AI ate your search traffic.
For years now, digital marketers have been playing a profitable game of painting by numbers.
We built elaborate dashboards, celebrated our ROAS, and convinced ourselves that if we just tweaked the bidding strategy one more time, the leads would keep flowing.
It was a good run while it lasted, but the game is over.
The search engine as we know it is dying in real time. And with it, the performance marketing house of cards is beginning to collapse.
Marketing teams spent a decade optimising our way into only the most efficient of campaigns, and the cost has finally arrived.
Performance marketing was built on a very simple assumption: people search, they click, and they convert. Our one job was to intercept them at that perfect moment.
But what happens when that moment is removed? To say we are moving rapidly from an era of search engines to an era of answer engines would be a mistake, the transition has already been made.
When a user asks ChatGPT or Claude a question today, they don’t get ten blue links. They get a synthesised, definitive answer.
The user has what they need. They close the tab. The click never happens.
Search traffic is not just down, it’s disappearing. If your strategy was built on intercepting that traffic until they converted, you are fishing in an ever-shrinking pond.
If you check your performance dashboards over the next 6 months to find your branded search terms have grown leaps and bounds ahead of your other campaigns, you are witnessing this shift in action. Traditional search is becoming nothing more than a navigational tool, but you can only capture brand searches for demand that already exists.
Building it is not a job for a bidding strategy. It never was.
This shift changes everything about how brands are discovered.
On traditional search engines, even if you ranked on page two or three, you still existed. You could buy your way to the top of page one if you had the budget. There was always a chance to catch a stray click.
In the world of answer engines, visibility is binary. When a consumer asks an LLM to recommend a product or service, the AI provides an exclusive shortlist of three to six options.
As it stands, you can’t buy your way into these recommendations. Your presence in the answer is based on your presence in the training data. Brands that are widely discussed, culturally relevant, and highly recognised have the advantage.
