This is an interview with Jesse Drukker, the Marketing Director at Inboxtrail. Inboxtrail is the premier tool for managing email deliverability. This interview talks about email deliverability as it relates to various ESPs, best practices, and Inboxtrail.
Q: Why did you create Inboxtrail? What need or problem in the market made you realize there was a need for something like Inboxtrail? #
Inboxtrail was born out of the need to demystify email deliverability, which so many people struggle at managing. Many organizations don’t know when their emails get blocked or spam-filtered or what to do about it when it does happen, to the point that it’s common to see businesses say to their customers “Please remember to check your spam folder” when they send an important email. Of course, most of the time these people are not spammers, which really signals that something is broken in the industry!
So we created Inboxtrail to make it easy to test the deliverability of emails, to know when & where delivery problems occur, and to fix those problems – before the emails are sent out to users. We also created our ESP Comparison Center where we run daily delivery tests from each ESP and publish the data, since users need to know if their emails will be doomed from the start by their ESP’s sender reputation.
Q: Why do you think most ESPs don’t publicize their delivery rates? #
It comes down to the simple fact that ESPs generally get paid by the number of emails sent, not the number received! They often advertise high delivery rates even though in some cases our delivery test data indicates otherwise (and we’ve heard from a few ESPs who were not pleased to see their poor performance on Inboxtrail’s ESP center).
I get the impression a lot of ESPs prefer to sweep the question of deliverability under the rug, rather than bring doubt to the myth that an ESP’s services guarantee 100% inbox placement. In reality, whether we like it or not deliverability rates are always a collaborative effort between service provider and customer, and we hope our own publishing of ESP comparison data will help foster industry acceptance of this fact as time goes on. Postmark’s own moves on this front have been encouraging.
Q: How do you measure deliverability for different ISPs at Inboxtrail? #
Email delivery testing works like this: We give the user a list of email addresses distributed across the top ISPs (Gmail, Hotmail/Outlook, Yahoo, AOL) – this is known as their seed list. The user sends an email to their seed list, and we show them delivery results in real-time from each seed account: did their email land in the Inbox, the Spam folder, Gmail’s Promotional or Social tabs, were there duplicates sent, or was it never received at all?
Of course for accurate results, users need to send their seed list the specific email they want to test (since content matters) and send it from their ESP or email server (since sender reputation matters). This will accurately measure what their inbox placement rate will be when emailing their users.
Q: Do you track the difference between deliverability of marketing email and transactional email? If so, have you noticed a difference between the two when it comes to delivery rate? #
Yes… and get this: Almost all the delivery tests we run that use promotional ESPs (ie. marketing email providers) show that the emails land in Gmail’s “Promotional” tab. And remember that Gmail has the largest market share of all ISPs, so this is a big deal. We run daily tests on our own opt-in email (which new Inboxtrail users need to get in order to activate their account). This is arguably the most important email we ever send a user, and our tests show that if we sent it from a promotional/marketing ESP, the chance of it being opened drops by a lot!
This is actually one of the strongest examples of why it’s important to segment email providers by content: send transactional email from a transactional ESP, separate from the newsletter or product updates and other email marketing.
Q: Which ISPs do you see the most deliverability issues with? #
The most volatile are definitely Hotmail and AOL. Their stats (spam folder placement, blocked emails) often spike in a manner that might have nothing to do with the sender. This becomes obvious when one sends the same email every day, and suddenly it starts getting placed in the spam folder, but then without any intervention on the part of the sender, it starts landing in the inbox again after a few days. This signals internal ISP issues.
For this reason, senders shouldn’t panic if their delivery tests suddenly show problems on Hotmail or AOL. Immediately modifying content can sometimes backfire. We always recommend waiting a few days to see if the problem improves on its own. In contrast, Gmail is consistent. If they start blocking or spam filtering emails, it’s usually for a reason and intervention is probably needed.
Q: Do you think it’s best to use a dedicated IP or multiple shared IPs to send transactional email? #
That’s a tricky subject, but the short answer is: shared IPs. Historically a dedicated IP was seen as advantageous because it let the sender be identified by ISPs and build a reputation independent of any other senders. But at the same time it means starting out with no rep at all, and building a reputation for a fresh IP isn’t easy. New IPs need to be warmed up, they tend to get greylisted by default, and one needs to wait a while to apply for whitelisting.
Plus, more recently IPv4 address exhaustion has meant dedicated IPs will often have been recycled. This means they may come to the sender with a pre-existing reputation which could be bad! ISPs have also reacted to this by caring a bit less about IPs, and factoring in other variables such as outbound weight of the links within emails. All this means that it’s best to use shared IPs, assuming they’re managed by an ESP with good deliverability.
Q: What other metrics besides deliverability are a good indicator of the quality of an ESP? Do you plan to start testing for those metrics? #
There are definitely other important factors that define an ESP:
Speed. Do they have the infrastructure to handle high volumes & avoid delays? All that thought senders put into the best time of the day to send an email goes out of the window if their ESP has a queue of several hours.
Quality Control. ESPs should develop advanced systems for early detection of abuse before the bad apples affect their reputation and harm the deliverability of the ESP’s user base as a whole. An ESP should not be afraid to cut off users who will negatively impact everyone else, even if it costs them short-term revenue by losing the offending customers.
Integrations. Does the ESP have a versatile and well-documented API? Are there pre-existing solutions for integrating it with a wide range of CRMs & marketing solutions?
Customer support. This is more than a friendly voice on the telephone. How does an ESP communicate ongoing issues to their users? Every ESP one day faces downtime, and this serves as a test for the transparency of their operations.
We’re working on rolling out an expansion of our ESP Center that includes this deeper level of analysis, since these are all important factors to consider when choosing a provider.
Have any questions or thoughts about this article? If so, please leave them in the comments.