Average Email Delivery

May 4th, 2016 by David Simms

Categorized as: Other

Email Announcement Changes

You conduct a mass mailing and a certain percentage of emails go undelivered because they simply don’t exist. This is normal, but you may be wondering if that percentage is within a normal range when compared to other organizations. To answer the question, I looked at week-by-week statistics for the percentage of emails that were not delivered because they don’t exist beginning on July 20, 2014 through October 19, 2014. I took these readings on a week-by-week basis rather than take the entire date range all at once because by breaking it down week-by-week, the highest and lowest weekly readings, which may have been aberrations and skewed the results, may be thrown out to get a truer reading. That truer reading is 3.34%.

The highest for any given week was 6.94%, the lowest 1.81%. There’s a twist. To protect its reputation as an email sender, ElectionsOnline, through its partnership with SocketLabs, maintains what’s called a suppression list. If any piece of email sent by ElectionsOnline is determined to have been sent to a non-existent address, that address is added to this suppression list, which ensures there will never be another attempt to send to it. This is a best practice because one of the things that can damage an organization’s reputation as an email sender is to repeatedly attempt to send email to an address that we know, or should know, doesn’t exist. If we assume that some of the emails sent during the sampled data range were on the suppression list, then the delivery rate would have been improved by virtue of the suppression list since attempts to email a bad address would have been prevented. In other words, the 3.34% could have been higher without the suppression list.

It can not be easily determined if any of the emails sent during the three months from which bad email data was gathered were already on the suppression list. It is possible if a client conducted an election previously that a bad email could have been added to the suppression list during that previous election and was still being included for a subsequent mailing. While that would be an interesting and more in-depth analysis, it’s unlikely there’s a high enough ratio of email addresses on the suppression list to mailing list that the 3.34% figure would move a whole lot in either direction.

It is realistic for an organization to get its undelivered email rate well below the 3.34% being reported here. Below 1% is not unreasonable. Remember that ElectionsOnline is sending email on behalf of clients. While some of these clients may be very diligent about their data hygiene practices, others may do nothing at all so there’s a constant flow of voter rosters coming into ElectionsOnline that are sure to contain bad emails. However, for an organization that works only with its own restricted collection of emails, you have the opportunity to exercise hygiene on that data and keep it in an always-current state. Doing so improves your own reputation as an email sender since, as stated above, one of the things that damages an organization’s sender reputation is to repeatedly attempt to email addresses known to be bad. ElectionsOnline helps in this effort by making a failed email delivery report available to you after a mailing, but your organization should have its own process for capturing and purging from your systems any email address identified as non-existent. Doing so will improve your own sender reputation, which in turn improves the effectiveness of all other activities that rely on putting an email inside someone’s inbox.