Last Click Attribution: How It Works, Examples and GA4 Behavior

Last click attribution is a model that gives 100% of the credit for a conversion to the last channel the buyer clicked before converting. Every earlier touchpoint, such as the blog post that introduced your brand or the social ad that brought them back, gets nothing. It is simple and good at showing which channels close sales, but it consistently undervalues the channels that start journeys.

This guide explains how the model works, how it compares to first click, multi-touch and data-driven attribution, what Google Analytics 4 actually does (it is not quite pure last click), and how to tell when the report is misleading you. It includes a worked example with real arithmetic and a template you can copy.

What Is Last Click Attribution?

Last click attribution, also called last touch or last interaction attribution, is a single-touch attribution model. It looks at the path a buyer took before a purchase, sign-up or lead form, finds the final click, and assigns the whole conversion and its value to that click’s channel, campaign or page.

The logic behind it is an assumption: the last thing someone did before converting was the thing that made them convert. Sometimes that is true. Often the last click is just the most convenient door into a decision that was already made.

How last click attribution works, step by step

Picture one buyer for an online furniture store:

1. Monday: they read your guide on choosing a desk chair after a Google search.
2. Wednesday: they click an Instagram ad and browse two products.
3. Friday: they sign up for your newsletter to get a discount code.
4. Sunday: they click the discount email and buy a chair.

Under last click, the email gets 100% of the sale. The guide that started the search and the Instagram ad that brought them back get 0%. If you only read last click reports, your content and your social ads look like they sell nothing.

Why Marketers Still Use It

Last click has survived decades of criticism for practical reasons:

• It is easy to set up and explain. You only need to know where the converting session came from. Anyone in the company can follow the logic.
• It needs little user tracking. Because only the final session matters, you do not have to stitch together one person’s visits across weeks and devices. That makes it lighter on privacy than user-level multi-touch models. It still depends on a referrer or UTM tags on the converting visit.
• It shows the closers. For branded search, retargeting, cart-recovery email and promo campaigns, the last click is a reasonable signal of what tipped the purchase. Bid adjustments and creative tests on those campaigns work fine with it.
• It is consistent. The same rule applies to every conversion, so week-over-week trends are comparable even when the absolute credit is off.

The Drawbacks of Last Click Attribution

• It ignores everything before the final click. Discovery and consideration touchpoints get zero credit, no matter how many there were.
• It moves budget toward closers. Branded search and retargeting harvest demand that other channels created. Cut the upstream channels because they “don’t convert” and the closers eventually have less to close.
• It undervalues brand and content. Blog posts, video, social, PR and awareness campaigns rarely get the last click, so they look worthless in the report.
• It breaks down in long sales cycles. In B2B, a deal can involve months of webinars, emails and sales calls. Crediting the final form fill to a single click says little about what drove the deal.
• It cannot see clickless influence. Podcasts, TV, events, word of mouth and ad impressions shape decisions without producing a click, so a click-based model can never credit them.
• It is fragile. Cross-device journeys, browser privacy features and missing referrers mean the “last click” your tool sees is sometimes not the last click that happened.
• Every platform has its own version. Google Ads, Meta and your email tool each report conversions from their own data, and some count ad views as well as clicks. Add their numbers together and you often get more conversions than you actually had. Your analytics tool is the referee: it picks one source per conversion.

Last Click Attribution Example: One Set of Orders, Four Models

The fastest way to understand last click is to run the same orders through several models. The five orders below are illustrative, not data from a real store. Total revenue is $1,000.

Here is how each model credits revenue by channel:

How the numbers are built:

• Last click: B and E end on Direct, so Direct takes $300 + $250 = $550. Email takes A ($200). Paid search takes C and D ($150 + $100 = $250). Organic search and ChatGPT get nothing.
• Last non-direct: Direct is skipped, so B goes back to ChatGPT ($300) and E goes back to organic search ($250).
• First click: organic search opened A and E ($200 + $250 = $450), ChatGPT opened B ($300), paid social opened C ($150) and paid search opened D ($100).
• Linear: each touch gets an equal share. In E, organic search gets $250 / 3 = $83.33 and Direct gets two shares, $166.67. Organic search in total: $100 (A) + $50 (C) + $83.33 (E) = $233.33.

Three lessons from this small table:

1. The total never changes. Switching models does not create or delete revenue. It only moves credit between channels. If your totals change when you switch, you are looking at a different data source, not a different model.
2. Organic search swings from $0 to $450 depending on the question you ask. Under pure last click, the channel that started two of the five purchases looks worthless.
3. Direct is where credit goes to hide. $550 of the $1,000 lands on Direct under pure last click. That is why most analytics tools, including GA4, skip Direct when another channel is available.

Last Click vs First Click Attribution

First click is the mirror image of last click. It gives 100% of the credit to the touchpoint that started the journey. Neither is “right.” They answer different questions.

A practical habit: judge each campaign with the model that matches its job, then look at the other view before you cut its budget. A channel that gains a lot under first click is an opener. One that gains under last click is a closer. One that barely moves is the only channel its buyers ever touched.

Last Click vs Multi-Touch and Data-Driven Attribution

Multi-touch attribution splits credit across several touchpoints instead of giving it all to one. Here are the common models side by side:

The trade-off is always the same. Multi-touch models need you to connect a person’s visits over time, which privacy features, cookie limits and cross-device behavior make harder. Last click needs only the converting session, so it is more robust but less complete. Data-driven models are only as good as the paths they can see.

Does Google Analytics Use Last Click Attribution?

Several guides call last click the default in GA4. Google’s own documentation says otherwise, and the details matter when you read reports.

• The default for key events is data-driven. Event-scoped reports (Source, Medium, Campaign, Default channel group in key event and conversion reports) use the model you select, and by default that is data-driven.
• The last click option is “paid and organic last click.” GA4 offers three models: data-driven, paid and organic last click, and Google paid channels last click. First click, linear, time decay and position based were removed in November 2023.
• Direct is excluded. Per the same page, all GA4 attribution models exclude Direct visits from credit unless the whole path to the key event was Direct. So GA4’s “last click” behaves like last non-direct click: in the example above, it would match the “Last non-direct” column, not the “Last click” one.
• Session and user dimensions always use last click. Session source / medium (Traffic acquisition) and First user source / medium (User acquisition) use paid and organic last click and do not change when you switch the reporting model.
• Lookback windows are limited. Acquisition key events default to 30 days (or 7). All other key events default to 90 days, with 30 or 60 as options. Model changes apply to historical data, while lookback window changes only apply going forward.

In practice, this means two GA4 reports can show different numbers for the same channel in the same week. A Traffic acquisition report uses session-scoped last non-direct click. An Advertising or key event report may use data-driven. Neither is broken. Check which scope a report uses before comparing.

What Last Click Does to SEO Content and AI Assistant Traffic

Last click is particularly harsh on two channels that tend to start journeys rather than finish them.

Informational SEO pages

Someone who finds your “how to choose” guide on Google is usually early. They often come back later by searching your brand, clicking an email or typing your URL. Under last click, the guide gets nothing, and the conversion is credited to branded search, email or (outside GA4) Direct. Product, pricing and service pages that convert in the same visit look fine. Blog content looks dead.

The fix is not to abandon last click but to read it at the landing-page level. Ask: which organic landing pages start sessions that convert, and how much is each conversion worth? Then compare that with a first-click or assisted view for informational pages. Our guide to organic conversion attribution covers landing-page reporting in detail, and SEO conversion tracking covers the setup.

ChatGPT, Perplexity and other assistants

AI assistants are often a research step: someone asks ChatGPT for options, clicks through to your site, leaves, and comes back later to buy. If the assistant passed a referrer, the first visit is identifiable as AI traffic. Under pure last click, the purchase still goes to whatever came last. Order B in the example shows it: ChatGPT gets $0 under last click and $300 under last non-direct or first click.

It gets worse when the assistant sends no referrer. That visit arrives as Direct, and because GA4 skips Direct when anything else is in the path, the AI visit can disappear from attribution entirely. No model can credit a source the tool never saw. To see what is identifiable, read AI conversion tracking.

When the Last Click Is Wrong: Failure Modes and How to Fix Them

Before you argue about models, make sure the last click is being recorded correctly. These problems steal credit from the real channel under any model, but last click suffers most because it trusts one session.

Payment processors are the classic case for GA4’s unwanted referrals list. If one of them is your biggest “channel,” your last click report is crediting checkout, not marketing.

When Last Click Is Good Enough: A Decision Rule

What to Use Alongside Last Click

• A second attribution view. Put last click next to first click (or data-driven) for the same period. Large gaps show which channels open and which close.
• Incrementality tests. Pause a campaign in some regions or for a period and compare against a control. This is the only way to measure what a channel actually causes, rather than what it is credited with.
• Marketing mix modeling. A statistical model of spend against sales over time. It needs no user-level tracking and catches offline channels, but it needs months of data and is usually worth it only at larger budgets.
• Self-reported attribution. A “how did you hear about us?” field on forms or a post-purchase survey picks up podcasts, word of mouth and AI recommendations that leave no click.

Last Click Attribution Template

If you want to build last click and first click views in a spreadsheet from exported paths, use one row per conversion and these columns:

Fill touches left to right with no gaps, then use these formulas (they work in Google Sheets and Excel):

For a last non-direct column, take the last touch that is not “Direct,” and fall back to Direct only when every touch is Direct. That reproduces how GA4 treats Direct.

Turning Last Click Into Revenue by Channel and Page

Attribution only helps if each conversion carries a value. A lead form, a demo request and a purchase are not worth the same, and counting them equally makes every model misleading. Assign a value to each conversion type (see how to calculate conversion value), then report value by channel and by landing page. SEOConversion does this for organic search and AI assistants: it reports conversions and their value by landing page, and when an assistant sends no referrer, the visit stays Direct rather than being guessed.

FAQ