Commentary

Deliverability: The More Things Change, More They Stay the Same

I’ve been working in email marketing for over 16 years now, with a heavy focus on deliverability for the last 13 of these. Over this time, I’ve seen the anti-spam techniques used by mailbox providers change dramatically -- along with the tools available to email marketers. However, getting mail in the inbox is still a problem. Why is that?

Our report on deliverability shows that global inbox placement rates remain a problem, with about 20% of messages not going into the inbox when measured by seed addresses that are mailed out with each campaign.

The problem is larger when you look at “real world” inbox placement by using a panel of real mailboxes used by real humans. Looking back one, five or ten years ago, you would have seen deliverability as a serious problem in each of those time frames.

So why is deliverability still a problem after all this time? With the email marketers that I spend time with, I see four major causes:

New deliverability rules have been added to the old rules, and marketers haven’t caught up. The anti-spam engines being used by the largest mailbox providers are extremely complex. These engines look at tens of thousands of “features” for each messages, including reputation metrics tied to the IP address and the sending domain. In the past, they looked at only a small number of features when making filtering decisions.

advertisement

advertisement

In addition, these algorithms are making an individual delivery verdict for each individual recipient, instead of a blanket decision for all recipients.

In many cases, marketers are using the old approaches to solve a much more complex problem. For example, marketers might try removing email addresses that are likely to “report as spam,” when they really need to be concerned about subscribers who aren’t engaged with a particular domain but are engaged with a lot of other domains, and who repeatedly delete messages without reading.

Many marketers still don’t measure or benchmark their own deliverability. If you don’t know you have a problem, you’re not going to pay attention. Most marketers still don’t have a good understanding of their true inbox placement. These marketers depend on reports from their email service provider that shows the number of messages accepted for delivery, not mail placed in the inbox. (Note that some ESPs do provide this data to their clients.)

Common deliverability “playbooks” don’t take marketer’s most important metrics into account. The most common pieces of advice about delivery problems are based on improving inbox placement, not in improving ROI from email marketing (revenue, clicks). Getting to 100% inbox placement can require some very drastic approaches that require (in some cases) sending a lot less email and potentially driving less return from the marketing investment. Many marketers prefer to avoid this sour-tasting medicine.

Marketers are very, very busy. Email marketers are expected to drive a lot of revenue for their business, but frequently do this with very limited resources.

What should a marketer do? I would recommend starting with sensible “baby steps.”

Understand if you have an opportunity to improve email marketing performance by improving inbox placement. Use third-party delivery monitoring services to determine your true inbox placement. Do you have a problem at the largest mailbox providers on your lists? Do you have a problem with your most valuable customers?

If you have a delivery problem, treat fixing the problem as an experiment where you are trying to optimize revenue (or other metrics depending on your business model) over the medium term. There are a large number of standard plays to improve inbox placement. These include changing frequency (including not sending to certain addresses), setting clear expectations for new subscribers during signup, utilizing a series of welcome emails to onboard new subscribers, and hundreds of other approaches. Each of these has a cost. Measure the benefit you are getting in the experimental group or time frame relative to those costs. This takes time, which is why you only want to invest in this effort if you know you have a significant problem.

Here’s hoping that next year, we see a big improvement in inbox placement.

Next story loading loading..