcome.

When it works: Linear is the fairest simple model. It acknowledges that multiple touches matter without making complex assumptions about which touch mattered more. For teams just moving beyond single-touch attribution, linear is a defensible starting point. It is also useful when you genuinely do not know which touches are most influential and want to avoid the bias of first-touch or last-touch.

When it breaks: Equal distribution is almost never accurate. The first touch that created awareness and the last touch that triggered conversion almost always matter more than the middle touches that maintained engagement. Linear attribution under-credits high-impact touches and over-credits low-impact touches. A prospect who clicked a banner ad by accident and spent two seconds on the site gets the same credit as the sales demo that closed the deal.

Best fit: Teams transitioning from single-touch to multi-touch who need a simple, defensible model while they gather data for something more sophisticated. Linear is also useful for content teams who want to show that middle-funnel nurture content contributes to revenue, even if the exact weight is debatable.

Linear adoption sits at approximately 14% in 2026. Most teams that start with linear move to time-decay or position-based models within two quarters once they have enough data to justify the switch.

Time-decay attribution

Time-decay attribution gives more credit to touches that happened closer to the conversion. The logic is simple: recent touches had more influence on the decision to buy today than touches from six months ago. The exact weighting varies by implementation, but the standard formula gives each touch exponentially more credit as it gets closer to conversion.

When it works: Time-decay is the best simple model for short-to-medium sales cycles. A D2C brand with a 14-day consideration window, a SaaS company with a 30-day trial-to-paid flow, or any business where recency correlates with purchase intent. Time-decay also handles long nurture sequences well — the early touches get some credit, but the touches that pushed the prospect over the line get the most.

When it breaks: Time-decay under-credits brand-building and demand-generation activities that happened far from the conversion date. A prospect who first encountered your brand at a conference six months ago, forgot about it, and later searched for your brand name gets almost no credit for the conference touch. If your business depends on long-term brand awareness, time-decay will systematically under-invest in the activities that create it.

Best fit: D2C ecommerce with 7-30 day cycles. SaaS free-to-paid with 14-60 day windows. Any business where the decision to convert is heavily influenced by recent activity rather than long-term brand exposure.

Time-decay adoption is approximately 12% in 2026. It is the most common upgrade path for teams moving from last-touch — it keeps the simplicity while acknowledging that earlier touches matter.

U-shaped attribution

U-shaped attribution (also called position-based attribution) gives the most credit to the first touch and the last touch, with the remaining credit distributed across the middle touches. The standard split is 40% to first touch, 40% to last touch, and 20% divided equally among everything in between.

When it works: U-shaped is the most honest simple model for B2B SaaS with defined funnel stages. It acknowledges that creating the prospect and closing the prospect are the two highest-impact moments, while still giving some credit to the nurture activities that kept the prospect engaged. The 40-40-20 split is arbitrary but directionally correct for most B2B journeys.

When it breaks: U-shaped assumes the first and last touches are always the most important. That is not always true. In account-based sales, the middle touches — the demo, the security review, the procurement negotiation — can be more influential than the initial awareness touch. In product-led growth, the "last touch" before conversion might be an in-app notification, which gets 40% credit while the onboarding sequence that drove activation gets almost nothing.

Best fit: B2B SaaS with 30-90 day sales cycles and a clear handoff from marketing to sales. U-shaped is the default recommendation for teams with fewer than 300 monthly conversions who cannot yet run data-driven attribution.

U-shaped adoption sits at approximately 9% in 2026, but it is growing rapidly as B2B marketing teams move beyond single-touch models. For a deeper look at how attribution connects to the broader marketing-channel ROI calculation, see our guide to calculating marketing channel ROI.

W-shaped attribution

W-shaped attribution adds a third high-credit milestone to the U-shaped model: the lead creation touch. The standard split is 30% to first touch, 30% to lead creation, 30% to last touch, and 10% divided among the remaining touches. The "W" shape refers to the three peaks of credit distribution across the journey.

When it works: W-shaped is the best model for B2B companies with a clear marketing-qualified lead stage. It recognizes three critical moments: the touch that created awareness, the touch that converted the prospect to a lead, and the touch that closed the deal. This aligns well with how most B2B marketing and sales teams think about their funnel. It also defends the middle-funnel content — the webinars, the gated downloads, the demo requests — that U-shaped under-credits.

When it breaks: W-shaped requires a clear definition of "lead creation." If your CRM has fuzzy stage definitions, or if leads are created through multiple channels simultaneously, the model becomes ambiguous. W-shaped also breaks in product-led growth models where there is no discrete "lead creation" event — the user signs up and starts using the product without a traditional MQL stage.

Best fit: B2B SaaS and services with a defined MQL stage and a sales-assisted conversion process. W-shaped is the model we recommend most often for B2B companies between $2M and $20M ARR that have clean CRM data but not enough conversion volume for data-driven attribution.

W-shaped adoption is approximately 6% in 2026, but it is the fastest-growing position-based model among B2B marketing teams. Most teams that adopt W-shaped do so after running U-shaped for one or two quarters and realizing they need more credit in the middle of the funnel.

Data-driven attribution

Data-driven attribution (also called algorithmic or machine-learning attribution) uses statistical models to calculate the actual incremental contribution of each touchpoint. Rather than applying a fixed rule like "40% to first touch," the model analyzes historical conversion paths and estimates how much each touch increased the probability of conversion.

When it works: Data-driven attribution is the most statistically defensible model when you have the data to support it. It captures interaction effects that rule-based models miss — the fact that a webinar followed by a demo request converts at a higher rate than either touch alone. It also adapts over time as your buyer journey evolves, rather than requiring manual model updates.

When it breaks: Data-driven attribution has three failure modes. First, it requires volume. Most implementations need 300 to 400 conversions per month to produce stable results. Below that threshold, the model overfits to random noise and produces unstable credit allocations that change month to month. Second, it requires clean data. If your identity resolution match rate is below 60% — meaning you cannot reliably connect the same prospect across devices and sessions — the model is computing on incomplete data. Third, it is a black box. Rule-based models have the virtue of being explainable. Data-driven models often produce counterintuitive results that stakeholders reject because they cannot see the logic.

Best fit: High-volume B2C and D2C businesses with 400+ monthly conversions. Mid-market B2B SaaS with 300+ opportunities per month and clean CRM data. Enterprise companies with dedicated data teams who can validate the model's outputs against holdout tests.

Data-driven adoption is approximately 7% in 2026 but growing 44% year over year, making it the fastest-growing attribution category. Google Analytics 4 defaults to data-driven attribution for properties that meet the conversion threshold. The shift from rule-based to data-driven is the single biggest change in attribution practice over the past three years.

Decision matrix: which model fits your business

The right model depends on three inputs: your sales cycle length, your monthly conversion volume, and your business model. Here is the decision matrix we use with operators.

A practical rule: if you are unsure which model to choose, start with W-shaped for B2B and time-decay for D2C. Run it for two quarters. Then compare the attributed revenue by channel against your intuition and your finance team's view. If the model says content drives 5% of revenue and your sales team says half their deals mention a blog post, your model is wrong — not your sales team.

The most important decision is not which model you choose. It is that you choose one model, document it, and require every team to report against it. A mediocre model used consistently beats a perfect model used inconsistently. For more on building the operating cadence that makes attribution data actionable, see our weekly revenue review template.

How Fairview handles attribution across models

Fairview does not force you into one attribution model. The platform connects to your CRM, ad platforms, and finance tools, then lets you switch between first-touch, last-touch, linear, time-decay, U-shaped, and W-shaped views in the operating dashboard. Data-driven attribution is available on the Scale plan for teams with sufficient conversion volume.

The Margin Intelligence feature in Fairview goes a step further. It does not just attribute revenue to channels. It attributes profit. A channel that drives $100K in attributed revenue but costs $95K to operate shows up differently than a channel that drives $80K in revenue at $40K cost. Most attribution models ignore cost. Fairview includes it by connecting ad spend data from Google Ads and Meta Ads, revenue data from Stripe, and cost data from QuickBooks or Xero.

When a channel's attributed profit drops below the threshold you set, Fairview flags it in the weekly operating report with a named next-best action. The action is specific — "Review Google Ads spend by campaign" or "Shift $3K from paid social to content" — not a generic alert. This is the difference between attribution that reports and attribution that drives decisions.

The setup is straightforward. Connect your CRM, your ad platforms, and your finance tool. Fairview normalizes the data, applies the attribution model you select, and surfaces profit by channel, campaign, and SKU. First integration is live in under 10 minutes. See pricing and tiers for which attribution models are available on each plan.

Key takeaways

• The attribution model you choose determines which channels get budget and which get cut. It is a decision with real financial consequences, not a technical default.
• Seven models matter: first-touch, last-touch, linear, time-decay, U-shaped, W-shaped, and data-driven. Each distributes credit differently and fits different business models.
• Short-cycle D2C should use time-decay. B2B SaaS with 30-90 day cycles should use W-shaped or U-shaped. Enterprise B2B needs custom account-based models. Data-driven requires 300+ monthly conversions.
• The most common failure is not choosing the wrong model. It is letting different teams use different models and wondering why they disagree. One model, documented, used by all.
• Attribution should connect to profit, not just revenue. A channel that drives attributed revenue but erodes margin is not a channel to scale. Fairview's Margin Intelligence connects attribution logic to actual cost data.

Conclusion

Marketing attribution is not a solved problem. It is a managed trade-off between accuracy, explainability, and data requirements. The model that is perfect for a D2C brand will mislead an enterprise B2B company. The model that is statistically optimal will fail if your team does not trust it. The model your CRM shipped with is almost certainly not the model your business needs.

The right next step is diagnostic, not prescriptive. Look at your average sales cycle. Count your monthly conversions. Ask whether your current model matches both. If it does not, pick the model from the decision matrix that does, run it for one quarter, and compare the results against what your sales team and finance team already believe. The goal is not perfect attribution. The goal is attribution that is honest enough to make better budget decisions.

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

First-touch attribution credits 100% of a conversion to the first marketing interaction a customer had with your brand. Last-touch attribution credits 100% to the final interaction before conversion. First-touch favors top-of-funnel channels like SEO, content marketing, and brand advertising. Last-touch favors bottom-of-funnel channels like direct search, retargeting ads, and sales outreach. Neither tells the full story in a multi-touch buyer journey, which is why most operators now use multi-touch models.

What is multi-touch attribution and when should I use it?

Multi-touch attribution distributes credit across multiple touchpoints in the buyer journey rather than assigning 100% to a single interaction. Linear gives equal credit to every touch. Time-decay gives more credit to recent touches. U-shaped gives 40% each to first and last touch, with 20% to the middle. W-shaped adds a third high-credit milestone at lead creation. Use multi-touch attribution when your sales cycle involves more than two or three touches, which applies to most B2B companies and many D2C brands with consideration phases longer than 14 days.

What is data-driven attribution and do I need it?

Data-driven attribution uses statistical models to calculate the actual incremental contribution of each touchpoint based on your historical conversion data. Unlike rule-based models that apply fixed formulas, data-driven attribution learns from your specific buyer journeys. You need 300 or more conversions per month for stable results. Below that threshold, the model overfits to random noise. Data-driven attribution is the most statistically defensible approach for high-volume businesses but requires clean data, reliable identity resolution, and a team that trusts algorithmic outputs.

How do I choose the right attribution model for my business?

Match the model to your sales cycle length, conversion volume, and business model. Short cycles under 14 days with high volume can use last-non-direct or time-decay. B2B SaaS with 30-90 day cycles should use W-shaped or U-shaped. Enterprise B2B with 180+ day cycles needs custom account-based models. Teams with 300+ monthly conversions can run data-driven attribution. The most important rule is to pick one model, document it, and require every team to report against it. A consistent mediocre model beats an inconsistent perfect model.