A small catalog is easy to keep an eye on. Ten or twenty products, one person can reasonably remember what each listing looks like and notice when something's off. That confidence tends to carry forward even as the catalog grows past the point where it's actually still true.
The breakdown happens quietly
Nobody decides one day that manual checking has stopped working. It just gets slower to notice things: a broken image sits for a month before anyone happens to click into that specific listing, a missing price survives a bulk import unnoticed. The catalog didn't get worse overnight, the amount of it that any one person can actually keep in their head just quietly shrank.
Big catalogs solve this with headcount
Large stores usually have someone, sometimes a whole team, whose job includes reviewing listings for exactly these kinds of issues. That's a reasonable solution at that scale, and also not a realistic one for a small store where the same one or two people are already handling inventory, orders, and support.
The gap that leaves
Small stores end up in an odd spot: too large to catch every issue by memory, too small to justify hiring someone whose job is checking listings. That gap is exactly what an automated scan is meant to fill, running the same handful of checks a careful person would run, at a scale a person realistically can't sustain on their own.