What Are Attribution Models?
Attribution models are rule sets that determine how credit for a conversion gets distributed across the touchpoints a user interacted with before converting. Every conversion has a path — a sequence of clicks, ad impressions, email opens, and organic visits — and attribution models decide which of those interactions deserve credit.
The core problem is simple: most conversions involve more than one interaction. A prospect might discover your brand through an organic search result, return via a retargeting ad, read a comparison article from an email newsletter, and finally convert through a direct visit. Each of those touchpoints played a role, but without a structured framework, teams default to crediting whichever touchpoint is easiest to measure — usually the last click.
This creates blind spots. Content that introduces prospects to your brand gets no credit. Nurture campaigns that keep leads engaged look worthless on paper. Teams cut the channels that actually fill the top of funnel, then wonder why lead volume drops two quarters later.
Attribution models solve this by establishing consistent rules for credit assignment, so marketing teams can evaluate channel performance with the full customer journey in mind rather than a single snapshot.
Types of Attribution Models
Six models cover the majority of real-world use cases. Each makes a different assumption about which touchpoints matter most.
First-Touch Attribution
All credit goes to the first interaction. If a user initially found your site through an organic search for "attribution models explained," that organic visit gets 100% of the conversion credit regardless of what happened afterward.
Last-Touch Attribution
The final interaction before conversion receives all credit. This is the simplest model to implement and the default in many legacy analytics setups.
Linear Attribution
Credit is split equally across every touchpoint. If a user had four interactions before converting, each receives 25%.
Time-Decay Attribution
Touchpoints closer to the conversion receive proportionally more credit. An interaction that happened two days before purchase counts more than one from three weeks ago.
Position-Based (U-Shaped) Attribution
The first and last touchpoints each receive 40% of credit, with the remaining 20% distributed evenly across middle interactions. This model values both discovery and closing while still acknowledging the nurture phase.
Data-Driven Attribution
Machine learning analyzes your actual conversion paths to determine how much credit each touchpoint deserves. Rather than following fixed rules, it compares converting paths against non-converting paths to identify which interactions statistically increase conversion probability.
Model Comparison
| Model | How It Works | Best For | Limitation |
|---|---|---|---|
| First-Touch | 100% credit to first interaction | Measuring brand awareness and top-of-funnel discovery | Ignores everything after initial contact |
| Last-Touch | 100% credit to final interaction | Short sales cycles with one or two touchpoints | Undervalues awareness and nurture channels |
| Linear | Equal credit to all touchpoints | Teams starting with multi-touch attribution | Treats a casual blog visit the same as a demo request |
| Time-Decay | More credit to recent touchpoints | Longer sales cycles where recency signals intent | Discounts early-stage content that planted the seed |
| Position-Based | 40% first, 40% last, 20% middle | Balanced view of discovery and conversion drivers | Arbitrary 40/40/20 split may not match actual influence |
| Data-Driven | ML-weighted based on actual paths | Accounts with enough conversion volume (300+/month) | Requires substantial data; opaque logic |
How to Choose the Right Model
Picking the right attribution model depends on three factors: your sales cycle length, channel mix complexity, and data maturity.
Short sales cycles (under 7 days): If most conversions happen in one or two sessions, last-touch attribution is often sufficient. The path is short enough that the final interaction genuinely represents the decision driver. This applies to many e-commerce purchases, free-tool signups, and newsletter subscriptions.
Medium sales cycles (1-4 weeks): Position-based or time-decay models work well here. Prospects typically interact with your brand three to six times before converting. You need visibility into what brought them in and what closed them, but middle-of-funnel nurture matters less.
Long sales cycles (1+ months): Linear or data-driven attribution becomes necessary. With dozens of touchpoints over weeks or months, fixed-rule models produce misleading results. If you have enough conversion volume — GA4 requires roughly 300 conversions per month for data-driven attribution — use DDA. Otherwise, linear attribution at least avoids the worst biases of single-touch models.
Channel mix complexity: The more channels you run, the more you need multi-touch attribution. A team running only organic search and email can get reasonable signal from last-touch. A team running paid search, display, organic, social, email, and affiliates needs a model that accounts for cross-channel interactions.
Data maturity: If your analytics setup is new or tracking is incomplete, start with position-based attribution. It provides a reasonable multi-touch view without requiring machine learning infrastructure. As your data matures and conversion volume grows, graduate to data-driven.
Attribution in GA4
Google Analytics 4 made data-driven attribution (DDA) the default model for all properties, replacing last-click as the baseline. This is a meaningful shift because it means most GA4 users are already running a multi-touch model, even if they haven't configured it explicitly.
How GA4's DDA works: GA4 uses conversion path data from your property to build a model that evaluates how each touchpoint contributes to conversions. It analyzes both converting and non-converting paths, applying algorithmic credit based on statistical impact rather than position in the path.
Comparing models in GA4: Navigate to Advertising > Model Comparison to see how credit distribution changes across models. This report lets you compare data-driven, last-click, and first-click side by side for each channel. Significant differences between models indicate channels where single-touch attribution was giving misleading signals.
Conversion paths report: The Advertising > Conversion Paths report shows the actual sequences users follow before converting. Use this to identify common patterns — for example, if organic search consistently appears as an early touchpoint but rarely as the last, first-touch or position-based models better capture its contribution than last-click.
Lookback windows: GA4 defaults to a 30-day lookback window for acquisition conversion events and 90 days for other conversion events. If your sales cycle exceeds these windows, touchpoints beyond the window get excluded from attribution entirely. Adjust the window in Admin > Attribution Settings to match your actual buying cycle.
Common Attribution Mistakes
Ignoring assisted conversions. The most frequent mistake is evaluating channels only by last-click conversions. A channel that appears in 500 conversion paths but only closes 50 of them looks weak in last-touch reporting. Check the assisted conversions report before cutting any channel — if a channel has a high assist-to-last-click ratio, it is driving value upstream that last-touch hides.
Switching models without establishing a baseline. Teams often change attribution models and then compare current numbers against historical data from the old model. This creates phantom trends — channels appear to improve or decline based entirely on the model change, not actual performance shifts. Always run both models in parallel for at least one full sales cycle before switching your reporting baseline.
Treating attribution as absolute truth. No attribution model perfectly captures reality. They are approximations. Data-driven attribution is the most sophisticated option available in GA4, but it still operates within the constraints of trackable touchpoints. It cannot credit a podcast mention, a word-of-mouth recommendation, or a LinkedIn post someone read without clicking. Use attribution data to inform decisions, not dictate them.
Optimizing for a single conversion event. Attribution data changes significantly depending on which conversion you measure. A channel that performs poorly for purchase attribution might excel at driving email signups that later convert. Map attribution across your full conversion tracking funnel rather than a single endpoint.
Frequently Asked Questions
What is the difference between first-touch and last-touch attribution?
First-touch attribution gives 100% of conversion credit to the first interaction a user had with your brand — the channel that introduced them. Last-touch gives 100% to the final interaction before conversion — the channel that closed them. In practice, first-touch overvalues awareness channels and undervalues conversion-focused ones, while last-touch does the opposite. For most teams, neither model in isolation provides an accurate view, which is why multi-touch models like position-based or data-driven produce more actionable insights.
How much data do I need for data-driven attribution?
GA4 requires approximately 300 conversions per month for a specific conversion event to generate a data-driven model. Below that threshold, GA4 falls back to a rules-based model. If you run a low-volume site, focus on your highest-volume conversion event first — such as form submissions or signups — and use position-based attribution for lower-volume events until you build enough data.
Does attribution work across devices?
GA4 uses Google signals and User-ID to stitch cross-device journeys when users are signed in. Without these signals, a user who researches on mobile and converts on desktop appears as two separate users, and the mobile touchpoints get no credit. Enable Google signals in your GA4 property settings and implement User-ID if you have authenticated users to improve cross-device attribution accuracy.
Should I use the same attribution model for all channels?
Use one model as your primary reporting standard to maintain consistency, but analyze performance across multiple models to identify blind spots. The Model Comparison report in GA4 exists specifically for this purpose. If organic traffic shows strong first-touch credit but weak last-touch credit, that tells you organic is an awareness driver — and you should evaluate it accordingly, even if your default reporting uses a different model.