14 April 2015
Published 14 April 2015
Recently I reviewed some material from my early days in catalog marketing and magazine publishing. The context of "deliverability" back then, as it still is today in the physical world, was all about understanding the postal system. Things like sectional center facilities (SCFs), zip code sort, zip+5, carrier route coding, bar codes, containerization, in-plant post office, merge/purge reports and site-penetration were just some of the things that were used to make decisions about getting your catalog or magazine delivered to the right people on time.
All of that seems to have been stripped away when deliverability was netted down to one item: the email address.
Or does it?
Some marketers would have you believe this answer. Especially those that work for small to mid-sized companies; or those that handle small to mid-sized volumes of messages.
I am not convinced. It seems too simple and fundamentally naive.
Getting your message delivered to your customers at the right time and in a place where they are receptive (device / location / time of day, etc) still takes work - and lots of it.
Consider for a moment the following scenario:
Your company markets through a marketing automation provider. It could be ExactTarget, Pardot, Marketo, Unica, Constant Contact, Act-On, Genius, Lyris, Eloqua, InfusionSoft, Responsys, Hatchbucket, Sailthru and a host of others.
Next assume that you do not use a dedicated IP for your marketing but instead used a shared IP hosted by your provider. What happens if one of their customers sends out large volumes of email with deliverability rates (delivered to a valid email inbox) of 90% and Open Rates of .25% and click-through rates of .0025%?
How do the spam houses respond to your email when it comes from this same IP address and your deliverability is 99%, Open rates of 63% and click-through rates of 25%? Well, if your volume represents just 2% of the total volume being sent from that IP address you'll be treated like big volume sender that is on the same IP.
Clearly the scenario above suggests that a dedicated IP gives you more control over the reputation of that IP. This is a good thing.
But what about high volume mailers? Those that use one or more dedicated IP addresses?
Here are a few questions to consider that may be useful in making better marketing decisions based on delivery analytics:
- Number of hops per email sent. What are the typical number of hops required to get your message to each of your users? Why would anyone care about this? If you're sending a high volume of messages that are time sensitive (think financial services or emergency alerts sent by government agencies) how many hops and how much time transpires during the delivery process could make a difference. In the old days of "snail mail" number of days in the mail was a big factor when aligning inventory management, labor resources and customer expectations. Although the time measurement may have shrunk into the range of nano-seconds, the same issues exists and delivery analytics need to be visible to the the marketing manager.
- Site penetration details. How much value can your company derive from delivery analytics (e.g. number of messages delivered to an IP address and the physical location associated with that IP address; or number of messages delivered to a specific domain.) In the olden days, site penetration details from a merge/purge house could be extremely valuable in making profit & loss decisions at the point of mailing. The same may be true today based on visibility of resources across your delivery infrastructure (number of hops, internet exchanges, physical location of destination IP addresses, time of day delivery, sensitivity of content, etc.)
- Load balancing across the internet. Most of the time we think about load balancing internal networks. What I'm suggesting here is load balancing your delivery infrastructure (not sending infrastructure) but the routes across the global internet in order to align resources in a fashion similar to the old days of snail mail. Granted I am speaking of high volume, time sensitive messages, but the tools need to be available to pull it off in order for the marketer to make highly informed decisions. And if the tools are available, then they can be scaled down to the masses.