COVID-19 Outbreak (Update: More than 2.9M cases and 132,313 deaths in US)

This is true.

But also it's a high standard to say that those who believe the observed conditions call into question the official data, must be able to support it with their own data. In that context, anecdotal evidence can be useful, if it is from a credible source. If the same anecdote appears in multiple locations, it can be a basis to draw conclusions. Yes, they're caveated conclusions of course, they're not necessarily reliable - but not dismissible out of hand.

But you're right that it's not data and so it's only so useful. It's all very challenging for sure.
This is true. However, the more degrees of separation in the anecdotal chain, the less it should be considered reliable information. If a medical professional shares anecdotal evidence with me that numbers are being under reported, that's much more reliable than someone saying they know someone that shared the anecdotal evidence with them, and that's more reliable than someone who says they know someone who knows someone that someone else showed anecdotal evidence.

Unless one sees the anecdotal evidence themselves, it's all hearsay. That doesn't mean dismiss it out of hand, but it also means don't accept it as true out of hand.