| Email Delivery Fall-offs and Open Rates |
|
|
|
Written by John Wagner
Has anyone noticed a fall off on email deliveries lately? July 31, 2006
We seed our mailing lists with our own accounts and over the past several months have noticed a drop off in mail being delivered to the regular accounts, most of it instead ending up in the spam files of the major ISPs. Specifically, AOL and Yahoo, although we suspect there are others as well. We used to get a read-through rate (number of emails actually opened) of 30-50% in mailings of ten thousand at a time, now we are lucky to see a read through rate of 10% in the same size mailing. When we call specific long time customers on the phone, they tell us they did not get our email, or they found them in their spam files (which most people never look at).
Accordingly, business has dropped off proportionately as well. We applied to both AOL (which comprises about 40% of our list) and Yahoo (20%) to be whitelisted. They sent us a form to fill out, we did, and have yet to hear from either one. In the meantime, we still find our mails being sent to the spam files. We set up an SPF a long time ago, one of the AOL requirements and still... nothing.
I suspect that this all has to do with both AOL and Yahoo's agreements to use Goodmail but I still see it as a form of blackmail. I would like to hear if other LED-ers are finding similar results.
John Wagner, VP
jewelex.com
Written by Steven McCall August 1, 2006
We
too have seen this with our clients. However it hasn't been recent for
us; it's been going on for quite sometime. I never thought of GoodMail
as much as I did HTML. Are your newsletters or emails mostly text or HTML?
I'm
going to start sending out text emails again (did this years ago) with
the only link being to the HTML newsletter which we'll now display on
the site. Hopefully this will increase the read-through or open-rate of the emails.
We'll see, of course.
Steven McCall
sundownerfacts.com
Writtten by Adam Audette August 1, 2006
I can speak to this. One of the lists we host recently
underwent huge changes. At the beginning of the summer it still had
over 180,000 subscribers. Now it has, well... keep reading.
It is an old list, having been published since 1998 or so (I'll
just call it "List X"). A large portion of the subscribers to List X
were the result of co-registration, a sometimes dubious enterprise in
my opinion, that took place in early 2001. Early last month we began a
controlled audit of all of our lists, including the LED and List X.
What we found wasn't a shock to me, as I had suspicions about List X's
subscriber integrity, but it was extremely surprising to the publisher.
As
we rolled out auto-deletions across our lists, the solid ones withstood
the audit and the subscriber numbers barely changed. For example, the
LED began with just under 41,000, and after the auto-delete cleanings
retained over 99% of its subscribers. Today numbers have climbed over
the 41k mark.
But List X didn't fare nearly as well.
It experienced massive attrition as auto-deletes bulldozed over 130,000
subscribers off the list! Today it has roughly 40,000 real-live email
addresses, and 5% of these are currently being monitored as suspicious.
The
interesting thing about this is in the results: 75% of the auto-deletes
were AOL and Yahoo subscribers, no doubt the bulk of the old
co-registration strategy. Since I was suspicious of so many AOL and
Yahoo subscribers being removed, I examined the removed addresses, and
with the help of our email distribution software, bulk confirmed that
all of them were in fact bad. Since we are on whitelists for AOL and
Yahoo, I was glad to see that the recent Goodmail situation hadn't been
the factor hiding behind the curtain.
The lesson in
all this is to build email lists the hard way: one subscriber at a
time. How this relates to email delivery fall-offs I'm not sure, but I
hope this sheds some light on another area of the publishing situation.
I
think there's a larger picture here as well, one we've covered in depth
before: the problems inherent in email delivery amidst so much UCE and
spam. It is very difficult to effectively filter email these days. Just
last week I missed an important email from Yahoo and actually received
a phone call from the company alerting me that it may have landed in
the spam folder. An auto-response from Google Analytics also landed in
the trash bin. And I am notoriously picky about reading and monitoring
my email, as I have been running discussion lists for years where every
post is valuable and needs to be received, read and filed.
Solutions?
First of all, good old RSS. Well, it's not old, but it's good! Real
good. I'm setting up feeds for the LED Digest and the archives so that
content updated on the site can be monitored more efficiently. I'll
also encourage email subscribers to add the RSS feed to their bookmarks.
Secondly,
subject lines. They are very, very important. Emails need to be easily
recognized. The From and Subject headers should be easily identifiable
and catchy. Readership improves on the LED when there's a catchy or
alluring tagline - sometimes more difficult a task than I ever thought
it would be.
Thirdly, whitelists and authentication. As John Wagner mentioned,
getting on AOL and Yahoo's "good guy lists" should be priorities. An
SPF, or Sender Policy Framework, is also important and should be
implemented. As John mentioned, AOL requires a published SPF in order
for you to be whitelisted. What does it do? Basically, it authenticates
your identity as an email publisher. There are pluses and minuses to the technology. SenderID and several other new versions are in development (Hotmail uses SenderID apparently).
Here's an article on authentication (old, but still useful). Here's another dated article on improving AOL subscriber readership. Check out OpenSPF but do not use their Wizard to create your SPF record. It's known to be busted. When you've got your record properly set up, run it through this handy SPF query tool. If you need specific help with this, feel free to contact me.
Just some ideas from me. I'd love to hear more from LEDer's about this.
Adam Audette, LED Digest Moderator
Written by Pepper Kay August 1, 2006
"We seed our mailing lists with our own accounts and over the past several months have noticed a drop off in mail being delivered to [AOL and Yahoo]..." - John Wagner
Yes...
I have items in my AOL spam folder every morning that are NOT spam...
these items are the SAME items every morning and there is nothing the
least bit controversial or noteworthy... I have been in touch with AOL
Tech Support at least a half dozen times with no success ...
I
continue to receive 'advice' from them that is a rehash of everything I
have already done - already do - and will continue to do so to
alleviate the problem ...
AOL Tech Support either does not have a clue OR couldn't care less.
Thanks,
Pepper Kay
Written by Dr. Mani Sivasubramanian August 2, 2006 Hi Adam
"Early last month we began a controlled audit of all of our lists including the LED and List X."
If it isn't too involved and intricate, could you please post a little more about the exact process this involves?
"But List X didn't fare nearly as well. It experienced massive attrition as auto-deletes bulldozed over 130,000 subscribers off the list!"
When I 'audit' my lists (or as I call it, 'trim off the flab'), I use the 'human reader' test. I
send out a series (usually 3) of direct response emails, and have a
re-subscribe process on each (either give away a PDF for which they
need to opt-in, or rarely, just send them to a form to re-join another
list). This process decimates my list - often
resulting in one smaller by 75% or more. But the small readership left
behind at each stage is increasingly loyal, responsive, and a group any
list owner would love to be in touch with.
So, while
technical issues of email deliverability matter, there's one more
'layer' of filtering to overcome to be effective email marketers.
All success
Dr. Mani
... the Ezine ANTI Marketer ;)
<Moderator Comment>
"If it isn't too involved and intricate, could you please post a little more about the exact process this involves?"
Sure,
Dr. Mani. It's all controlled via our distribution software, and while
the details are pretty fancy and technical, the way it initiates audits
is pretty straightforward. We use L-Soft's Listserv and LSMTP, the
high-powered versions of both that cost well into the 5-figure range.
These systems were purchased back in the Adventive days (for those of
you who remember) and we spared no expense for quality and reliability.
Anyway,
what Listserv does is pretty neat: it scans email lists using a passive
server-level protocol that can be fine tuned and tweaked as needed. We
normally have it on a moderate setting that removes emails from the
list only if they bounce repeatedly 100 times or 5 days in a row. These
auto-deletes are removed from the list but saved in a plain text flat-file and can be manipulated from there -- bulk jobs either to confirm
that they're bad addresses or whatever. Usually I just delete the file,
since they've never not been bad emails. I sometimes take a random sample of 20 or 30 of them to be sure.
Hope that's helpful.
Adam
Written by Janet Pickard August 2, 2006
Hi Adam,
In issue no. 2215 you spoke of auto deletes in email lists. What is this and could it help me? I
have a large list, also from way back when. I send my email through
netatlantic.com. I pay per subscriber. This list is now double opt-in,
but for many years just opt-in.
My instincts tell me
that 90% of this list is no good. When I check my logs for the day the
email is sent, that is about what my traffic increases (every week for
years now). Netatlantic says 97% have reached the subscriber. That just
isn't so. Since I pay per subscriber, I would love to find a way to delete the bounced or "qmail" type subscribers.
Any suggestions? Many thanks!
Best,
Janet Pickard
chesscentral.com
Written by Charles Gartrell August 2, 2006
We had noticed this issue a while back, even for smaller lists
(few hundreds), with large numbers of bounce-backs and what have you.
This coupled with the GoodMail systems did not bode well for our
customers (which generally operate small to modest size mail lists). In
a few cases with a list of around 9000 we even saw delivery ratios at
less than 10%. We did not have a bunch of happy campers to say the
least as result. Since delivery and open rates are for the most part
more important to our customers than speed, we tried something a bit
different.
First we dropped use of our mail list
software and developed a custom application. This new system (at least
to us) does the following - sends both HTML and text messages to each
address, customizes each message with the subscribers name and email
address, and sends no more than 50 messages per hour (not exactly true
- it sends messages seperated by quarter of a second between each for a
group of no more than 50 addresses) until the entire list is completed.
Now
admitedly this is (very) slow but it works - bounce backs are very
small, bad addresses are few and open rates are very good (we have one
list over 90%).
Is this a universal solution? Nope...
this approach is only effective for small to modest lists (< 15000
or so) but it does work and our customers are happy again. How long
this will work is another matter - major ISPs will muddy the waters
again I'm sure.
Charles Gartrell
bearweb.com
Written by Rich Dudley August 3, 2006
Since I was on vacation last week, I may have missed if anyone brought this up or not. There is a difference in deliveries and opens. Most e-mailers measure deliveries as those which don't bounce back. If a spam filter catches the message, it may seem as delivered, but probably won't be opened. This would inflate your delivery rates, since most of the trapped messages won't reach the intended recipient.
Also, it used to be a fairly common practice for e-mail admins to have a catch-all account, to filter any unknown addresses. I think this practice has declined, and unknowns are let bounce. This will decrease your delivery rates, relative to when a human was filtering messages.
Opens are usually tracked when an image -- often a 1x1 px transparent GIF -- is downloaded. Most new e-mail software, including Outlook 2003 and Gmail, block the download of images by default. Even though people may be reading your messages, you may not know it.
Rich Dudley
bloomeryweddings.com
Written by Nathan Holley August 3, 2006 Open Rates - Open Can of Worms! LEDers, I'm surprised no one has jumped on this -- open rates? Please... don't get me started. Too late, I started. This is a huge can of worms that really needs to be fully explored. This post is a start, anyway... Open rates are unreliable at best, and totally worthless at worst.
First,
what are they and how are they tracked? Open rates are so-called
because mailing list providers, publishers, and software developers
have coined the term for calculating the amount of "read" emails on a
list. In other words, how many people actually got the email, opened
it, and... well, at least opened it! Lots of ways
to do it, but probably the most common is a simple graphic (invisible)
embedded in the message. The email is opened, the graphic is called,
and the logs show a "view." This view is tracked by server log
analytics and the total amount of the list is compared with the total
tracked views, thus the open rate is achieved.
There
are many problems with this. Firstly, ISPs and email providers across
the universe have become more restrictive about what they let through,
and blackhole lists, authentication, and other strategies apply the
first layer of filtering. Second layer filtering may happen on the
client side, where either a third-party service, a script for IMAP
email, or Eudora's filering tool for POP is employed.
Just an overview - but the summation is: lots of hops along the way to intercept the communication.
What's the problem with the open rate? Briefly:
I
still think open rates can be helpful, though. For instance, using open
rates as a general barometer of list health can be beneficial. Just
don't put too much reliance in them.
Dr. Mani spoke about his "trim the fat" technique for lists,
and I think it makes a lot of sense. It is, however, a pretty
aggressive approach and will probably lose you some good subscribers.
For old lists, like Adam pointed out,
"list fatigue" sets in. Old subscribers often equal low open rates,
while newer ones are more prone to reading your emails. But don't
equate older subscribers who are less likely to read your email with
worthless subscribers. They aren't.
For instance, I did a search of the LED archives for Bruce Clay, because I enjoyed his recent post on reciprocal linking
and SEO conferences. I found 7 posts by him since 2003 (it appears the
entire collection of LEDs isn't up, since I have earlier issues of LED
Digest in my email folders somewhere). The last time he contributed
here (previous to his recent post) was October of 2004 (see "The Future
of SEO" in LED issue 1878).
So for nearly 2 years he's been quiet, no doubt busy and not reading
every issue of LED. But the point is, he posted again. And I'm sure he
will in the future.
The LED is an established resource
- it has credibility. But the same philosophy applies to smaller,
younger lists too (like most of mine).
BTW - I placed my search here: http://list.audettemedia.com/archives/led.html ... but I also recently visited the LED site and found another archive here: http://www.led-digest.com/content/view/126/129/ ... Adam, which one should we use?! I get confused by too many options.
Later,
Nathan Holley
the invisible man
<Moderator Comment>
Nathan
-- use them both! The list.audettemedia.com archive is attached to our
list software and updates automatically when an issue goes to the list.
The led-digest.com archives are better organized, better integrated
with the site and include both full issues and threaded discussions. My
plan is to add them all in threaded form, and I've been working like a
crazed monkey on it for the last few weeks. Tell me what you think.
-adam
Comments (0)
 Write comment
|
