|
|
|
|
|
|
by paganel
4891 days ago
|
|
|
False positives always suck, and for good reason. The thing is that it's not that easy to get it right, no matter how good the algorithms or the data-sets are. Just in the last two days I've had gmail decide to tag as spam two messages received from different people, with whom it is true I had not communicated before via email, but whose messages had nothing spammy about them and were of great interest to me. And now I'm left wondering if a email that I sent to a third person a week or so ago wasn't tagged as spam as well. That person has not responded yet, and I'm reverse-engineering in my head if maybe me including two links to Imdb in the said email could have been enough to tag it as spam, or if I finally decide to follow up with a second email and probably
making a fool of myself would it be wise to in include the word "spam" in the message, as in "Hey, did you happen to check your Spam folder?" Like I said, false positives are a bitch, for both users and developers |
|
|
So how does the company HANDLE false positives? Perhaps they send out a notice which includes contact information and requests specific information be provided for an "appeals" process of some sort (ideal). Perhaps they say "email us if you object" and then mostly maintain radio silence unless you happen to know somebody (Google). Perhaps they delete the data automatically so there is no possible way to recover (maybe Meetup?).
False positives are a problem and I'm not going to get overly upset with a company when a false positive is triggered. But if you pretend that your process is perfect and do not ALLOW for the possibility of false positives, then I have a problem working with you.