by Darren Fell
The first and most important area in deliverability is the reputation of your sending servers or those of your ESP.
Many think that the reputation or 'sending history' is associated with the sending domain name, but in most cases it isn't. As we know with 'phishing' attacks, sending addresses can be faked so reputation systems look at the root delivery address that can never be altered; the IP addresses of the delivery servers.
So, have you got a good reputation?
There are a number of services out on the internet that allow you to interrogate the IP addresses you are using to deliver your campaigns. If you don't know how to find your IP address that, you'll see it within the first four lines of the email header when you view the 'source' of the email campaign.
Services like Sender Score give a rating of between 1-100. For an incredibly simple rule of thumb, scores above 80 are good and anything below 30 indicates that the reputation of your sending servers could be pretty poor. If it's this bad it will instantly have an effect on your deliverability and it's highly likely your emails will be blocked across numerous ISPs.
Well there are a myriad of possible reasons, but it all comes down to poor management of either the internal corporate system, or the ESP is providing the service through a shared set of delivery servers – hence using shared IP addresses. The latter is akin to giving the spare set of keys to your brand new car to your 17 year old brother who has just passed his test.
The trouble with shared systems is that numerous 'spam' complaints are registered across many customers. Most often don't have particularly good data policies - ie poor quality lists, so the reputation score drops and free email services like Hotmail, Yahoo, AOL and other ISPs eventually block the IP addresses. Funnily enough, some ESPs can experience intermittent success in getting through to certain ISPs with some email campaigns but not all. This is experienced when, down to sheer luck, your campaign goes out on a delivery server that hasn’t yet been blocked… Separate IP addresses are all well and good, but again these can be blocked.
Is there a better way? Yes of course. You see, it's all in the management of the process.
White-listing teams really need to know what they are doing and should have numerous tools at their disposal. For example, MSN provides a web-based system called SNDS, once you have the sending IP addresses assigned to their feedback loop. It provides an incredibly useful 'traffic light' sequence across the IP addresses for Hotmail, showing green for very low spam complaints, amber for medium level complaints and red for high complaints. Red will be blocked if it hasn’t been already.
If the white-listing team have these tools at their disposal, customers can be alerted to reduce the frequency of their email campaigns, or can be questioned by account management teams about their high complaint levels. For example, the main question asked is ‘where did you get that data from?’
This is the second key area and looks at making sure the actual email is as perfect as it can be; in the eye of the spam filters. There are two main areas here: the quality of the HTML and the spam scoring of the email.
The first is easily addressed. Double-check that the HTML is perfect. It is so simple for designers to produce something that is fabulous and that works in every browser. Trouble is, email clients are not as clever as browsers and often can't recognise mistakes. Sourcing an HTML validator will spot simple errors like unclosed off tags. Lines of HTML code stretching beyond one thousand characters can cause problems in Microsoft exchange servers - incredibly simple issues like this can actually prevent delivery as the spam filters will pick the email up and spit it straight back; particularly with web-based email systems like Hotmail or Yahoo.
Ensuring your email has a low spam score is relatively easy these days. The ESP, if you are using one, should have a spam checker built in, so it is important that you double check your email before sending it. Typically, Spam Assassin is used and email campaigns will be checked live against Spam Assassin's database. This contains endless recordings of what constitutes spam and apportions various scores for each 'spam' trait.
For example, all upper case letters in the subject line will get a score of 1. Using spaced characters for impact will get another score, and so on. The spam checker should then make it as easy as possible to identify the issues with your email and allow you to change it so you fly though corporate spam filters with low scores.
This is possibly one of the most interesting areas in deliverability and is fast becoming a critical part, particularly if your lists have a high number of email addresses for ISPs like Hotmail, AOL and Yahoo etc, which have a 'report spam' feature.
The theory behind this is that you simply need to keep reminding the end recipient that they actually did sign up to your great email communication, and while you are doing it, it's worth reminding them which email address you are sending the communication to, just in case they have signed up a few times under different addresses.
“But why?” I hear you ask. “They signed up legitimately and we even used double opt-in.”
Ok, but are you aware of exactly how recipients manage the ever-increasing torrent of emails flying into their inbox?
A recent survey conducted by the Email Sender Provider Coalition (ESPC), across a fairly small sample of 2,200 people, was intended to gauge consumer behaviour towards spam. The recipients involved used some of US's top ISPs: AOL, MSN/Hotmail, Yahoo!, Lycos, Excite, Gmail, Netscape and Compuserve, and some startling results were uncovered.
The survey found that 80% of the email-savvy users in the sample were particularly familiar with the 'report spam' button as a way of managing their email inbox!
What is the direct result of people clicking on the 'report spam' button? Yes, you've guessed it. It directly affects the reputation of your delivery servers.
To give you an idea, over a 0.13% complaint rate from a delivery IP address into AOL, over a 24 hour period, leads to a number of warning notification emails being sent through to the ESP’s white-listing team or your internal IT team. If these are ignored (which would be madness), a block is placed so no emails get through. Depending on the state of the IP address reputation at this point, as the deliveries are stopped, the reputation slowly rises and eventually lifts, with luck.
If this is the case then we need to continue to build that trust and remind end recipients again and again, 'You signed up to us!'
The question is, how do we do this? All that’s required is a simple line at the top, before the creative, that spells it out:
'You signed up to www.your_brand_website.com and we are sending this email to: johnjones@companyX.com'
Will it stop them clicking on the 'report spam' button? Possibly, but combined with the other following tips, it will almost undoubtedly make a difference.
Ok, so looking further into the ESPC survey, the following statistics were revealed:
- 73% based their decision on the 'from’ field.
- 69% based their decision on the 'subject’ line.
- 79% used the report spam button when they didn't recognise the sender
- 20% admitted to using the report spam button as a quick way to un-subscribe
The importance of the ‘From’ field…
If the majority of decisions are based on the sending address or 'from’ field, it really needs to be clear.
Using an ESP's system that simply provides a cover address over their broadcasting address could mean that it doesn't comply with Sender ID regulations on spam filters and you go straight to the junk folder.
Using a system that broadcasts using the ESP’s domain to the end recipient doesn’t look like the brand they signed up to. So why shouldn't they move immediately and click on that dreaded button?
The ultimate way is to use email masking, where a sub-domain is set up for your brand URL and points back to the ESP. This enables you to send an email campaign so it's easily recognisable as you, without the issues of getting caught by Sender ID equipped spam filters. Here's an example. From: Brand area description [firstname.lastname@example.org]
If the ESP does it properly your send and reply addresses will be instantly recognisable. Expect to see every link in your plain text and HTML parts reflect your brand exactly and not the ESP's domain, which will always cause suspicion, especially with phishing attacks still in abundance.
As a last note on this sub-area, if your sending or reply address has the longest line of alphanumeric digits you've ever seen; please don't think this is standard in the industry. It's not. If anything, many people will believe this is spam as it's so unclear who the sender really is. Here's an example to feast your eyes on:
Here’s a final suggestion to build trust. Perfecting subject lines is a whole topic in its own right, but we’ll leave that for another article. Taking all elements of the above into account, the following provides the ultimate set of details that should be used to create trust. Many will raise an eyebrow at my suggestion to put the ‘un-subscribe’ option at the top of an email, but the question I'd pose is do you trust your communication? If you do not think it is fantastic material and wouldn't want to receive it yourself, then you have your answer. Getting recipients to click on your ‘un-subscribe’ at the top is far, far better than them clicking on the spam button, which as we now know directly affects your sending reputation.
Here’s an example of the copy necessary to go at the top of the email, before the creative:
You signed up at www.emailmarketingmanual.com and we are sending you this email to: email@example.com
If you cannot view the images within the email click here for the web based version.
If you no longer wish to receive this please click here and we will immediately stop you sending further communication from www.emailmarketingmanual.com.
Given the area discussed above and the realisation that people actually manage their email using the spam button, it really starts to make you think that if you consistently send your email campaigns at the wrong time, people will simply manage your email with that button.
First off, take a gut reaction call as to the best time to conduct the send. This is exactly how you should start but the idea is then to test your theories with an enterprise-level reporting system within a top-end ESP that accurately shows when all the activity occurs. You can also use intelligent time sending software which pinpoints when recipients are most likely to open your email. Ideally, recipients should be engaging with your email campaign within the first few hours of the delivery, so it’s clearly worth hitting your recipients when they are going to be the most receptive.
I accept this is an obvious one but you'd be surprised how major brands make mistakes over the data they use, or rather how they manage the data. Get this key part wrong and you really won't be surprised when you get stacks of complaints...or at how fast you get blocked, left, right and centre!
Brokered lists have to be used on occasions and that in itself is a separate discussion. This is about making sure that the organic lists you build follow a well managed, single opt-in route. For example, when people un-subscribe, provide a facility that does it automatically and instantly.
It is so surprising the amount of emails you see that say: 'your un-subscribe will take 5-10 days to be effected'. Oh dear! What I see happening is some brands thinking they can make the most of this time period and literally blast the poor individual with emails every day, until the unsubscribe is actioned. Seriously, don’t do it. If there’s one way to tarnish your brand, this is it.
Provide this unsubscribe facility, which should be standard in every ESP offering, and make sure it is instant. I believe that a click through to a confirmation web page is sufficient; the need to send an email to a special address is somehow lacking in the assurance the end recipient needs.
If of course your data capture is managed through a double opt-in process, then perfect. This method requires a welcome mail (confirmation email) to be sent to the user to check they are who they say they are. They simply click on a confirmation link within this mail and they're then validated. This avoids the whole saga of people who didn't sign up receiving emails that they then complain about. In short, double opt-in requires a bit more effort but is well worth it for the assurance that the data is perfect.
What does the creative have to do with deliverability? Well, it follows the same principles as ensuring the subject line and ‘from’ field are explicitly clear to the recipient, taking them one step closer to reading the email.
If the creative has been designed for the preview pane and a fantastically crisp, powerful graphic appears here or a glimpse and the key call to action can be seen in this area, this will, with all the other elements combined, encourage the end recipient to read the email properly. Again, all incredibly simple but it works.
Overall, don't consider email to be the lesser of the marketing channels. Design your emails as though they are being created for your best above the line campaigns (if you still do them) and I promise you, you will see the results improve and improve.
Phew! We've covered some considerable ground and have hopefully enhanced your knowledge of deliverability. Of course, if you still manage your own internal email solution then you probably realise the magnitude of successfully managing your IP ranges and the amount of time it will take to get properly white-listed (where possible) or simply get on the response paths. You could argue (and I would say this) that this is exactly the reason why outsourcing is so successful and far more cost-effective. Whatever you choose to do, much of deliverability, as you now know, is in your hands.