An attribution model in GA4 defines the mathematical rule used to distribute conversion credit across paid and organic touchpoints along the customer journey. By default, Google Analytics 4 applies Data-Driven attribution, which uses machine learning to estimate how each interaction changes the chance of a purchase. In contrast, the last-click model in GA4 is called Paid and organic last click (also named Last non-direct click). It assigns 100% to the last channel that is not a direct visit, so cold search, video ads, and social campaigns can look unprofitable. Understanding how these models differ prevents small business owners from mistakenly shutting down profitable top-of-funnel campaigns.
Why Marketing Channels Clash Over a Single Sale
In modern marketing, customer journeys are rarely a straight line. Consider a homeowner looking for an HVAC contractor or a buyer shopping for high-end furniture: they might first watch a YouTube review, search Google a few days later, click a retargeting ad on Instagram, and finally buy after a branded Google search. If you evaluate performance strictly by the final click, that branded search gets full credit while every earlier channel receives zero.
According to a 2025 consumer study by Google and Measure Protocol examining more than 22,000 online purchases in the United States, 8 in 10 purchase journeys involve multiple touchpoints. Relying solely on Last Click measurement creates a costly blind spot. Business owners often assume upper-funnel ads are wasting dollars and turn off YouTube or cold search campaigns. They risk a later drop in new visitors. The drop is not guaranteed: compare the account before and after.
When learning how to use Google Analytics 4 to guide your growth, understanding the role of every interaction protects your marketing budget from premature cuts.
Here is how a four-step customer journey is evaluated under different attribution methods:
| Customer Journey Touchpoint | Last Click Attribution | Data-Driven Attribution |
|---|---|---|
| First visit: Cold YouTube video ad | 0% credit | Fractional credit for starting the journey |
| Second visit: Google organic search | 0% credit | Fractional credit for intermediate research |
| Third visit: Social retargeting ad | 0% credit | Fractional credit for re-engaging prospect |
| Fourth visit: Branded Google search | 100% credit | Fractional credit for closing the transaction |
How Data-Driven Attribution Works in GA4
Data-Driven attribution in GA4 uses Google machine learning algorithms to evaluate all user journeys on your website. The system analyzes both paths that ended in a sale and paths where visitors abandoned without taking action. By comparing thousands of journey sequences, the algorithm determines which specific touchpoints statistically increase the likelihood of a conversion.
The core advantage is fractional crediting. As detailed in the official Google Analytics documentation, instead of picking an arbitrary winner, the system assigns each click an influence weight from 0 to 1, summing up to 1.0 per conversion event. For example, if a homeowner finds your plumbing company through local search and later books an appointment via remarketing, search might receive 0.4 conversions while remarketing receives 0.6.
Data-driven attribution is the default model in Google Ads and in GA4. The help pages do not publish a minimum data volume. When an account has little data of its own, the model may use aggregate data from data sharing settings. Rules-based models (First click, Linear, Time decay, and Position-based) have been unavailable since November 2023. Fewer than 3% of Google Ads web conversions used them in February and March 2023, as the Google Ads help page states.
Why Last Non-Direct Click Is Not a Separate GA4 Model
Attribution reports in GA4 offer three models: Data-driven, Paid and organic last click, and Google paid channels last click. Last non-direct click is the second name of Paid and organic last click, not a separate menu item. Direct traffic gets no credit unless the path is only direct visits. Direct gets 100% only when the whole path is direct.
It is equally important to distinguish cross-channel website measurement from ad platform figures. Unlike isolated ad tracking governed by specialized Google Ads attribution models, GA4 measures every touchpoint across search, social, referral, and email. To maintain clean tracking between platforms, business owners must correctly link Google Ads and Google Analytics.
In GA4, reporting attribution settings only apply to event-scoped traffic dimensions such as Source, Medium, Campaign, and Default channel group. User-scoped or session-scoped dimensions follow fixed attribution rules and remain unchanged when you adjust property settings. For retail brands, an accurate GA4 ecommerce setup transmits actual transaction values, enabling machine learning models to allocate real dollars rather than arbitrary scores.
Model Comparison Report: Practical Analysis
To audit campaigns, open the model comparison report: Advertising, then Attribution, then Attribution models. The help page calls this report Key event attribution models. According to Google Analytics guidance, it compares models and shows the percentage change in key events and revenue. It does not publish percentage cutoffs.
A practical audit compares Data-Driven attribution with Paid and organic last click. Read the direction of the revenue shift, not a round number:
- Channels with a clear gain under Data-Driven: These are demand generators at the top of the funnel. They look expensive under last click because that model gives them no credit. A shift in credit does not show that cutting them reduces new customers. Compare the account before and after.
- Channels with a clear drop under Data-Driven: These are usually branded search and direct traffic. They capture demand that other channels already created.
- Channels with almost no shift: Campaigns with a short cycle, where the first and final clicks are the same visit.
Pair this with the attribution paths report. If half of the conversions fall on the first day and the rest are spread across the week, launch remarketing immediately. If a second conversion spike appears between 12 and 30 days, prospects deliberate slowly, requiring long-term retargeting workflows. This clarity underpins effective ad analytics for business owners who focus on total marketing return on ad spend.
Lookback Window: 30 vs 90 Days
The lookback window determines how far back in time an ad interaction can be eligible for conversion credit. If an ad click took place outside this window, GA4 excludes it from attribution calculations. You can manage this setting under Admin > Data display > Events > Attribution settings. The key event lookback window is on that same page.
By official platform rules, acquisition events (first_open and first_visit) use a default 30-day window, adjustable to 7 days. All other key events default to a 90-day window, with options for 30 or 60 days. Keep in mind that changes only apply to future data and do not retroactively recalculate past conversion history.
Choosing between 30 and 90 days depends on your sales cycle. For low-ticket impulse goods, a 30-day window is sufficient. For B2B contracts, home services, dental implants, or legal representation, shortening the window distorts reality. In a multi-stage sales process, initial inquiries arrive quickly, but contracts close weeks later. A narrow reporting window captures only the quick lead and misses the final revenue, making top-of-funnel channels appear far weaker than they truly are.
How to Protect Profitable First-Click Campaigns: Three Rules for Business Owners
In my 19 years in Google advertising (since 2006), I frequently see business owners cut ad spend based strictly on last-click cost per acquisition, unintentionally destroying their customer pipeline. To safeguard your growth, follow three rules:
- Compare models before turning off ads: In the model comparison report, pick a date range and key events that carry revenue. Set Data-Driven beside Paid and organic last click and read the change in events and in revenue. Do not decide on a window shorter than the sales cycle. If the channel shows a clear revenue gain, it feeds the whole pipeline.
- Align reporting windows with your true sales cycle: Do not judge cold search or video ads by the first few days. If your typical buyer takes 20 days to make a decision, the financial impact of first-touch ads will only show up a month later.
- Track total business revenue over isolated channel CPA: Keep your eyes on overall marketing efficiency. If shutting down top-of-funnel ads lowers your branded search cost per click but total business revenue drops, you have cut off the supply of fresh paying customers.
The owner should keep Data-Driven: it is the default model in Google Ads and in GA4. Use Paid and organic last click beside it when the path is short and repetitive, or when data is thin. For a long sales cycle, keep the 90-day window and decide on revenue.
When this does not apply
Despite the analytical advantages of Data-Driven attribution, there are specific scenarios where switching models delivers little practical benefit:
- Emergency services and instant purchases: In emergency plumbing, lockouts, towing, or local food delivery, the customer usually acts at once and does not build a long path. The two models then sit close, and Paid and organic last click is enough as a conservative check.
- Few conversions or broken tracking: Google does not publish a minimum volume. When data is thin, or consent breaks the link between sessions, I do not turn a campaign off on a short window: that is my practice, not a platform rule. Collect key events first.
- No order value: The revenue column in the comparison report is calculated like purchase revenue. Without the order amount you see conversions, not money. Send revenue before you split the budget.
- Offline sales without a CRM feed: If the form is online but the payment stays offline and never returns to GA4, analytics sees only the lead. Do not rank channels by that lead.
If you also run Google Ads, the step-by-step version for business owners is in my book.
My e-book “19 Google Ads Secrets”: how to get the most out of Google Ads without draining your budget, drawn from real campaigns and tests.
See the book →Questions and answers
Does changing the attribution model in GA4 automatically alter Google Ads bidding?
Changing the reporting attribution model in GA4 does not automatically adjust smart bidding inside Google Ads. For your ad campaigns to optimize toward fractional values, you must import the corresponding GA4 key events into Google Ads and select your preferred attribution model directly within the Google Ads conversion settings.
Can I restore the Linear or Time Decay attribution models?
No, you cannot restore Linear, Time Decay, First Click, or Position-based models. Google permanently retired all static rules-based models across GA4 and Google Ads in November 2023 because fixed mathematical percentages failed to reflect modern non-linear consumer journeys.
Why do decimal numbers appear in my conversion columns?
Fractional numbers indicate that Data-Driven attribution is active. The machine learning algorithm distributes a single conversion across multiple contributing touchpoints according to their measured influence, resulting in decimal metrics like 0.45 or 1.25 rather than whole integers.