Attribution

advancedwidespread

Attribution is determining which ad, search, article, or AI answer led to a purchase or inquiry, and how much value each contributed.

In plain terms

Imagine your team scores a goal. The scorer's name appears in the newspaper, even though five people passed the ball before that, and the goalkeeper started the attack. Attribution in marketing asks the same question: who deserves the credit? The one who last touched the ball, or everyone who participated in the action?

Before you buy something, you usually encounter the company multiple times: you see an ad, search for it, read a review, receive a newsletter, and finally place an order. Attribution tries to trace this path so the company knows which channel is worth spending on and which is not.

In the AI era, the concept has two new layers of meaning. First, if a chatbot recommends a product to you, the company often cannot see that you came from there, so the impact of AI is hard to measure. Second, in AI answers, attribution also refers to the system indicating which source it took a claim from.

How it works

The company records touchpoints: which ad you clicked, which search brought you in, which newsletter you opened. It uses cookies, tracked links, and login data for this.
It stitches interactions into a single customer journey, preferably across multiple devices (phone, laptop).
An attribution model distributes credit. "Last click" assigns everything to the final step, "first click" to the first, multi-touch models share it, and the data-driven model statistically estimates how much each step contributed.
Results are compared with expenditures: which channel generated the most revenue per forint spent.
Where tracking fails (e.g., during conversations in an AI assistant or when cookies are rejected), estimates are made using surveys, controlled experiments, or statistical models.

An everyday example

You ask an AI assistant which robot vacuum cleaner is worth buying, it recommends a brand, and the next day you type the name into a search engine and order it. In the online store's statistics, this appears as a visit from a search engine, with no trace of the AI's recommendation. This is why stores increasingly ask at the end of the order: 'Where did you hear about us?'

Who is affected

marketerBased on attribution, it decides how much money goes to which channel, thereby validating the results of its own work. → Marketing

entrepreneur, executiveIf it relies on a poor model, it can easily shut down the channel that actually brings in customers.

developerThis person implements measurement code, consent management, and data linking, meaning the reliability of the measurement depends on them. → Code

everyoneAttribution tracks your clicks and browsing path, so it matters what you consent to.

A common misconception

MythThe attribution report shows exactly which ad led to the customer's purchase.

In realityAttribution is an estimate, not proof: it only sees what can be measured, and correlation is easily mistaken for causation. The impact of a friendly recommendation, a podcast, or an AI response typically remains outside it.

Limits and risks

Cookie restrictions, ad blockers, and data protection regulations mean a growing portion of the customer journey is invisible.
Recommendations in AI assistants and AI search engines often work without clicks, so traditional measurement cannot track them.
Different models declare different winners from the same data, so model selection is a business decision, not an objective truth.
Advertising platforms' own reports tend to claim credit for themselves, so the same purchase may be recorded by multiple platforms.

Similar, but not the same

conversion tracking – Conversion tracking records that a purchase or other desired action (e.g., subscription) occurred; attribution indicates what it was due to.
source attribution – Source attribution shows where the AI got its claim from, while marketing attribution shows which channel brought in the customer.
marketing mix modeling – This estimates channel impact from aggregated spending and traffic data without individual tracking, whereas attribution starts from individual customer journeys.

Related concepts

Broader

marketinganalitika, performance measurement, digital marketing

Narrower

Last-click attribution, multi-touch attribution, data-driven attribution

On the edge

conversion, customer journey, source attribution, generative search engine optimization, AI visibility, zero-click search

A brief history

The question is older than the internet: the saying that half of advertising spend is wasted, but no one knows which half, is attributed to 19th-century American merchant John Wanamaker. With the spread of online advertising, clicks became measurable, and for a long time the last-click model was the default. From 2021, Google gradually made data-driven attribution the default in its advertising and analytics systems, and in 2023 removed most rule-based models; only the last-click model remained selectable.

Where things stand

Attribution is a standard practice in digital marketing, but by the end of 2026 it is transforming: due to AI search engines, chatbots, and agents purchasing on behalf of users, more decisions are made outside measurable clicks. Companies are therefore supplementing tracking with experiments, surveys, and statistical models.

Status: 2026 November

In the news

Appears in 7 Agent Age stories; the curve shows weekly occurrence in the news magazine.

Perplexity blocks Time's ads aimed at AI agents, calling them deceptiveNov 2, 2026 · Marketing
Hotels lose attribution as AI-driven zero-click search erodes guest trackingOct 19, 2026 · Marketing
OpenAI introduces visual ChatGPT ads in image generation and expands ad measurementOct 5, 2026 · Marketing
Semmelweis University faculty rules treat use of AI content without attribution as plagiarismOct 1, 2026 · Education & Family
Measurement and budgets: CFO pressure, marketing mix modelling and B2B traffic loss to AI searchSep 22, 2026 · Marketing
Hungarian forums on health AI: nurses' congress, MTA symposium and FESZ programmeSep 21, 2026 · Health
IAB develops AI ad measurement framework that could redraw attributionAug 24, 2026 · Marketing

Also known as

Attribution · marketing attribution · attribution model · attribution modeling · attribution · marketing attribution · multi-touch attribution · MTA

This entry was machine-edited and has undergone machine expert review; the “In the news” section updates automatically from Agent Age stories.