Why your competitor gets named and you do not
The useful version of this question is not why they were chosen, because nobody can tell you that. It is which sources the answer rested on, and that you can find out this afternoon.
Start where the working is shown
Run the question in Perplexity first, because it cites as a matter of course. Open every source behind an answer that names your competitor and not you. You are not looking for a ranking factor. You are looking for a list of who currently speaks for your trade in your town.
Do it for ten questions and the list stops being a list and becomes a pattern. In most local markets it is three or four directories, one or two review platforms, sometimes a local news site, and rarely anybody's own website.
Then sort what you find into three piles
- Records you control
- Your own profiles on those platforms. If your competitor's listing is complete and yours has three fields filled in, that gap is closable this week and costs nothing.
- Records you can influence
- Review volume and detail, supplier and manufacturer dealer listings, trade body pages. Slower, entirely legitimate, and these are usually where the real difference sits.
- Things you cannot buy
- A local paper writing about them. Twelve years of accumulated mentions. Being the firm everyone's parents used. Naming this pile honestly is what stops a strategy being built on wishes.
What we found when we did it
Captured 12 September 2026 · Austin TX, desktop, logged out
We ran this on one question, how long does a furnace last
, and captured
every source Google's AI Overview used. Eleven references across nine domains: one Reddit thread
cited first and five years old, four manufacturers, three pages from a single filter retailer,
and three contractors last.
Contractors took three of eleven slots on the one question every contractor in the country has blogged about. The competition was Carrier, Trane, Lennox, American Standard and a forum conversation between homeowners. The full capture is here, and it is one observation on one day in one market, which is exactly how much weight to put on it.
The finding that usually lands hardest
Most contractors expect the answer to be their website. It very rarely is. When the cited sources are overwhelmingly third-party records, the implication is that months of blog work is aimed at the smallest slice of the thing being read, while a directory profile that takes an hour to complete is doing more.
A worked version: ten questions produce 63 citations across four engines. Twenty-two go to one review platform, fourteen to a national directory, nine to a local news site, four to contractor websites. The remaining fourteen are scattered. The contractor's content budget was pointed at the four.
What you cannot conclude
Correlation only · the sources are visible, the weighting is not
That fixing the gap will get you named. The source list tells you what was read, not how it was weighted or why one company was chosen over another, and the same question asked again may answer differently. What you have is a well-founded hypothesis and a list of concrete, cheap actions, which is a considerably better position than guessing. It is not a guarantee, and anyone offering one is describing a system they do not have access to either.