GA4 Attribution Models: Which One Should You Use?

GA4 Attribution: What Changed from Universal Analytics

Google Analytics 4 fundamentally changed how attribution works compared to Universal Analytics (UA). If you migrated from UA and assumed everything works the same, your data interpretation is probably wrong.

Key Differences

| Feature | Universal Analytics | GA4 |

|---------|-------------------|-----|

| Default model | Last-click (non-direct) | Data-driven |

| Available models | 7 models | 3 models |

| Attribution scope | Session-based | Event-based |

| Cross-device tracking | Limited | Improved (Google Signals, User-ID) |

| Lookback window | Fixed by model | Configurable (30-90 days) |

| Conversion counting | One per session | Configurable (once or each) |

The shift from session-based to event-based tracking is particularly important. In UA, a "session" was the unit of analysis — a user visits your site, and the session gets attributed to a source. In GA4, each event (page view, form fill, purchase) can have its own attribution, allowing more granular analysis.

For businesses spending $10,000 to $200,000+ per month on paid media, understanding GA4's attribution is critical because it directly impacts how you evaluate channel performance and make budget decisions.

GA4 Attribution Models Explained

Data-Driven Attribution (Default)

Data-driven attribution (DDA) uses machine learning to analyze your actual conversion paths and determine how much credit each touchpoint deserves. It does this by comparing conversion paths to non-conversion paths and identifying which touchpoints are most correlated with conversion.

• GA4 collects all user journeys (both converting and non-converting)
• The algorithm analyzes which touchpoints appear more frequently in converting paths
• Credit is distributed proportionally based on each touchpoint's incremental contribution

• You have at least 300-400 conversions per month (Google's minimum for reliable DDA)
• You run campaigns across multiple channels
• You want an objective, data-backed credit distribution
• You do not have strong opinions about which model "should" be used

• It is a black box — you cannot see the exact calculation behind credit distribution
• Results change as your data changes (not always consistent month-to-month)
• It can be biased toward channels with more data points
• Small conversion volumes produce unreliable results

Paid and Organic Last-Click

This model gives 100% of the conversion credit to the last channel the user interacted with before converting, excluding direct visits. If someone clicks a Google ad, then visits directly, the Google ad gets credit.

• You need simple, consistent reporting
• You want to match how ad platforms report their own performance
• Your business has short sales cycles (1-3 days)
• You are comparing to historical UA data (which used last-click)

• Systematically undervalues upper-funnel channels
• Over-credits bottom-funnel channels (branded search, retargeting)
• Does not reflect the multi-touch reality of most customer journeys

Google Paid Channels Last-Click

This model gives all credit to the last Google Ads click. If there was no Google Ads click in the journey, it falls back to cross-channel last-click.

• You want GA4 conversion data to closely match Google Ads reporting
• You are primarily a Google Ads advertiser

• Obviously biased toward Google's own channels
• Not useful for cross-channel analysis

Configuring GA4 Attribution Settings

Step 1: Set Your Attribution Model

• Go to Admin (gear icon)
• Under your property, click Attribution Settings
• Choose your reporting attribution model:

- Paid and organic last-click

- Google paid channels last-click

The model you select here affects all standard reports, explorations, and the attribution reports section.

Step 2: Set Your Lookback Window

GA4 lets you configure two lookback windows:

• Options: 30 days (default)
• This controls how far back GA4 looks for the first touchpoint that brought a new user

• Options: 30, 60, or 90 days (default: 30 days)
• This controls the attribution window for repeat conversions

| Business Type | Acquisition Window | Other Conversions Window |

|--------------|-------------------|------------------------|

| E-commerce (low AOV) | 30 days | 30 days |

| E-commerce (high AOV) | 30 days | 60 days |

| SaaS / B2B | 30 days | 90 days |

| Lead generation | 30 days | 60 days |

Step 3: Configure Conversion Counting

For each conversion event, decide whether to count:

• Once per session — Counts a maximum of one conversion per session (better for lead generation forms where a duplicate submission is not a new lead)
• Every event — Counts each conversion event separately (better for e-commerce purchases where each transaction is real revenue)

Reading GA4 Attribution Reports

The Model Comparison Report

Location: Advertising > Attribution > Model comparison

This report is one of the most valuable in GA4. It shows your conversion data side-by-side using different attribution models, so you can see how credit shifts between channels depending on the model.

• Select two models to compare (e.g., Data-driven vs. Last-click)
• Look for channels where the difference is significant:

- If a channel gets less credit under DDA than last-click, it may be getting inflated credit from being the final click

| Channel | Data-Driven Conversions | Last-Click Conversions | Difference |

|---------|------------------------|----------------------|------------|

| Paid Social | 245 | 180 | +36% (undervalued by last-click) |

| Organic Search | 310 | 290 | +7% (slightly undervalued) |

| Paid Search (Brand) | 420 | 510 | -18% (overvalued by last-click) |

| Display | 85 | 40 | +113% (heavily undervalued by last-click) |

| Email | 190 | 230 | -17% (overvalued by last-click) |

This data tells you that display and paid social are doing more work than last-click reporting shows, while branded search and email are getting too much credit under last-click.

The Conversion Paths Report

Location: Advertising > Attribution > Conversion paths

This report shows the actual paths users take before converting. Use it to understand:

• Average path length — How many touchpoints does a typical customer interact with before converting?
• Common sequences — Do most customers start with social and end with search? Start with display and end with direct?
• Time to conversion — How many days between first touch and conversion?

The Advertising Snapshot

Location: Advertising > Advertising snapshot

This provides a high-level overview of channel performance under the selected attribution model. It includes:

• Channels ranked by conversions
• ROAS by channel (if revenue tracking is configured)
• Conversion trends over time

GA4 Attribution Pitfalls

Pitfall 1: GA4 Data Sampling

GA4 samples data when your queries exceed thresholds. In standard reports, sampling starts when the data exceeds roughly 10 million events. In explorations, it can start much sooner.

Pitfall 2: Consent Mode Gaps

If you have implemented consent mode (required in the EU and increasingly common in the US), GA4 does not collect full data from users who decline tracking. GA4 uses behavioral modeling to estimate conversions from non-consenting users, but this modeling adds uncertainty.

Pitfall 3: Cross-Device Gaps

GA4 offers three identity methods for cross-device tracking:

• User-ID — You provide a logged-in user identifier (most accurate)
• Google Signals — Uses Google account data for signed-in users
• Device-ID — Falls back to client-side IDs (least accurate)

If most of your users are not logged in, GA4 cannot accurately track cross-device journeys. A user who clicks an ad on mobile and converts on desktop looks like two separate users.

Pitfall 4: Not-Set and Unassigned Channels

If your UTM tagging is inconsistent, GA4 categorizes traffic as "(not set)" or places it in the wrong default channel group. This pollutes your attribution data.

Pitfall 5: Comparing GA4 to Platform Data

GA4 will almost always show fewer conversions per channel than the ad platform itself reports. This is not a GA4 error — it is a feature. GA4 deduplicates conversions and distributes credit, while platforms claim full credit.

| Platform | GA4 Reports vs. Platform Self-Reported |

|----------|---------------------------------------|

| Google Ads | 10-25% fewer in GA4 |

| Meta Ads | 25-45% fewer in GA4 |

| TikTok | 30-50% fewer in GA4 |

| LinkedIn | 20-40% fewer in GA4 |

Advanced GA4 Attribution Techniques

Technique 1: Custom Channel Groupings

GA4's default channel groupings may not match how you think about your marketing. Create custom channel groups:

• Go to Admin > Channel Groups
• Create a new custom channel group
• Define rules that match your campaign structure

Example: Separate "Brand Search" from "Non-Brand Search" by creating rules based on campaign name UTMs.

Technique 2: BigQuery Export for Custom Models

Export GA4 data to BigQuery and build custom attribution models that go beyond what GA4 offers natively:

• Custom weighted models that give specific credit percentages to different channel types
• Revenue-weighted attribution that values high-AOV conversions differently
• Segment-specific models that apply different attribution logic to different customer segments

Technique 3: Combining GA4 with Server-Side Data

Use GA4 as one input in a broader measurement framework:

• GA4 data-driven attribution for cross-channel directional insights
• Server-side tracking data for more complete conversion counts
• Platform-level data for within-channel optimization
• Incrementality tests for causal validation

Technique 4: Attribution by User Segment

Use GA4 Explorations to analyze attribution by user segment:

• New vs. returning users (new users often have longer, more complex paths)
• Mobile vs. desktop (path lengths and channel mixes differ by device)
• High-value vs. low-value customers (your best customers may come from different channels)

Making GA4 Attribution Actionable

Attribution data is only valuable if it changes decisions. Here is how to act on GA4 attribution insights:

• Quarterly model comparison reviews — Compare DDA to last-click every quarter. Identify channels that are systematically over or undervalued.

• Budget reallocation — If DDA shows that paid social drives 30% more credited conversions than last-click indicates, test increasing social budget by 10-15%.

• Path analysis for creative strategy — If most conversion paths start with social video, invest in better video creative for the awareness stage.

• Funnel optimization — If paths show a common drop-off point (e.g., users click social ads but do not return from email), optimize that stage.

If your GA4 attribution data is confusing, inconsistent, or not matching what you see in your ad platforms, your setup may need tuning. Get a free growth audit from Digital Point LLC, and we will review your GA4 configuration, attribution settings, and tracking implementation to make sure you are getting accurate, actionable data.