This chart from Return Path shows the factors at play when ISPs decide whether or not an email is spam. As you can see, reputation is all important.
The diagram below, from dotMailer, shows the path of a marketing email.
As it goes to the intended recipient’s inbox, is filtered into the junk pile, or bounced back by the ISP:
Pre-broadcast review for spam characteristics
Check for use of spam keywords and other spam characteristics in emails
For each campaign there should be a process to check the spam score for emails to see that your email doesn’t exceed the threshold.
Here’s an example of a spam report:
Don’t use IP address for links.
You shouldn’t use an IP address in links to pages or images. It’s unlikely you will, but worth mentioning, since it’s what phishers and spammers often do.
For example, do not include a link as http://18.104.22.168/page/. This will cause Outlook to disable the links and label it as spam.
Send emails to seed email accounts to check delivery
They should at least through your own firewall, Gmail, Hotmail, Yahoo! Mail. Additionally, use an inbox delivery tool checker before broadcast.
One or two addresses per ISP isn‘t going to give you the statistically valid data, you need at least 10 and sometimes more depending on the size of the ISP.
Mail server and IP address reputation
Configure email server to authenticate quality
The setup of the broadcasting mail server is not generally the concern for marketers managing the campaign. But you should check that your email Service Provider or in-house broadcaster is using these techniques:
- Setup physical address on sub-domain. The sub-domain of the main client sent is used for a physically separate broadcast service, e.g. : http://email.domain.com, for example: http://email.amazon.com (This is not essential, but good practice).
- A & MX records An „A‟ record is needed for click and open tracking. The „MX‟ record is for emailing. This setup is needed to verify sending IP address and for replies/bounces to be routed back to the server.
- Setup Sender Policy Framework.
- Setup Yahoo! Domain Keys.
Act on emails reported as spam
Best practice is to setup a feedback loop system between the ISP / webmail provider and the broadcasting email service provider or in-house server and then emails reported as SPAM are removed from the list in the same way as if they are bounced.
This prevents repeated reporting of emails as spam.
Check for blacklisting
If you notice problems with your deliverability, check that you are not on a blacklist.
Consider impact of presence of emails in spam traps.
It is not usually possible to check for addresses in your list that are potentially causing problems in spam traps.
If an email in your list is discovered in a spam trap this can lead to a permanent block, so your are most likely to find this if your deliverability falls dramatically.
Remove inactive subscribers from list
Send an email to people who have neither opened nor clicked in the last year asking them to confirm they still want the email. Then you are less likely to be reported as SPAM and this is a potential way to remove emails in spam traps.
Regular cleaning of inactive addresses, which have not interacted within a 12 month period will help to prevent deliverability issues caused by emotionally unsubscribed recipients.
Prompt users to add you to their address book
You should encourage recipients to add your address to their “Safe senders” list or address book, so that even if message has characteristics of spam it is not placed in junk box.
Not many will do this, but it is still worthwhile since it can guarantee inbox delivery. Here’s an example from Comet:
Since the proportion of recipients who respond to this are so small, it is arguably more worthwhile to do this as part of a signup process or welcome campaign rather than including it as a standard template feature.
Add unsubscribe links and instructions
Every email should contain unsubscribe instructions by law, and you should carefully word your unsubscribe method so it doesn’t look like that commonly used by spammers.
Using a ‘Statement of Origination’ or explain with text like ‘why am I receiving this email?’. Spammers often use fake statements of origination, so it’s best to check the text against a filtering system before delivering the campaign.
Your statement of origination should explain either that the recipient has opted in, ideally with the place and time of opt-in, or that they are receiving it since they are an existing customer.
Placement of unsubscribe links is another important consideration.
They are most commonly found at the bottom of emails, but there are some good arguments to make it easy to unsubscribe through prominent links, as with this example from East Coast.
A prominent link means recipients are more liklely to opt out in this way, rather than using the ‘report spam’ button on.
Review deliverability for all campaigns
While most marketers will mainly be interested in opens, clicks and transactions when reviewing their campaign reports, you should always look at delivery also.
Compare each campaign to your past history of delivery percentage to see whether you have a problem and then act on it.
You should separately review inbox deliverability since an email may not bounce, but has been placed in the junk mail folder in which case it is less likely to be acted upon by the recipient.
Review inbox deliverability for web mail companies and major clients
You can tell whether you have an inbox delivery problem by comparing the relative delivery for different groups of recipients.
Better still, look at open and click rates for different groups of subscribers and this will give you an ideal of inbox delivery or rendering problems in a specific email client such as Hotmail.