sqldriver (9/2/2014)
Ed Wagner (9/2/2014)
sqldriver (8/28/2014)
The email addresses I was processing were not from a web form, they were provided to us either by clients, or by data marts. Nothing was deleted, it was just given a 'status flag' that meant the email inspection found something wonky. A certain percentage of the contact information was expected to be inaccurate or incorrect. So if the project was running a long time without a high response rate, and the PM didn't want to buy more sample, the remainder email addresses could be released.I had to do a lot of similar stuff with phone number cleaning/checking and formatting so the dialer could make the calls correctly.
Market research is a bummer.
Hey, you know it's your fault that the invalid email addresses and phone numbers don't work. Because they're invalid, that shouldn't impact the response rate, should it? Your situation is certainly not unique. Guess who else works in market research? 😀
What do you mean you don't know if these Thai phone numbers are cell phones? We can't have cell phones on the predictive dialer! CASRO will exile us!
Yes they will, unless you have permission to use cell phone numbers. I think that predictive dialers are falling out of favor anyway industry-wide, given all the restrictions. There are many little quirks with market research data. And watch where (geographically) that data is stored if you any business with European respondents.
I'm in the process of writing a decent email validation, but I'm a ways away from having anything yet. I will not be using the regular expression solution, as we already have one in place and it's a nightmare. I see phase 1 in my head, so it's just a matter of getting the time to write it, get it working and making it fast. Ideally for phase 2, I'd like to "call up" the email server, ask if the account lives there and store the server's response. I don't see that part yet, but one thing at a time.