Same Search, Completely Different Results: Google vs ChatGPT Side by Side

We typed "best plumber in Manchester" into Google and ChatGPT at the same time. Google returned its usual mix of local pack results and organic listings. ChatGPT returned a neat list of five businesses with descriptions and links.

Two businesses appeared in both. Out of fifteen total across the two platforms. That's it.

The other thirteen? Completely different. Google's top plumbers don't exist in ChatGPT's world, and ChatGPT's picks are buried deep in Google or missing entirely. Two search engines, same query, two different realities.

Then we tried "best plumber in Didsbury." Didsbury is a suburb of Manchester, about four miles south of the city centre. The results shared zero businesses with the Manchester query. Not two. Not one. Zero. One hundred percent different businesses.

We tested this across Manchester suburbs -- Chorlton (0% overlap), Sale (0% overlap), Stockport (0% overlap). Every single suburb returned a completely separate set of businesses. A plumber who dominates the "Manchester" query might as well not exist when someone in Didsbury asks for help.

The phrasing finding was even stranger. We ran "emergency plumber Manchester" and compared it to "best plumber Manchester." Zero overlap. Not a single shared business. "Boiler repair Manchester" -- also zero overlap with the generic query. "Who fixes blocked drains in Manchester" returned five drain specialists and zero general plumbers.

We tested ten different phrasings for plumbers in Manchester. Across all ten, 55 unique businesses appeared. The average overlap between any two phrasings was 5.1%.

Each query is a separate competition with separate winners. A business checking "do I show up on ChatGPT" by typing one query is seeing roughly 10% of the picture. The other 90% is invisible to them.

This is fundamentally different from Google. On Google, if you rank for "best plumber Manchester," you probably show up for "plumber Manchester" and "top plumber Manchester" too. The results overlap heavily. On ChatGPT, they don't. Every phrasing is its own race, every suburb is its own market, and nobody is monitoring any of it.