As demands for digital messaging grow more complex, the need for feature-rich and sophisticated email infrastructure becomes increasingly apparent. Commercial/Carrier grade MTAs must offer innovative, intelligent features that will accommodate the changing topography of the messaging world. The increase in queue sizes makes delivering time sensitive emails more difficult, as does possible congestion due to temporary blocks or deferrals by ISPs. Given these challenges, it is important to make sure that critical emails can be delivered successfully.
Queue Prioritization, a new, more advanced feature, allows clients to indicate a “priority” in the message header. These priority levels affect when a message is sent to a given recipient (or set of recipients). In other words, certain segments of a sender’s list would receive the sender’s email message first. Previously, this feature was only available at the domain level; now, a sender has more granular control and can manage which individual recipients receive messages first.
Increasing queue sizes signify a healthy economic indicator for our industry; however, senders are challenged daily by how much time a massive queue will take to complete processing. With queue prioritization, senders can now target the individual list segments that are most likely to engage. Concentrating on these list segments helps to reduce campaign costs, substantially improve subscriber experience, reduce spam complaints, and dramatically increase customer lifetime value all while reducing inevitable churn.
For example, senders first might consider segmenting their list to identify the most engaged subscribers. Then, this noteworthy enhancement at the MTA level will offer priority queuing based on engagement patterns (senders set priority levels for different list segments). Prioritization occurs on a scale from 0 to 100; for a list segment with a very high potential for engagement, the sender might assign a priority level of 100. Segments that are less likely to engage would receive lower priority numbers.
Using various criteria to determine engagement level, senders can determine in advance who receives emails first. Queue Prioritization is ideal for the increasingly popular time-sensitive emails. Sys Admins now can set a directive, which allows more engaged recipients to receive their emails before the less engaged, much like passengers board an airplane.
One of the advantages of prioritizing emails is that prioritization can serve as a helpful classification system for marketers, which they can use to maximize subscriber engagement. Due to prioritization, marketing now has an additional data point to encourage open rates and CTR. Moreover, marketers can dedicate emails to clients who are equivalently less engaged, thereby communicating a sense of urgency. At the same time, marketers can send out other messages geared toward rewarding subscribers who are active members of the list.
Alternatively, this feature can be used to examine the reverse phenomenon. For instance, perhaps a marketer might want to test his most unengaged subscribers. To do this, he would set a higher priority for this (unengaged) segment, as a reactivation technique. In the message, he would inform them that they are the first to receive such exclusive messaging/offers. Once again, the airplane boarding analogy applies. In this case, the passengers in “economy” class, who are seated in the rows furthest back, might for once receive special privileges. Senders who continuously refine list segments and dive into critical data can explore how soon the subscriber interacted with the message after it was sent. This data point will allow senders to define their most ardently engaged subscribers. Using this information, senders now can better determine who their most valuable clients are. Identifying these clients allows senders to deliver targeted re-engagement campaigns. The content of these campaigns is formulated when senders parse this particular data point, based on various engagement metrics. With the advent of queue prioritization, the pendulum has swung in the direction of email marketers, as it is now easier for them to determine which segment of their lists should receive exclusivity.
An example below illustrates how queue prioritization works by assigning a number to the level of importance each list segment should have in a particular queue.
It would look something like this:
mail-from: newsletter@yourdomain.com/ recipient-priority=30
mail-from: transactional@yourdomain.com/recipient-priority=79
mail-from: passwordreset@yourdomain.com/ recipient-priority=100
rcpt-to /@engagedcustomer.com/ recipient-priority=90
With queue sizes constantly increasing, setting a priority for your more evolved and engaged segments becomes increasingly significant. Queue prioritization is an innovation that allows marketers to more easily accomplish this important task.
This post was inspired by Ken Pfeiffer of Blue Sky Factory and Wayne Mehl in conjunction with the dev team at ExactTarget.
Fred Tabsharani
Last 5 posts by Fred Tabsharani
- ESP Forums: A Scarce Commodity - June 28th, 2011
- Email Delivery Policy Management: And, the Highway Patrol - May 31st, 2011
- Email Mythology Part II: Gods of Deliverability - April 19th, 2011
- Email Mythology, Part I: Goddesses of Email Deliverability - March 21st, 2011
- Deliverability Reporting: Gateway to Dividends - January 31st, 2011
I like the idea and I agree that given the expected increased in digital messages some prioritization will need to take place. However, I don't see a lot of marketers putting a lower priority on their messages. Shouldn't all marketing messages be important all the time?
I expect to see the majority of messages with a priority of 100!
Note that senders organizing email streams into multiple buckets (usually aka VirtualMTAs or virtual bindings depending on the MTA vendor) can achieve high priority for all messaging classes simultaneously. However, some email shops have to utilize the same bucket for different communication types. This is where queue prioritization technology makes the biggest impact.
For example, a publisher using a dedicated IP address account with a particular ESP can dispatch millions of messages targeted to disparate audiences where breaking news alerts maintain proper timeliness, while newsletters and sponsored offers are delivered with a slight delay. Not to forget that, while all of this email is being processed simultaneously, creative production staff will still have their pre-launch tests for other mailings turned around in only a few seconds.
Nice. This doesn't encompass everything you're talking about above (especially Mr. Fletcher's example use case), but we prioritize our delivery queue based on engagement in a similar way – http://eepurl.com/cUHek
Highly engaged subscribers are moved to the "front of the line," and we use better IPs. Subscribers with low engagement are put in the back of the queue, and we use other IPs.
It's automatic (we force it), but users can manually segment based on engagement too.
But this is not w/out risk.
Users can *theoretically* build segments of their list consisting of low-engagement recipients in order to try to "re-engage" with them (we've even written a tutorial on how to do it), but the ironic thing here is that our anti-abuse algorithms end up suspending their accounts for high concentrations of abuse complaints, bounces, etc. You can check out the comments here for issues people face when actually trying this, and some workarounds (and even automation ideas):
http://blog.mailchimp.com/how-to-reactivate-inactive-subscribers/
From a very, very high level, what we were trying to do here is lay the foundation for a process where we can reward good behavior, and penalize bad behavior. Basically, if we can get people to stop emailing bad addresses, we help the email ecosystem.
This is an excellent idea! We are actually whiteboarding some similar ideas here, though since we're not in the business of making MTAs it is a pre-MTA injection prioritized queue (ie divide contacts up into prioritized batches and schlep them to the MTA in a certain order and wait a certain period of time) and obviously being able to appropriately tag recipients and have the MTA sort out the details is ideal.
Please let me know when it makes it into a beta of Port25
.
Great to see this advancement at the MTA level, we actually already do this and it works great. Different messages in the same email can be sent to different status of users. So engaged users can be sent X offer, active users y offer etc.
Also we prioritize on activity as well automatically so it doesn't need to be set by the client (although it can be).
Ben,
I was addressing the first commenter's statement that all mail should be the highest priority. In a perfect world, this would be the case. However, some senders are somewhat limited based on the way they have to inject email into their email delivery system.
We've found that more sophisticated senders will typically use the multi-bucket approach to optimize email stream delivery to their most engaged recipients. However, few ESPs would allow/support this with basic accounts.
Russ
Ben,
I was addressing the first commenter's statement that all mail should be the highest priority. In a perfect world, this would be the case. However, some senders are somewhat limited based on the way they have to inject email into their email delivery system.
We've found that more sophisticated senders will typically use the multi-bucket approach to optimize email stream delivery to their most engaged recipients. However, few ESPs would allow/support this with basic accounts.
Russ