For most, the answer to the question in the header would be ‘Yes’, but I am not sure how.
Google has spent 20 years conditioning us to rely on the ‘Last click’ a customer makes in their purchase journey, and monetising the generation of that last click.
Any marketer worth the label knows that there is a range of factors that influence a buying choice that have little to do with the last click.
That way of thinking was always lazy. Now, with AI search answering questions directly and starting a journey that in my view leads to strangling traditional SEO, it’s also muddle-headed.
Economists have two simple tools that expose how broken your attribution really is: the Lorenz Curve and the Gini Coefficient. Together, they can connect what is happening right now with SEO, AI answers, and the shift to “Answer Engine Optimisation” (AEO).
The Lorenz Curve is a way to show how unevenly something is shared. Economists use it to show income inequality. We can use it to show attribution inequality.
On the horizontal axis you line up all your marketing touchpoints: brand advertising, brochures, display ads, Google ads, word of mouth, showroom visits, social proof, installer van on the street, and so on. On the vertical axis you show how much “credit for the sale” each touchpoint gets.
If all touchpoints shared credit evenly, the curve would be a perfect diagonal line. That says: every channel mattered equally. When you only count that last click, usually a Google ad, everything else in the marketing mix looks like a rounding error. Reality never looks like that.
When you graph that unrealistic assumption that the last touchpoint is the key, the Lorenz Curve bends sharply down and across instead of sitting near the diagonal. The harder that curve bends, the more you are lying to yourself about what led to the transaction.
Now we give that bend a number.
The Gini Coefficient is a number between 0 and 1 that tells you how unequal the distribution is. Zero means “perfectly even”. One means “one thing took it all”.
A low Gini means you’re spreading credit across the full customer journey. You acknowledge brand, reputation, trust, word of mouth, proof, and follow up.
A high Gini means you are giving credit to one or two digital clicks and pretending everything else did not exist.
Call it your Attribution Gini.
When your Attribution Gini is high, you’re under-investing in the work that actually created demand. You’re starving the slow compounding stuff: reputation, perceived quality, remembered expertise, physical presence, referrals. You are funding only the “closer” at the goal line, and not the team that marched the ball 90 metres up the field.
Think about how people actually buy. A neighbour mentions you. They see your installer van parked in a nice suburb. They see your name in a social media group, hear your name in a casual conversation, or meet you in a group of some sort. When they have a relevant issue, there is some level of ‘mental availability’ built up by these non-attributable mentions. So, they visit your website, and probably those of your competitors, then right at the end, they Google your brand name, click a link, and request a quote. Google dashboards give overweighted attribution to that last step in the process. It’s like giving most of the credit for dinner to the waiter who carried the plate to the table, and none to the farmer, the chef, ambience, or location of the restaurant.
We all know it is nonsense, but it is a nonsense we have accepted because it is easy, relies on numbers which pleases the engineers and accountants who run the place, and we did not have an easy alternative.
Consider the following example, a marketing program with seven touchpoints that appeared in a set of successful sales, with the google allocated sales impact.
- Google Ad (last click). Drove 60% of sales.
- Instagram Ad. Drove 15% of sales
- Email follow-up. Drove 10% of sales
- Website research visits/. Drove 5% of sales
- Brochure as PDF download. Drove 5% of sales
- Brand advertising / PR coverage. Drove 3% of sales
- Word of mouth. Drove 2% of sales.
If you graph that on a Lorenz Curve, you get a big bend, as demonstrated in the header. Then you calculate the Gini Coefficient and it’s high. The google dashboard reports that almost all the credit goes to one channel.
That will never feel right, but to date we have run with it.
The migration of search from the familiar SEO ‘tricks’ that suit the last click environment to single answer responses to longer queries is a profound change. So far, the share of LLM generated queries is a small percentage of total searches, 1-3% depending on the source, but is rising at geometric rates.
Those who are successful in the future will figure out how to ensure their brand is returned when an AI initiated search is done. This requires a rethink of the way questions are asked, and puts far more weight on the communication channels currently largely ignored by the old SEO rules. Google now shows an AI generated “overview” at the top of many searches. Chat-style engines like Perplexity, and AI assistants baked into phones and browsers, give a direct answer and cite a few brands as proof. Users tend not to scroll and browse. They ask, they get told, they decide.
LLM’s gives us the ability to dig deeper into the drivers of attribution that we have ever had. We are moving into the world of ‘Answer Engine Optimisation’. Do not be left behind.
