Archives
April 2008
4/29/08
4/8/08
4/3/08
A Quick Note on Invalid Recipients
4/2/08
Thursday, April 3, 2008
9:26:00 AM EDT
Invalid recipients are a fact of life for any mailer. You can minimize them with confirmed/double opt-in, but you still have typos and people who change their email addresses. We get that - it's cool.
At the same time, invalid recipients are used industry-wide as a metric of how good a mailing list is, and high invalid recipient numbers are associated with unsolicited mail and namespace mapping. That's not cool.
We've always factored invalid recipients into a sender's reputation, but we want to give you a heads up that we're refining the process a bit. As such, we would like to remind mailers that high numbers of invalid recipients is a Bad Thing and will result in poor delivery and removal from the whitelist.
Now, I know you, and you are asking, "What's a high number?" That's easy. A high number is a number higher than that which we consider indicative of a good mailer. If your invalid recipient stats are consistent with mailers similar to you, this shouldn't matter much to you. If you generate more invalid recipients than mailers similar to you, you will see a negative impact on your delivery in the near future.
David
Systems Programmer, AOL Postmaster Team
Written by davidzakar Blog about this entry |
Add to del.icio.us | 
digg this
9:26:00 AM EDT
A Quick Note on Invalid Recipients
Invalid recipients are a fact of life for any mailer. You can minimize them with confirmed/double opt-in, but you still have typos and people who change their email addresses. We get that - it's cool.
At the same time, invalid recipients are used industry-wide as a metric of how good a mailing list is, and high invalid recipient numbers are associated with unsolicited mail and namespace mapping. That's not cool.
We've always factored invalid recipients into a sender's reputation, but we want to give you a heads up that we're refining the process a bit. As such, we would like to remind mailers that high numbers of invalid recipients is a Bad Thing and will result in poor delivery and removal from the whitelist.
Now, I know you, and you are asking, "What's a high number?" That's easy. A high number is a number higher than that which we consider indicative of a good mailer. If your invalid recipient stats are consistent with mailers similar to you, this shouldn't matter much to you. If you generate more invalid recipients than mailers similar to you, you will see a negative impact on your delivery in the near future.
David
Systems Programmer, AOL Postmaster Team
Written by davidzakar Blog about this entry |
Add to del.icio.us | digg this
This entry has 6 comments: (Add your own)
-
Some marketers seem to get overly sensitive about the FBL complaints they get. We tell them to think of it as---well---feedback. Like a focus group, telling them how much their content sucks (or not).
For some though, even one complaint is some kind of consipiracy to destroy their business.
So out of curiosity, do you guys ever have problems with users who just report too much spam? I'm talking about "spam-o-chondriacs" who report every email they receive as junk, even though they opted-in? -
It's important to understand that telling people what line to toe isn't really in anyone's best interest. That just leads to people gaming the system, and once that happens we have to start moving the existing lines or making up new ones...which causes even bigger problems for the people who are just trying to do the right thing who sometimes get caught in the crossfire.
What David is saying is that lines will be drawn based on how everyone who sends mail to us actually performs. That means that they will change over time based on everyone's behavior and what you need to worry about is staying ahead of your competitors. There are plenty of anecdotal thresholds floating around the web, and you'll find that ours don't differ too greatly from them since we are all studying roughly the same data. If I knew what the invalid recipient rate for the majority of the senders on the internet was going to be in 1, 3, or 6 months...well, I would probably be a professional gambler.
The point to take away from this should be that if you are having delivery issues its because you're behind the curve. If you make some improvements to your processes and you don't see an improvement in delivery, then you aren't there yet. We don't expect anyone to be perfect, we just expect you to try to be better than the other guys. If you don't know how well the other guys do...well...that sounds like a problem some sort of organization of senders would be better at solving than us. :) -
Well, do keep in mind that it's a bit unfair to tell people, "there's a goal, but I'm not gonna tell ya what it is -- Haw haw!"
I think reasonably speaking, though, one's goal as far as non-deliverable rates should be 0%. Sometimes you'll be much higher than that, especially if transitioning from a previous mail platform. But, overall, if you mail regularly, handle bounces properly (in that you remove bounced addresses from lists, so they don't receive future mailings), you're going to have a near zero bounce rate. -
Okay now I completely agree that invalid recipients are a huge problem for both senders and recipients, and I appreciate that you're giving senders some warning about changes to your policies.
But this..
> "What's a high number?" That's easy. A high number is a number
> higher than that which we consider indicative of a good mailer.
..isn't really much help.
You might as well say "We're changing some stuff, some of you guys are gonna get blocked. Unlucky!"
I have a pretty good handle on the percentage of invalid recipients we send to AOL, but I have only a very vague idea of how that compares to other senders. Since you're holding all the cards here, how about a bit of information we can work with?!
For the record, the blog is a great idea. I'm sorry my first comment is a whine :)
4/9/08 5:21 PM
As an email delivery vet, I can offer this guidance mark: 10%.
While it's been some time since I've seen the study, my last recollection is that an email list will churn at least 10% per year. With the DSL-cable wars, introduction of free AIM, and Gmail, this number has probably increased over the past 2 years but should still be a hard and fast number to keep below. Based on my experience, anyone who I've seen hit that mark with any significant volume is likely in need of serious data hygiene.
>> If your invalid recipient stats are consistent with mailers similar to you, this shouldn't matter much to you.
This statement needs clarity; my take on this is that 'similar to you' means; IP address and volume per day/week/month, etc. It certainly shouldn't mean; CPG/financial/travel, etc. So it's not 'you' as much as it is 'your ID and history'. (I also don't think AOL is using domain names, but I'd love to see an AOL blog post about domain reputation sometime soon!)
Our friend, Ken Magill, offered an important question on DirectMag.com yesterday; "Would someone please explain how one would “game” a simple list-hygiene metric other than by cleaning up their file? Either a list is acceptably clean or it’s not, no?"
Unfortunately, it's not that simple. In fact, even legitimate marketers have to 'game' the system in order to maintain delivery. The simple example is the tried-and-true long-time legitimate emailer who 'finds' an email list on the shelf that hasn't been mailed in a long while. (You'd be surprised how often this occurs.) With AOL's new bounce threshold requirements, trying to mail the file all at once is likely going to get it blocked immediately. As a result, the mailer will have to split the file into multiple segments and mail it on the same days as the 'tried and true' list to minimize the percentage of boun