George Schlossnagle, CTO, Sparkpost
Dan Deneweth, Sr. Director, Strategic Services, Oracle
Kurt Diver, Manager, Email Delivery Consulting Services & Email Operations, Twilio SendGrid
John Stephenson, VP Campaign Assurance, Epsilon
Tony Patti, VP Deliverability & Compliance, eDataSource
Jackie (00:05): I want everyone to give us all the juicy details: what is the biggest deliverability blunder that you’ve encountered? Obviously you don’t need to publicly shame your clients, we don’t want to put you in that situation, but the more details the better.
Tony (00:34): A while ago I was working for an ESP and we had an international client from France. At the time the EU didn’t have the regulations it has today; things were a lot looser, and one thing that was particularly loose was mailing lists, buying them and renting them (both B2B and B2C). The concept that this was not a good idea was completely foreign; they had no idea. List brokers we’re doing really good business there. So, we have this client who was France’s largest entertainment site, she had emails that you’d sign up for and you’d get a regular email every day telling you what your astrological predictions were for that day. We were having a lot of deliverability issues on the platform she was on and we got down to starting to correct it. We were on an uptick, getting into more inboxes and someone in her company sold her a list. She didn’t ask us. Then one day she went from 95% inboxing, plummeting down to 25% and generated two Spamhaus listings. The biggest lesson there was don’t buy or rent lists. On a side note, we all thought it was kind of humorous, but if she was a psychic why didn’t she know that list had some spam traps on it?
John (02:00): The one that I thought of when this question was posed was an international client. They had a list, an active list of about 1 million. But when I say active, they were pulling a 5% open rate on average, which I consider anything south of 10% to be deficient in some way. You need to get some type of engagement and audience interest to maintain a strong product reputation and therefore inboxing. They were already off to a bad start, they were actually getting 60% of the list distributing to Gmail which is pretty darn high, but they had this great initiative. Their active list was 1 million, but they had this other list which was great, and this list was reportedly given to them by 1 million other people. But it wasn’t really a true opt in, it was given directly to the company from these people but not an opt-in. All the questions I asked at follow-ups were vaguely positive but mostly vague. Now, my questions came after they deployed this obviously, or they wouldn’t have deployed. I caught it on the second deployment because some of the ISPs took a while to punish them for this. So, they basically doubled their send volume, but it was terrible because half of the people had no interest in their product. So very quickly they went from a high reputation in Gmail to a bad reputation; they actually grade it and bad is the worst reputation. Their open rates in Gmail went to around 5% as I stated, to about .5% or lower and stayed there for a long time. They didn’t really follow much of my advice to dig themselves out, which was to just start targeting actives. They also took a hit in Outlook and Yahoo but those were a lesser concern because of the list distribution. I think that follows with a lot that’s been said about sending to engaged users. It took them a long time to dig out because for a long time they didn’t want to follow my advice. Even when they did follow my advice, sometimes Gmail algorithms don’t really change their minds that quickly so this was totally unnecessary. I kept asking our partner, “Isn’t this client freaking out? Half of their sales have dropped off.” The answer I got back was they didn’t really track this line of business as closely and they don’t see any drop off. Which, you know, my other thought was maybe they should stop email marketing altogether if this doesn’t impact them negatively, then nothing will, so they might as well not market if this isn’t impacting them. So yeah, that was the worst one.
Kurt (04:50): So, mine is kind of cautionary as well. How many of you track where your messages land in Gmail and which tab they are going to? And, do you notice an impact when a message isn’t updated for a monthly promotion? Is there a decline in engagement? That’s what we had with one of our clients. Going back, they had a great deliverability, they were a big company. But they grew their list without telling us, and they didn’t do it gradually. They bumped it up by 4x which caused messages to slip from the updates tab, where they had great engagement, to the promotions tab. They saw a major impact and it’s really hard to sway Gmail’s opinion to put your messages back in updates. So, you can contact your Google AdWords person, they’ll talk to the postmaster, they’ll make a slight change but in the end the recipients have spoken and now your messages are now classified as promotion. So, you are not getting paid, there’s nothing you can really do other than reducing your frequency and content to try to move your mail back to the update side. So, just more cautionary as you increase your list or make big changes there’s those repercussions. We are able to identify it because we are one of the few customers that actually track ROI by message and know what each email is worth. But again, we use eDataSource and Chris was able to go back and tell us message placement over time, so we’d go back and overlay our data and we saw that change in tab placement.
Dan (06:14): We all love this topic of blunders and we all have lots of ideas; we could talk all day on this. There is one that I think stands out and it’s actually a common practice and lots of people are doing it. Some of you in this room probably have done this at some point and it’s not a stupid thing to do, it’s just something that people overlook, and it has to do with acquisition of emails. But it’s not the method of acquisition that I’m concerned about most, it’s what you do with the emails once you have them. And for me the biggest blunder people make is not properly vetting new email addresses when they come onto your list.
This is a big deal, I have spent a lot of time with our clients on this. It’s usually in the form of a conversation with spamhaus because this is how you get Spamhaus listed, by new email addresses coming onto your list either from your website, in-store or from a partner. We all love new sign-ups right? It’s the lifeblood of our business. It’s what we need to do good email marketing. But, what do we know about these email addresses really? Very little if you think about it. It might be an invalid address that someone typed in, they may have intentionally given you an incorrect address. It could be associated with a person other than them; they type in someone else’s address. Now you’re mailing to someone you never asked for. Or, worse, the one we care most about for deliverability, is it’s a spam trap. How do spam traps get on your list? Well, people make a typo. It seems unlikely, but it happens every day. It’s very important to think about people making a typo on the domain that matches the Spamhaus domain. It’s not just Spamhaus but all kinds of blacklist operators; that’s how you get blacklisted. If you have been blacklisted, you know a thing or two about spam traps.
I was talking to a client last week, Spamhaus listing, and they said, “You guys must have changed something. It’s got to be your fault.” I said, “OK, let’s talk. Tell me what you do with a new email address when it signs up.” “Well, we mail to it once, then wait a month and we put it into the general population.” I said, “OK, what happens with the general population?” “They’re eligible for three messages a week for two years.” Now these addresses are spam traps, and I mean it’s really hard to know that right? But, what’s happening is that you’re releasing spam traps into your general population. They don’t bounce, right? So you’re not getting a bounce, you start mailing to them and you’re happy. Then more come in the next day, and more come in the next day, and before you know it you’ve got so many spam traps on your list that you’re mailing to them three times a week for years and it really snowballs. So, if you don’t have a problem now and you think you’re okay, what happens as time goes on?
The good news is we know a thing or two about spam traps and we know how to address this. Spam traps don’t typically open or click. So, the solution is to properly vett these with a rule designed specifically for new addresses, I call it a Unified Lateral Respondent rule. See, you have addresses that sign up; they never open and click; you have limited interaction with them. That’s how you protect yourself and your whole program. I think of this as the “Achilles Heel” of deliverability. In fact, when you read a lot of blogs and look at people’s lists it’s usually not on it. I don’t see a lot of people talking about it, but I see so many clients get blacklisted because of it and it’s easy to fix.
What happens is someone signs up and they do not open or click, either they’re a spam trap or they’re not interested. Either one is bad; the first one will get you blacklisted, the second one will just drive down your engagement rate leading to other deliverability problems. If you close this hole you can actually be more aggressive with the other mail. But I see this as the open hole for a lot of senders that have everything else buttoned up. They’re smart, they do everything right, then missed this one thing. I mean Spamhaus and blacklistings are the worst things you can get and it is devastating to senders, just when they thought they had everything. So that’s the hole and I think that it is a blunder because people should be closing the hole and I would say for all of you when you get back to the office, check and make sure you have a rule like that and how many times you mail to them is your choice, each touch point to the spam trap is incremental risk of getting blacklisted. A good rule of thumb is five, or not more than 10, but two years, that’s just a disaster so look into that when you get back home.
George (11:10): Awesome, that’s great. Before I go into mine, I think that rule combined with what you were talking about in regards to taking aggressive steps to keep an unengaged user off your list, are by far the best ways to eliminate problems with hidden traps. So, I guess the thing that I wanted to talk about was maybe the flip side of that, something we’ve actually seen multiple times from people coming onto our platform from other platforms. You have customers that could use the suppression list capabilities within their existing provider. So, they’re using that for suppressing mail to hard bounces or suppressing mail to people that unsubscribed or complained. Then, when they pick up and move to a new provider, something that is unfortunately common is not bringing that with you. And if you don’t bring that with you then all of a sudden you’re emailing to a high number of nonexistent addresses, potentially mailing to people who have asked you never to mail them again, people that have complained and you can really generate a huge number of problems, and we’ve seen that. Actually, fortunately, I can’t think of a high-profile case where the result ended in a block, but certainly a deeply degraded inbox placement at Gmail, Yahoo and others. It’s simply a very dangerous thing to do. So, I think the gold standard is if you’re getting hard bounces and complaints, and you want to remediate those while removing people from your list, even if your provider provides that facility themselves you’re better off addressing it in your own list so that you have the intelligence incorporated into your own platform. But, if you don’t do that at least please bring that list with you, export your suppression list from your provider we can bring it to a new platform to save yourself a lot of headaches.