Reading only Google Search Console hides how you are actually found
If you only ever open one search console, you have no idea what fraction of the picture you are looking at.
Three sites, same period, both consoles side by side. The smaller engine delivered roughly a fifth of the impressions and converted them at two to three times the rate, ending up at between 18 and 39 per cent of all search clicks.
None of that was visible from the larger engine's console, which is the only one anyone had been opening.
Why the click rate differs so much
Not because one engine has better users. Result pages differ in how much they answer the query in place, how many paid and generated blocks sit above the organic results, and how far down a given position actually appears. The same position is worth a different amount on each. Comparing your position between engines tells you little; comparing clicks tells you plenty.
The part that matters more
Assistant products ground their answers in a search index, and it is frequently not the largest one. So the smaller engine's index is often the gate to being cited in AI answers.
That makes its console two reports rather than one: ordinary search performance, and observed citations in AI answers, including which queries produced them. That second report is usually the single most useful file available to anyone trying to understand their AI visibility, because it records questions real people asked rather than prompts you chose to track.
What you can miss
One site in that comparison had quietly fallen out of the smaller engine. Daily impressions dropped by around two thirds and stayed there. Its AI citations stopped within days of the same date.
The larger engine showed nothing, because it was still picking a canonical URL and still serving the site. An entire discovery channel went dark for three weeks and the dashboard everyone was watching looked normal.
Verify your site in both consoles, submit the sitemap to both, and pull both when reviewing performance. If you keep a spreadsheet of search numbers, give it an engine column. A single blended number will hide exactly this failure.
Small numbers, careful conclusions
Second engines usually mean smaller absolute figures, and small figures move for boring reasons. Before declaring a collapse, check whether the base is large enough to support the claim. In the case above it was 99 impressions before and 18 after, which is a real change but thin evidence, and worth saying so rather than overstating it.
Need help with any of this?
These notes are free and always will be. If you would rather someone just set it up, or you are stuck on something similar, get in touch at hello@opsira.io.