Attribution Report: Types, Models, Example and How to Read One

An attribution report shows which marketing touchpoints get credit for your conversions, such as leads, sign-ups or sales, and for the value attached to them. It takes the paths people followed before converting, applies a rule called an attribution model, and totals the credit by channel, campaign, keyword or landing page. Where a standard report tells you what converted, an attribution report tells you what helped it convert.

This guide covers the report types you will meet in GA4, Google Ads and HubSpot, the models and windows behind them, a worked example with real arithmetic, why the numbers rarely match your other reports, how often to check them, and a template you can copy.

What Is an Attribution Report?

Every attribution report is built from the same five parts. Change any of them and the numbers change, so it helps to know which one you are looking at:

• The conversion. What counts as success: a demo request, a purchase, a new contact, a closed deal.
• The touchpoints. The recorded interactions before it: ad clicks, organic visits, emails, referrals, and in CRMs also meetings, calls and form fills.
• The model. The rule that splits credit across those touchpoints, such as 100% to the last click or an equal share to each.
• The lookback window. How far back before the conversion a touchpoint can be and still earn credit.
• The dimension. What gets credited in the rows of the report: channel, campaign, ad group, keyword, landing page or asset.

The output is a table of credited conversions and credited value per row. Switching models moves credit between rows but never changes the total. If the total moves, you are looking at a different data source or window, not a different model.

Other meanings of “attribution report”

Two other things share the name. In finance, performance attribution explains why a portfolio beat or trailed its benchmark, for example how much came from asset allocation versus security selection. In web development, the Attribution Reporting API was a Chrome feature for measuring ad conversions without third-party cookies, covered further down. This article is about marketing attribution reports.

Attribution Report vs Standard Performance Report

A standard report counts results: sessions, clicks, conversions and revenue by source, usually tied to one session or one ad click. An attribution report looks at the sequence of touchpoints behind each conversion and decides how to share the credit.

You need both. The standard report tells you something changed. The attribution report tells you whether the channel you are about to cut was quietly feeding the one you are about to scale.

Types of Attribution Reports

Platforms slice attribution in two ways: by the conversion being credited, and by the view of the path.

By funnel stage (CRM reports)

HubSpot names its three attribution report types after the conversion they measure. Think of them as the top, middle and bottom of the funnel:

A channel can win the first report and lose the third. Content that creates many low-fit contacts will look strong in contact create attribution and weak in revenue attribution. That gap is the reason to run more than one.

By view of the path (ad and analytics reports)

Dimensions decide what you see in each row. HubSpot groups them into asset (for example landing pages), interaction, deal, UTM and other dimensions such as ad keywords and CTAs. Google Ads lets you switch between campaigns, ad groups, keywords, devices and more.

Attribution Models Used in Attribution Reports

The model is the rule that turns a path into credit. Here is how the common ones split a conversion:

No model is correct. Each answers a different question, which is why model comparison reports exist. For a deep look at the most common one, read last click attribution.

Lookback Window vs Conversion Window

The lookback window (also called the attribution window) limits how far back before a conversion a touchpoint can receive credit. A 30-day lookback ignores a blog visit that happened 45 days before the purchase. Set it too short and channels that start long journeys lose credit. Set it too long and stale touches get credit for decisions they had nothing to do with.

Google Ads separates the two ideas. In its attribution reports you can set a lookback window of 30, 60 or 90 days, while the conversion window is a setting on each conversion action that decides whether a conversion is recorded at all after an ad interaction. A practical rule: look at your path metrics (time lag) first, then choose a lookback window that covers most of your conversions rather than guessing.

Where to Find Attribution Reports in GA4, Google Ads and HubSpot

Google Analytics 4

• Go to Advertising, then under Attribution open Attribution models (model comparison) or Attribution paths (conversion paths), per Google’s GA4 attribution guide.
• All GA4 models exclude Direct visits from credit unless the whole path was Direct. So GA4’s last click behaves like last non-direct click.
• The reporting model and lookback windows live in Admin, under Attribution settings.

Google Ads

• Open the Goals menu, then Attribution. Tabs include Conversion paths, Path metrics, Assisted conversions and Model comparison, per Google Ads Help.
• Conversion tracking must be set up first, and path and credit data older than 2 years is deleted.
• These reports only see Google ad interactions. Organic search, email and other channels are not on the path.

HubSpot

• Go to Reporting > Reports, click Create report, choose Attribution report, then pick the Contacts, Deals or Revenue data source. You can also start from a template or an AI-generated report.
• Configure chart type, one or more attribution models, and dimensions, then add filters such as create date, campaign or interaction source.
• Contact create attribution needs Marketing Hub or Content Hub Professional or Enterprise. Deal create and deal revenue attribution need Marketing Hub Enterprise.

How to Build an Attribution Report in 7 Steps

Whatever the tool, the same steps decide whether the report is useful or decorative.

Step 1: Define the conversion and give it a value

Counting conversions treats a newsletter sign-up like a demo request. Give each conversion a dollar value so the report ranks channels by what they are worth. For leads, a simple version is:

Our guide on how to calculate conversion value walks through ecommerce, lead and sign-up cases.

Step 2: Make every touchpoint identifiable

Attribution can only credit what it can see. Add UTM parameters to every link you control (email, social, partners, QR codes), keep referrers intact through redirects, and record offline outcomes such as calls, meetings and closed deals in a CRM so they can be joined back to the original source.

Step 3: Choose the dimension

Channel is the usual start. For budget decisions inside paid media, use campaign or keyword. For SEO and content, use landing page: search engines hide most keyword data, so the page is the most precise organic unit you can attribute to.

Step 4: Pick a model, and a second one to compare

Use one model as your reporting standard and read a second one beside it. Last click with first touch is the most revealing pair: channels that gain under first touch open journeys, channels that gain under last click close them.

Step 5: Set the lookback window

Cover your typical time from first visit to conversion, based on path metrics, not habit.

Step 6: Add the metrics that drive decisions

At minimum: credited conversions and credited value per row. For paid channels add spend and value per dollar (return on ad spend). For CRM reports add lead-to-customer rate, so a channel that floods the funnel with poor-fit leads does not look like a winner.

Step 7: Visualize and decide

Bars for comparing channels, lines for trends, and a short note under each chart saying what you will do about it. A report that does not change a budget, a page or a campaign is a screenshot, not attribution.

Attribution Report Example, With the Math

The numbers below are illustrative, not data from a real company. A B2B software company tracks demo requests. It closes 20% of demos and its average first-year deal is $3,000, so each demo is worth 20% × $3,000 = $600. In one month it gets six demo requests, worth $3,600 in total.

Credited value by channel under three models:

How each column is built:

• Last non-direct (GA4-style last click): Direct is skipped. Paid search takes demos 1 and 4 ($1,200). Email takes 3 and 6 ($1,200, demo 6 because Direct is skipped). ChatGPT takes 2 ($600). Organic takes 5 ($600).
• First touch: organic opened 1 and 5 ($1,200). ChatGPT, paid social, paid search and email each opened one ($600 each).
• Linear: two-touch paths give $300 per touch, the three-touch path gives $200 per touch. Organic: $300 (demo 1) + $200 (demo 3) + $600 (demo 5) = $1,100. Paid search: $300 + $600 = $900. Email: $200 + $300 = $500. Direct: $300 + $300 = $600.

Now add spend for the two paid channels (also illustrative): $900 on paid search and $500 on paid social.

Paid search is $1,200 / $900 = $1.33 per dollar under last non-direct and $600 / $900 = $0.67 under first touch. Paid social goes from $0 to $600 / $500 = $1.20. Read only the last-click column and you would cut paid social, which started a $600 demo. Read only first touch and you would trim branded paid search, which closed one. The model comparison is the decision, not either column alone.

The same data by landing page

The blog guide never closes a demo, so a last-click report says it is worth nothing. It opened one $600 journey and assisted another. The comparison page converts on the spot and earns the same credit under every model. Look for that pattern, guides as openers and comparison or pricing pages as closers, in your own data before pruning content.

How to Read an Attribution Report

Work through these questions in order:

1. Which rows move most between models? Big gains under first touch mean an opener. Big gains under last click mean a closer. Rows that barely move are the only touch their buyers had.
2. Where is the value, not the volume? Sort by credited value. A channel with fewer conversions can be worth more if its conversions close.
3. Paid or earned? Separate paid search from organic search and paid social from organic social. Blended rows hide which tactic did the work.
4. How much lands in Direct? A large Direct share means the report cannot see part of your marketing. Fix tracking before moving budget.
5. Does it hold further down the funnel? Compare lead-stage attribution with revenue attribution. Channels that create leads but not revenue need a different fix than channels that create nothing.

Why Your Attribution Report Doesn’t Match Other Numbers

Mismatches are normal, but each one has a cause you can check. These are the ones that distort decisions most often:

The Google Ads differences come from Google’s own comparison of attribution reports and the Campaigns page. The rest are tracking problems that hurt every model, which is why it is worth fixing them before debating which model is right.

How Often Should You Check Your Attribution Report?

Check it as often as you make the decision it informs, and never on a window shorter than your time to convert. A rule of thumb:

If the last 30 days look bad for a channel that usually takes 60 days to convert, you are looking at unfinished journeys, not a failing channel.

The Attribution Reporting API Is Something Else (and It Is Being Retired)

Several results for this search describe the Attribution Reporting API, a browser feature from Google’s Privacy Sandbox. It let an ad on one site register a “source,” let the advertiser’s site register a “trigger” when someone converted, and had the browser send either event-level reports or aggregated summary reports to an ad tech server, all without third-party cookies.

Some pages still call it the future of conversion reporting. That is out of date. Google announced on October 17, 2025 that it is retiring the Attribution Reporting API on Chrome and Android because of low adoption, and said feedback from it will inform work on an interoperable attribution web standard. Unless you build ad tech, you never configured it directly, and it does not affect the attribution reports in GA4, Google Ads or your CRM.

Attribution Reports for Organic Search and AI Assistants

Most attribution reports are built around ads, because ad platforms log every click. Organic search and AI assistants are harder: Google and Bing hide most keywords, and assistants such as ChatGPT sometimes pass a referrer and sometimes do not. Two habits keep these channels from being undercounted:

• Attribute to the landing page. The page that started the session is the most precise organic unit you have. Report conversions and value per organic landing page, as in the example above. Our guide to organic conversion attribution covers this report in detail, and SEO conversion tracking covers the setup.
• Separate what you can see from what you cannot. Count AI referrals you can identify, and treat referrer-less visits as Direct rather than guessing. The AI conversion tracking guide explains which assistants send referrers.

If you want this view without building it by hand, SEOConversion reports conversions and their value from Google, Bing and AI assistants at the landing-page level, and leaves visits with no referrer as Direct instead of inventing attribution.

A Copyable Attribution Report Template

Paste this header row into a spreadsheet. Fill one row per channel (or landing page) per model, and keep the settings columns so anyone reading it knows which rules produced the numbers.

Period,Conversion,Value per conversion,Model,Lookback (days),Dimension,Row,Credited conversions,Credited value,Spend,Value per $1,Assisted conversions,Decision
2026-09,Demo request,600,Last non-direct,90,Channel,Paid search,2,1200,900,1.33,0,Keep brand terms
2026-09,Demo request,600,First touch,90,Channel,Paid social,1,600,500,1.20,1,Hold budget; test again

The example rows reuse the illustrative numbers from the worked example. Three rules for the template:

• Credited value per model must add up to the same total. If it does not, a setting changed.
• Fill the Decision column. A row with no decision is a row you do not need.
• Keep one tab per conversion type. Mixing demos and newsletter sign-ups in one total hides which one drives value.

FAQ