Because they are one page with the town name swapped forty times. That is a scaled-content pattern, and the usual outcome is not that some rank badly — it is that the whole folder gets discounted at once.
The pages that do work carry something genuinely local: real work done there, real prices, real availability.
Each page has its own record with its own search performance, so the twelve locations doing the work and the twenty-eight doing nothing are separable rather than averaged into one line.
Cannibalisation detection catches nearby towns competing with each other, which is the most common self-inflicted problem in local SEO.
It does not manage business listings, reviews or map profiles, which is a meaningful part of local search and lives elsewhere.
And it will not make forty thin pages rank. It will show you which twelve to invest in and which to consolidate.
Source tracking carries the page through to the enquiry and, where you record the job in Books, to the revenue — so a location page can be judged on jobs won rather than on impressions.
Technical audit and metadata checks catch the ordinary failures that quietly cost local rankings.
Products, not integrations. Each one reads the same record, so a join is a permission rather than a sync job with a mapping screen behind it.
Yes, per page rather than averaged across the folder.
Yes, cannibalisation detection covers exactly that.
No. That is a separate part of local search.
Half an hour on your own numbers is usually enough to say whether Growth is the right place to start.