Search Console is the best public window into Google search, and it is still a sample. For ninety days we compared API rows against server logs for a set of URLs we could hold still. The totals rarely matched. The gap was worse on long-tail queries and better on head terms, which means your “opportunity” list is biased toward what the sample likes to show.
That does not make GSC useless. It makes it a directional instrument. Treat a 4% click change as noise. Treat a step-change that also shows up in analytics as a thing. Put both on the calendar so you are not arguing with a sampled Tuesday.
If a vendor’s dashboard shows more precision than Google gives you, they invented it.