In my case, at this very moment, I'm going through antidepressant withdrawal and have random crying spells. Sure, depression might not be mediated by serotonin, but these drugs have an effect on mood. It might be that there's a different mechanism of action.

When they test drugs, they test them against a placebo.

Yes, but according to Irving Kirsch, the benefit of anti-depressants above placebo is very small for anyone who isn't severely depressed, and the observed small effects can plausibly be explained by patients unblinding themselves (due to the presence of other side effects in the treatment group), or by things like publication bias.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4172306/

I'll add that side effects alone can influence depression scores. Any medication that de- or increases sleepyness will change the score of any questionaire when taken at the appropriate time of day.