Last-Click Attribution Is a Participation Trophy. Here Is What to Measure Instead.

Google's AI Overviews now appear on more than 30% of search queries. ChatGPT search is pulling traffic that never touches a website. Meta's Advantage+ campaigns optimise across placements that traditional attribution cannot track. The measurement infrastructure most marketing teams rely on was built for a world that no longer exists.

Last-click attribution assigns 100% of the credit to the final touchpoint before a conversion. It was always a simplification. Now it is a distortion. When a prospect reads three blog posts, sees a LinkedIn ad, gets an AI-generated summary of your product, and then clicks a branded search ad to convert, last-click says the branded search ad did all the work.

The problem is accelerating. Rand Fishkin's SparkToro data shows that zero-click searches now account for roughly 60% of all Google queries. Users are getting answers without ever visiting a website. If your attribution model only counts clicks, you are blind to more than half of your search visibility.

Google's own response has been to push marketers toward data-driven attribution (DDA) inside GA4, which uses machine learning to distribute credit across touchpoints. It is better than last-click. It is not a solution. DDA still operates within Google's walled garden and cannot account for touchpoints it does not observe.

The more honest frameworks emerging from measurement consultancies like Measured and Analytic Partners focus on incrementality testing and media mix modelling (MMM). These approaches ask a different question: not "which click gets credit" but "what would have happened if we had not spent this money?"

MMM is not new. CPG brands have used it for decades. What is new is that the tooling has become accessible to mid-market companies. Google's Meridian (open-source MMM), Meta's Robyn and several commercial platforms now make it feasible for teams spending $50K per month or more on media.

Attribution shapes budget allocation. If your model says branded search drives 70% of conversions, that is where the money goes. But branded search is capturing demand that content, social and display created. Cut those channels based on last-click data and watch branded search volume decline three months later.

For Australian businesses, this is particularly relevant. The market is smaller, budgets are tighter, and the cost of misallocating spend is proportionally higher. Getting measurement right is not a nice-to-have. It is the difference between scaling what works and doubling down on what merely takes credit.

Start with a simple incrementality test. Hold out a geographic region or audience segment from one paid channel for four weeks and measure the difference in conversions. Run a basic MMM using Meridian or Robyn with 12 months of spend and conversion data. At minimum, switch GA4 to data-driven attribution and stop making budget decisions based on last-click reports. The goal is not perfect measurement. It is less wrong measurement.