I’ve sent 100+ email by match.com. What is so frustrating is the lack of feedback. I can tell if the email has been read, if my profile was viewed or if there was hesitation before they rejected me. Simply reading can be predicted by the last time they logged on and if they have a picture. But being an active account doesn’t mean they will respond positively. In fact, so far, I’ve gotten about a dozen profile views and 8 canned rejections.
To use quantitative models to improve the situation, I need some positive feedback. In other words, you can’t explain 100 no’s, but you can explain a small number of yes’es mixed in with a large number of no’s, say by using multiple regression.
So far the closest I can get to measurable positive feedback is a profile view or a delay between reading the email and sending me a rejection, and even there there isn’t much data to work with.
I tried using binomial distributions to model the situation. If I have a random number generator that returns 1′s and 0′s and it returns 100 zeros in a row, then I can ask myself, how likely is that given that the random number generator is calibrated to return a 1, say 5% of the time? I’m not sure if I did my math right, but 100 zeros in a row is what you could expect 95% of the time if the random number generator was calibrated to return a one 1 in 1000 times. In other words, if you get 100 no’s on match.com, if you don’t change your strategy, it will take 1000+ emails to get a single positive response.
There in lies the problem again, how can you improve without some positive feedback? One doesn’t even get feedback on how vehemently one has been rejected. Somehow, asking people to respond if the hate me with the fury of one thousand suns or only two hundred seems counterproductive.
Sigh. Maybe that is what match is all about, ripping off people who don’t have a good feel for statistics.