Attribution Models in Web Analytics

An attribution model decides which marketing channel gets credit for a conversion when users touch multiple sources before buying. This drives how you evaluate channel performance and allocate budget.

Most journeys aren't simple. Average path to purchase involves around seven touchpoints, each pushing the decision forward. Attribution models split credit across the path.

Core Principles

The model is the rule the analytics system uses to assign conversion credit. The full path equals 100%. Each touchpoint gets between 0 and 100%, depending on the model.

User Journey Example

Scenario: Online course purchase

Search query → blog visit (Organic Search)
Social network → ad view, subscription (Paid Social)
Email newsletter → discount received (Email Marketing)
Direct visit → final purchase (Direct)

Each model splits this conversion across the four channels differently.

Two Categories

Single-Touch ModelsMulti-Touch Models

Characteristics:

Assign 100% to one channel
Easy to grasp and implement
Ignore other interactions
Fit short sales cycles

Characteristics:

Split credit across channels
More complex, more realistic
Cover the whole journey
Fit long sales cycles

Single-Touch Models

First-Touch Attribution

100% goes to the first interaction.

graph LR
    A[Organic Search<br/>100%] --> B[Paid Social<br/>0%]
    B --> C[Email<br/>0%]
    C --> D[Direct<br/>0%]
    D --> E[Conversion]

    style A fill:#4CAF50,stroke:#333,stroke-width:3px,color:#fff
    style B fill:#f9f9f9,stroke:#ddd
    style C fill:#f9f9f9,stroke:#ddd
    style D fill:#f9f9f9,stroke:#ddd

When to Use First-Touch

New audience acquisition campaigns: evaluate brand introduction channels
Content marketing: analyze primary traffic sources
Branding campaigns: understand awareness-building channels

Limitations: Overvalues top-of-funnel channels. Ignores nurturing entirely.

Last-Touch Attribution

All credit to the final channel before conversion. Default in most analytics platforms.

graph LR
    A[Organic Search<br/>0%] --> B[Paid Social<br/>0%]
    B --> C[Email<br/>0%]
    C --> D[Direct<br/>100%]
    D --> E[Conversion]

    style A fill:#f9f9f9,stroke:#ddd
    style B fill:#f9f9f9,stroke:#ddd
    style C fill:#f9f9f9,stroke:#ddd
    style D fill:#4CAF50,stroke:#333,stroke-width:3px,color:#fff

Advantages:

Simple to implement and understand
Highlights channels that close
Standard across analytics platforms

Disadvantages: Misses the bigger picture for businesses with many ad sources, large budgets, or long deal cycles.

Last Non-Direct Click

Like last-touch, but ignores direct visits. Credit goes to the last identifiable marketing channel.

Why Exclude Direct Traffic

Problem: Users often return directly after first hearing about a brand through ads.

Solution: The model skips Direct and credits the last measurable channel.

Result: More accurate read on paid channel performance.

Multi-Touch Models

Linear Attribution

Splits value evenly across all touchpoints.

graph LR
    A[Organic Search<br/>25%] --> B[Paid Social<br/>25%]
    B --> C[Email<br/>25%]
    C --> D[Direct<br/>25%]
    D --> E[Conversion]

    style A fill:#2196F3,stroke:#333,stroke-width:2px,color:#fff
    style B fill:#2196F3,stroke:#333,stroke-width:2px,color:#fff
    style C fill:#2196F3,stroke:#333,stroke-width:2px,color:#fff
    style D fill:#2196F3,stroke:#333,stroke-width:2px,color:#fff

Application: Good for understanding total channel contribution and spotting undervalued sources.

Limitation: Touches don't influence decisions equally, so an even split misses real impact.

Time-Decay Attribution

Weights touchpoints closer to conversion higher. Earlier interactions decay exponentially.

graph LR
    A[Organic Search<br/>10%] --> B[Paid Social<br/>20%]
    B --> C[Email<br/>30%]
    C --> D[Direct<br/>40%]
    D --> E[Conversion]

    style A fill:#FFC107,stroke:#333,stroke-width:1px
    style B fill:#FF9800,stroke:#333,stroke-width:1.5px
    style C fill:#FF5722,stroke:#333,stroke-width:2px,color:#fff
    style D fill:#F44336,stroke:#333,stroke-width:3px,color:#fff

Time-Decay in Practice

B2B software purchase (3-month cycle):

Month 1: Google search (research) → 5%
Month 2: LinkedIn advertising (comparison) → 15%
Month 3: Email with commercial offer → 35%
Final: Direct visit for purchase → 45%

Best for: Long sales cycles, B2B, expensive goods with extended decisions.

Position-Based Attribution (U-shaped)

40% to first touch, 40% to last touch, 20% split across the middle.

graph LR
    A[Organic Search<br/>40%] --> B[Paid Social<br/>10%]
    B --> C[Email<br/>10%]
    C --> D[Direct<br/>40%]
    D --> E[Conversion]

    style A fill:#9C27B0,stroke:#333,stroke-width:3px,color:#fff
    style B fill:#E1BEE7,stroke:#333,stroke-width:1px
    style C fill:#E1BEE7,stroke:#333,stroke-width:1px
    style D fill:#9C27B0,stroke:#333,stroke-width:3px,color:#fff

Logic: Acquisition (awareness) and conversion (decision) channels matter most. Nurturing matters too, but less.

Best for:

E-commerce with active retargeting
SaaS with trial periods
Services with long consideration cycles

W-Shaped Attribution

U-shape plus a lead-generation milestone. 30% each to first touch, lead creation, and conversion.

Stage	Value Share	Rationale
First touch	30%	Attention attraction
Lead creation	30%	Interest manifestation
Conversion	30%	Decision making
Other touches	10%	Process support

Best for: B2B funnels with clear lead-gen stages.

Data-Driven Attribution

Machine learning analyzes historical paths to find each channel's real impact.

How it works:

graph TB
    A[Historical conversion<br/>data] --> B{Pattern analysis<br/>ML algorithms}
    B --> C[Identifying significant<br/>correlations]
    C --> D[Calculating weight<br/>coefficients]
    D --> E[Personalized<br/>attribution model]

Minimum Requirements

Data Volume: At least 15,000 clicks and 600 conversions over 30 days for statistical significance.

Technical Infrastructure: Advanced analytics platforms with ML capabilities.

Expertise: Data science team for setup and interpretation.

Advantages: Adapts to your business and audience.

Limitations: Complex to implement, demanding on data, opaque distribution logic.

How to Pick a Model

Sales Cycle Length

Short cycle (up to 7 days)Medium cycle (7-30 days)Long cycle (30+ days)

Behavior:

Impulse purchases
1-3 touchpoints
Quick decisions

Models:

Last-Touch for direct sales
First-Touch for acquisition channels

Behavior:

Considered purchases
Alternative comparison
3-7 touchpoints

Models:

Linear for even view
Position-Based to highlight key stages

Behavior:

Complex decisions (B2B, real estate, automobiles)
Multiple research phases
7+ touchpoints

Models:

Time-Decay for proximity weighting
Data-Driven for max accuracy

Business Model

Business Type	Decision Cycle	Recommended Model	Rationale
E-commerce (fast-moving goods)	1-7 days	Last-Touch	Focus on closing channels
SaaS B2B	30-90 days	Position-Based / Data-Driven	First touch and nurturing matter
Real Estate	90-365 days	Time-Decay	Decisions sit close to purchase
Educational Courses	14-60 days	W-Shaped	Clear interest → lead → purchase stages
Financial Services	30-180 days	Linear / Data-Driven	Many influence factors

Number of Channels

Channel Count Guidance

1-3 channels: Single-Touch may suffice.

4-7 channels: Multi-Touch becomes necessary.

8+ channels: Data-Driven gives best accuracy.

Data and Resources

Basic:

Standard Google Analytics
UTM tagging
Models: First-Touch, Last-Touch, Linear

Advanced:

Custom dimensions and events
CRM integration
Models: Position-Based, Time-Decay, W-Shaped

Enterprise:

Owned data infrastructure
Data science team
Models: Data-Driven, custom ML

Implementation

Tracking Setup

1. UTM-tag every campaign

UTM Tagging

https://example.com/landing?utm_source=facebook&utm_medium=cpc&utm_campaign=spring_sale&utm_content=video_creative&utm_term=running_shoes

Required parameters:

utm_source: traffic source
utm_medium: channel type
utm_campaign: campaign name

2. Define conversion goals

Macro-conversions: purchases, applications, subscriptions
Micro-conversions: downloads, content views, registrations

3. Connect CRM

CRM ties first touch through deal close, enabling LTV calculation.

Compare Models Side-by-Side

Channel	Last-Touch	First-Touch	Linear	Position-Based	Recommendations
Google Ads	45%	15%	25%	30%	Overvalued in Last-Touch
Facebook Ads	20%	35%	25%	30%	Undervalued in Last-Touch
Organic Search	15%	25%	25%	20%	Stable contribution
Email Marketing	12%	8%	25%	15%	Undervalued except in Linear
Direct	8%	17%	0%	5%	Overvalued without touch context

Conclusions:

Increase Facebook Ads (undervalued)
Reconsider Google Ads (likely overvalued)
Strengthen email (undervalued in most models)

Tools

Google Analytics 4

Built-in attribution and side-by-side model comparison.

Models available:

Last-click (default)
First-click
Linear
Time-decay
Position-based
Data-driven (with sufficient volume)

Setup: Reports → Attribution → Model comparison tool

Specialized Platforms

Adobe Analytics: Custom models and ML.

Google Ads Attribution: Bid optimization on attribution data.

Facebook Attribution: (Discontinued.) Cross-channel analysis between Facebook and other channels.

Alternatives: Mixpanel, Amplitude, Segment for flexible model configuration.

Custom Solutions

For large companies with unique needs:

graph TB
    A[Data collection from all sources] --> B[Data Warehouse]
    B --> C[ETL processes]
    C --> D[ML models for attribution]
    D --> E[Dashboards and reports]
    E --> F[Campaign optimization]

Limitations and Challenges

Cross-Device Tracking

Users research on mobile, buy on desktop. Cookie-based attribution can't link them.

Solutions:

User ID tracking for authenticated users
Probabilistic matching from behavioral patterns
Deterministic linking via email/phone

Privacy Regulation

GDPR, iOS 14.5+, third-party cookie deprecation reshape attribution.

New Limits

iOS 14.5+ App Tracking Transparency:

Users can opt out of tracking
Limited tracking in Safari
Facebook/Instagram attribution loses accuracy

Chrome third-party cookie phase-out:

Planned for 2025
Forces transition to first-party data
Complicates cross-site attribution

Server-Side Tracking

Server analytics is becoming standard for bypassing browser limits.

Advantages:

Independent of client-side limits
Greater data control
Higher tracking accuracy
Privacy compliance

Implementation: Requires technical expertise and infrastructure changes.

Future Direction

Privacy-First Attribution

Privacy Sandbox (Google):

Attribution Reporting API
Trust Tokens for fraud prevention
Topics API instead of third-party cookies

Apple Privacy-First Solutions:

SKAdNetwork for app attribution
Private Click Measurement for web
On-device machine learning

AI-Powered Attribution

Algorithmic Attribution 2.0:

Real-time model adjustment
Predictive attribution for future campaigns
Cross-channel optimization automation
Incremental lift measurement

Attribution distributes conversion value across marketing touchpoints. The right model depends on business specifics, sales cycle, channel count, and data volume.

As privacy tightens and journeys grow more complex, attribution evolves toward sophisticated, privacy-compliant solutions powered by AI and ML.

About AI participation in writing articles

This article, like many others on our site, was created, written and proofread by a team of developers. Of course, not without the participation of AI assistants. We don't hide this and believe that modern systems are already quite good at handling simple tasks and, relatively speaking, writing an article about Viewport yourself is quite strange. It won't come out significantly better and will take a lot of time. But providing basic understanding to beginner webmasters is necessary. Of course, after the article is written by assistants - there's always proofreading, and this is where not one or two people participate, and only after that the article is published.

Optimize Marketing Campaigns with Proper Attribution

Register for free testing of our web analytics tool and get detailed insights about your customer journey. Our attribution models help allocate marketing budget on data, not guesses.

Ready to take control of your web analytics? Try Statable free for 30 days — no credit card required, full feature access, GDPR-compliant by default. Start your free trial or view a live demo.