Forum Moderators: LifeinAsia
Is there ANY way to ensure the accuracy and correctness of email addresses on an ongoing basis? (Is there any way to "ping" that specific email address to ensure its validity?) Does anyone have any practical experience in handling a lot bad email addresses and how they are processed if they are invalid? In addition to that, what do sites do when they send out communications and they're inundated with auto-replies?
For example, how would the likes of Yahoo, GoGuides, JoeAnt handle a communication to all listed sites within the directory. If just 5% of the email addresses are bad, and 5% generate an auto-response, how do these sites manage all these incoming bounced/auto emails?
If its a bad email, we want to obviously try and fix the bad email so that we can continue to stay in contact with the company. If its an auto-reply, we'd (ideally) not like to receive it in the first place. If we have to receive the auto-reply, then we at least know it is a valid email address, so maybe we could just filter out these emails somehow.......?........
One thing that we are doing is when the company *initially* submits their information, we check for a proper email format (i.e, my@email.com) and we also check the domain name of their email adress to ensure its validity. We do all this upon their submission to our site, but how can we ensure the correctness of their address on an ongoing basis?
I hope this makes sense. Our current problem is that we receive a lot of bad email addresses, so if we send out a communication, we get a lot of bounced emails. The auto-replies just compound the problem...
Any thoughts would be greatly appreciated. Hopefully other individuals might have the same issues as myself.
After that, all you can do is to check for bounces, and remove/research/correct all those that return an error when you send your messages.
In theory, the SMTP protocol has a provision to ask the respective mail server for the validity of an address, but I wouldn't expect any server to actually still allow this, as it opens them up to address harvesters and other privacy violations.
Is there any way to "ping" that specific email address to ensure its validity?
Sure - there are lots of programs that do this, and some are free. You could run one occasionally to cull bad addresses. Worldcast has free a component that checks for valid addresses, and mailutilities.com offers a couple of programs to do this. I am sure there are others.
You can also set your autoresponder to delete certain addresses after X number of bounces.
All other methods are unreliable at best, or more likely completely useless. The way that most mail servers are configured nowadays, the validity of any particular address can't be determined by any less obstrusive method.
The SMTP standard knows about the VRFY command, which was originally meant for such "ping" purposes. But if you find any mail server that actually implements this today, then you should inform the maintainer of that server that he is jeopardizing the privacy of his users.
Sorry to sound so negative, Bradley. But there is no easy way out of your problem.
Double-opt in sounds like an avenue we might explore. One thing we are doing is making the user input their email address into a box, and then RE-input in another. Hopefully that will ensure they type it in correctly.
If it does bounce, I guess that means we will be giving them a telephone call.
There is no way to reliably determine whether an e-mail address is valid other than sending a unique token to that address, and have the owner return that token to you
That just sounds so high tech and cool. We use an old fashioned system - they send us an e-mail saying they want to subscribe, we send them one back saying "Are you sure," and they reply (hopefully) "Yeah, I'm sure." Then they go on the list.
Your unique token double opt-in thing sounds really cool, though. Maybe you could post some sample token code? ;)
The unique token is just a way to automate the process, so that the software can determine whether the reply is genuine. In practise, it's simply a random string appended to the subject header. Any decent list management software (majordomo, mailman, etc.) will do this for you.
What you're describing *is* double opt-in as well. Just that if you only get back an informal "yeah, I'm sure", you then need to process each subscription manually. That obviously works too, although it doesn't scale very well. The details how to manage double opt-in subscriptions without human interaction are just technicalities.
