Maybe I should have entitled this article “A holistic approach to email deliverability” (as a nod to natural medicine), but I didn’t dare to ;-). So, I didn’t.
Most of the time, email deliverability is poorly understood. Deliverability issues are solved through little corrective actions that often result in incidents even more complicated to solve. That’s why deliverability should be considered in its entirety. Any difficulty you encounter, even the smallest one, should bring about structural corrective actions rather than circumstantial ones.
In an ideal world, we shouldn’t need to monitor deliverability: indeed, it wouldn’t be necessary if all emails senders complied with good practice. Unfortunately, all emails senders aren’t quality senders, so spam filters do need to protect themselves against the downward slide of some users.
So much that even senders following good practices that respect their recipients can get caught about some details making them look like spammers.
Following the deliverability reputation principle, FAIs and webmails try to detect if a message is wanted or not by the recipient. To do so, they analyze a multitude of signals that makes reputation vary either in a positive or in a negative way. That’s why you need to work on these signals so that they’re always good.
Let’s put ourselves in the shoes of the advertiser, but also in the shoes of the sending infrastructure as a whole, be it the infrastructure of an advertiser (that has internalized its routing or deals with numerous internal clients) or a marketing platform (that needs to maintain its overall reputation and satisfy and assist its clients with deliverability).
1. Anticipate and limit risks
The best way to solve an incident is… not to encounter any incident! To make this possible, you’ll need to implement deliverability monitoring (and blocking) techniques that are stricter than spam filters and ISPs and webmails reputation engines. The idea is to eliminate all practices or signals that are similar to those used by spammers.
Spot engagement decrease
Engagement is one of the indicators that webmails use the most to know if a campaign is wished for or not by the recipients. If your opening rates or click-through rates are decreasing – in general or for a certain type of emails – you need to ask yourself why and to adjust. This is true for deliverability but of course from a marketing point of view too.
Notice that you are approaching critical thresholds
On top of engagement, other global statistics need to be observed; mainly the spam complaint rate, but also the hard bounce rate. Some deliverability monitoring tools can also help you checking the amount of spam traps that are concerned.
Monitoring the reputation numbers delivered by mailboxes providers is also interesting (Signal Spam, Microsoft SNDS, Google Postmaster Tools…).
Avoid any practice close to spammers’ practice
Some tools allow you to test your messages and get some feedback on possible blocking points BEFORE sending your message. To spam filters, emails content is less important than it used to be – in fact content analysis has completely changed, but that’s another story. These practices can be introduced for instance when you change the size of your emails, or when there is a mistake about the sender profile selected. Historically, this point was dealt with a ‘spam’ score based on the venerable ‘SpamAssassin’.
Limit bad technical configurations
A deliverability infrastructure should be stable and should not change unexpectedly. However, like men, technique can be unpredictable. An IP that disappears, a tracking that stops, a SPF record that changes, an email that is not properly signed… All those elements need to be regularly analyzed to avoid any unpleasant surprise when sending your email.
Preventively block flows of problematic emails
This issue mainly rests in the hands of the person or the company in charge of the deliverability infrastructure. If one of the elements listed above becomes an issue, there should ALWAYS exist an automatic blocking mechanism becoming active from a certain threshold. This blocking mechanism prevents signals from looking deteriorated in the eyes of ISPs and webmails. It needs to be as discerning as possible, in order to stop only the campaign type and risked sending profiles.
2. Spot deliverability incidents
If despite all your efforts to prevent deliverability incidents you’ve still headed for failure, it is crucial to check what kind of deliverability issue you’re faced with. Identifying the origin of the problem will help you solving it.
There are 4 types of deliverability incidents.
Bad emails classification
This is the most harmless of deliverability issues. If Gmail got the ball rolling years ago with its ‘Promotions’ or ‘Notifications’ tabs, the whole of the operators followed, and emails classification is now widespread. Obviously, a transactional email landing in the Gmail Promotion box is a problem. That’s why it is essential to track classification issues, thanks to dedicated tools or clients’ services feedback.
Besoin d'aide ?
Lire du contenu ne fait pas tout. Le mieux, c’est d’en parler avec nous.
Emails qualified as spams
This is probably the most common deliverability incident. Though your emails look fine, they still end in the spam box. It is important to understand that an email landing in the spam box is still a ‘delivered’ email! That’s why you won’t get any error message like bounce – this piece of information won’t directly appear in your statistics reports. There are numerous reasons for this but it is often linked to your reputation level. The easiest way to detect that you landed in the spam box is to check the opening rate by destination. If a destination (for example Outlook.com) drops from a 22% opening rate to a 10% opening rate when other destinations are stable, you’ve probably been identified as a spam by Outlook.com.
Another technique consists in using seed lists through dedicated solutions that provide several hundred addresses from most of the messaging services existing on the market. You integrate these addresses within your ESP and send campaigns towards these addresses at regular intervals. The inbox monitoring platform will automatically pick up the whole of the addresses to check if your emails landed in the inbox… or not. This will grant you access to a statistics report where you will see the delivery rate in spam box or inbox for each destination.
Blockings at the entry
A blocking ‘at the entry’ is a problem happening with the SMTP server. The message is blocked and should normally come back as a SMTP error message (block bounce). Therefore, you should observe a massive increase of bounces for one or several destinations. The other (additional) solution is to use a seed list and inbox monitoring, in which case emails are shown as ‘missing’.
This is a much more serious blocking than appearing as a spam and requires more thorough corrective actions.
Blacklisting
This is the fourth stage of deliverability incidents. Your IPs or domains can be blacklisted; this often happens when you’ve touched several spam traps. The only way to quickly detect this problem is to put your IPs and domains through a tool that checks all the blacklists on the market. A few blacklists are quite serious and can really impact your deliverability, whereas others are barely used. It’s a good idea to set up an automatic alert system.
3. Detect deliverability issues sources
If a deliverability incident is proven, you need to take action rapidly to block the flow of problematic emails, and to undertake an inquiry work that will allow you to identify the origin(s) of the incidents (to see which flows of emails are concerned and why they cause problem).
One-time blocking
If a temporary blocking happens (sometimes some emails are suddenly blocked or delivered as spams), make your inquiry step by step. You’ll probably notice on your dashboards a performance decrease on a specific day. Go on digging to find the name of the campaign, then the problematic message or target, then the destination.
From here, you may be able to spot the issue and understand what triggered the incident.
Multi-blockings and structural issues
This means you’re regularly exposed to performance decreases, returns between the inbox and the spam box, waves of block bounces and so on. In your inquiry, look at the history of your data management methods. You should check if any change happened within the configuration or infrastructure and monitor the roll-out of automated campaigns and scenarios.
The investment to restore the situation may be more complex.
4. Improve deliverability practices
Depending on the origin of the deliverability problem (see inquiry work), you’ll need to apply a ‘structural’ corrective action. You should never aim at resuming emails delivery as soon as possible, but rather aim at correcting once and for all the origin of the incident through a change of configuration or practice. Finding temporary solutions amounts to exposing yourself to new deliverability blockings that will be more and more difficult to solve.
There are different types of good practices:
- Deliverability infrastructure – Use of IPs, domains, authentication, DMARC…
- Data quality – Emails acquisition methods, inactive emails management, bounces cleaning…
- Emails content – Brand identification elements, hosting domains for images and tracking links, quality of the code…
- Targeting and marketing pressure – Targeting and personalizing strategy, sending frequency, staggering of sendings…
- Recipients engagement – Performance optimizing strategy (trials plans), messages ergonomics…
5. Communicate the right information to ISPs/Blacklists
It can be useful to communicate with ISPs and blacklists to ask them to unblock the situation, or at least give you some information. Nevertheless, if your problem is linked to your reputation, you’d rather wait patiently while your corrective actions hit their mark.
If you need to contact them, here is what you should communicate and ask for:
- How did you detect the issue?
- What’s the origin of the problem (you’re always the guilty one, not the ISP or the spam filter…)?
- What structural change have you implemented so that the problem doesn’t repeat itself?
- Ask for the unblocking.
Conclusion
Like in everyday life, better safe than sorry! Even with the best will in the world, it is sometimes hard to succeed. Therefore, you need to get ready, acquire good reflexes, use the right tools and benefit from a good support.
