A Data Science Central Community
I think it really depends on what you hope to do with the results.
As long as you take the results as "leads" that require further verification it is not an issue.
Being able to rank results (even with questionable methods/assumptions/sample sizes) can add value over not being able to rank at all.
But for formal statistical inference it is a different issue.