Ok, I admit it did take me about 20 mins to figure this out … but the search results are correct.
The bio field is saved as HTML — when sending a search query to the database, the database does not ignore HTML tags and such. So, when a search is performed, class names, image file paths, link URLs will also be searched and returned as hits it they contain a match.
In the case of
medium this is matching a class name of medium applied to the WP images added in the bio field. So, they are returned as matches.
This is not a bug or Connections limitation. Take for example this search applied to the whole site using the core WP search:
The second result is this page:
The word medium is not found in the page text but it applied as a class name to an image so it is returned as a result.
The same this for these pages:
In fact only the first search result has medium in the text body. All the other on the first page of results did not. I did not check the remaining search result pages.
I hope this explains it clearly enough that you can explain it to your client.
Hope this helps!
ps. I did search to see if HTML could be ignored, it can not. The only viable solution would be to save the bio twice. The second instance would have HTML stripped. It is this instance that would then have to be searched instead. Implementing this would be a considerable challenge at this point and not that much benefit.