Why Search Changed

For twenty-five years, search ran on a simple exchange: the engine sent you traffic, and in return you gave it content to index. Generative answers break that exchange. Instead of returning ten links and sending the visitor onward, the engine reads those links itself and hands the user a finished answer. The numbers say it plainly: 68% of US Google searches ended without a click in early 2026 (up from 60% in 2024), and when an AI Overview is present, clicks on the top result drop by roughly 58%.

This is an economic shift, not a headline. The click was the currency that funded the open web; when the answer becomes the destination, that currency deflates: publishers report 70–80% traffic declines on answer-prone topics. But the exposure is uneven. Informational and comparison queries ("what is…", "X vs Y") are absorbed readily, because a paragraph answers them well. Navigational and transactional intent ("acme login", "buy running shoes size 10") still reliably send clicks; nobody buys shoes from a paragraph. Knowing which half of your traffic is exposed is the whole point of this node.

You are not changing a setting; you are changing what you measure and what you promise. "Implementing" this node means finding out how much of your own traffic sits on answer-prone queries, and resetting expectations before the drop surprises anyone.