The Skinny Talks Gmail and Deliverability on the Salesforce Blog


I recently did a Q&A session with the Salesforce Core team about deliverability, and it was published today on the blog. It’s a bit remedial because it’s geared to a fairly broad audience, but I’m pleased with the results. Continue reading

Killing Spammers is Never Enough

The Messaging Anti-Abuse Working Group today has published it’s Best Common Practices document for E-mail Service Providers to use in vetting prospective customers. MAAWG is the foremost professional association in my industry, bringing together some of the best minds and well-known companies in the space to develop and promulgate strategies and policies designed to combat abuse of messaging networks (including, of course, e-mail). The practices detailed in the document should provide a much-needed benchmark among senders of high volume e-mail who are serious about curtailing abuse of their own networks, as well as those of their recipients.

Careful vetting of prospective clients isn’t just good for fighting abuse; it also makes good business sense for ESPs. Whenever an ESP brings on a new client, the ESP assumes a significant amount of risk. As the document explains, ESPs “are at the mercy of their worst clients’ worst practices.” Bad clients create reputational problems for themselves and for their own mail, and can damage the reputation (and therefore the deliverability) of all of the ESP’s other clients, as well as for the ESP itself. More than one large ESP has made itself a pariah within the industry by paying scant attention to the types of customers to whom they’d sold their services.

The existence of the document underscores an important industry truth that has been long understood, but to my mind has been historically under-emphasized: it’s never enough for ESPs to kill the spammers that appear on their networks. We must expend at least as much effort to ensure that we’re not giving birth to new spammers in the first place. After all, it makes little sense to keep bailing water out of the boat without troubling to plug any of the leaks.

And the document has lots of practical advice for the plugging of leaks that can be implemented right away. It contains a questionnaire that ESPs can copy and paste as-is for vetting new, high touch clients during a pre-sales cycle, and includes important advice for the monitoring of existing clients once they’ve successfully completed the on-boarding process. I’m very pleased to be able to report that my own employer, North Carolina-based iContact Corporation, has implemented the questionnaire to vet large managed customers, and is developing sophisticated tools that can automate the on-going vetting of existing customers who make use of our popular self-service options. We’re proof that the policies and recommendations contained in the document can be implemented to useful effect in a high-volume production environment.

I think it’s a pretty good document. If you work for an ESP, or for an agency that partners with an ESP, you should check it out.

On a more personal note, the adoption of the document represents a notable professional accomplishment. I’ve been shepherding the BCP for the last eight months, over two major rewrites and any number of less invasive drafts, and I’m very proud to have played a useful role in what I think has been an important collaborative effort among my professional peers. I sincerely believe that all of the folks who contributed so much of their time and expertise to the effort have much to be proud of here.

When Sender Scores Don’t Matter

E-mail marketers and other senders are often flummoxed by an apparent disconnect between the Return Path Sender Score on their outbound IP addresses and their own deliverability metrics. A frustrated ESP sales team member posed me the following question this morning that illustrates this rather common complaint:

A client received zero complaints on a campaign to 30,000 recipients, but saw a 80% soft bounce rate at Why would this happen? The IP reputation is solid! (98 score). Thoughts?

Continue reading

Half-True Predictions

The thought of blogging predictions for the New Year in Deliverability makes me cringe. I’m not exactly sure why; it may have something to do with the fact that everyone else has already done it – we’ve had a bumper crop so far this year. Mostly, I think, I’d hate to be proven wrong later. So, instead of offering my own predictions (which would doubtlessly sound much like anyone else’s), I’d like to take a look at two different trends in deliverability that came only half-true in 2010, but that are still worth your time to continue to watch in 2011. The two trends are Domain Reputation and Engagement Metrics.

Continue reading

Shaking Our Fists at E-mail Gods

Last week, Return Path released the results of their bi-annual deliverability benchmark report, and the news for senders is not encouraging.

Across Europe, only 82.2% of legitimate marketing e-mails reached subscribers’ inboxes, down from 85.4% reaching the inbox in the second half of 2009. More than one in eight commercial e-mails (13.6%) are going missing completely – not in subscribers’ spam folders or inboxes, blocked by ISPs before reaching their subscribers – compared to one in nine e-mails (11.1%) going missing six months ago. Moreover, the number of e-mails going straight to spam folders increased to 4.2%, up from 3.6% six months ago.

Return Path hasn’t yet released updated figures for North America, but in the last half of 2009, the numbers were similarly grim: 20% of e-mail in the United States and Canada is still not making it to the inbox while 3% of e-mail goes to the “junk” or “bulk” folder and another 16% goes missing altogether. Only incorrigible optimists are expecting anything other than worse news when the updated numbers are released.

What’s driving this steady erosion in overall deliverability performance? It’s tempting to place the blame squarely on the shoulders of the ISPs that do the actual filtering, but that’s not the whole story. By some accounts, up to 94% of the total, unfiltered e-mail volume worldwide is comprised of spam, most of which is generated by armies of compromised computers under the control of criminal spam gangs. ISPs are forced to use ever more aggressive tactics to protect their users from spam, and make sure that just the wanted mail is delivered. If an ISP can’t strike the right balance for their customers, recipients can change providers, often in just a few clicks.

This is an important point that’s often lost on senders who find themselves on the wrong side of the ISP’s filters: delivering mail that customers want is as much of competitive advantage for an ISP as blocking the spam. Within the last few months, major inbox providers like Hotmail and Gmail released new versions of their offerings that measure how recipients interact with  your mail – whether they open it, whether they reply to it or move it to a different folder, even how much time they appear to spend reading a particular message. (Yahoo! and AOL are preparing similar changes in the coming months.) The more engaging and relevant your mail appears to be, the more often it will land in the inbox.

It takes work to find out what engages your recipients – typically an iterative process that requires testing and careful measurement – but it’s a lot easier than convincing an ISP that your mail is wanted in the absence of supporting engagement metrics. One easy way to find out what recipients want is to ask them. Use a survey tool to find out if they want to know about sales or specials, industry-specific news, free webinars, or some mix of all of these – and how frequently they want them. Ask them about other content-areas of interest. If you’re a retailer of Irish-themed gifts an apparel, for example, readers might want to see short vignettes or links to stories about Ireland next to your offer. Then deliver what they ask for (and nothing else!).

E-mail frequency plays an important role in engagement, too. Use recipient-level tracking  to identify those recipients who haven’t opened or clicked in a while – they may be suffering from “inbox fatigue”. Give them a rest from future sends for a while, and come up with a sharp offer – like a coupon code or a discount for your webinar – to re-engage them with later. Avoid recurrences of fatigue with a subscription preferences page where they can specify their preferred frequency. If they still show no signs of engagement, you should suppress them from future sends.

Measure response to different subject lines and calls to action. Sometimes changing a single word in your subject line can make all the difference in open rates. Test different versions on small groups and identify the changes that result in the biggest gains in opens, and roll them out to the rest of your recipients.

It’s tempting to shake one’s fist at the sky and bemoan mistreatment at the hands of uncaring e-mail gods, but senders really are in control of their own destiny. You have the tools you need to make optimization for delivery and engagement easier than you think.

Are You Certifiable?

Author’s note: Since the publication of this story in August 2010, Goodmail has announced that they will be ceasing operations as of February 8th, 2011. I’ve left the references to Goodmail intact in the post, in hopes that the treatment of similarities and differences between Goodmail and other services may still be of use to the reader. –AJB

It’s no secret that ISPs and large inbox providers often rely on accreditation of a sender’s practices when making filtering decisions. Accreditation providers are the conceptual inverse of a black list – the two most widely-used even refer to themselves as “white lists”.

Today, there’s a range of certification and “seal of approval” programs available to most permission-based senders. Clients occasionally come to us looking for guidance on whether the expense associated with third party certification is worthwhile. The answer is, “it depends”. Here’s a quick run-down of two of the best-known players in the space, along with three newer entrants that may be worth a gander.

In the accreditation universe, ReturnPath’s certification program is the 800-pound gorilla, because they cover an estimated 1.8-billion e-mail inboxes. ReturnPath offers two levels of whitelisting: Safe and Certified. Both require an audit of the senders’ acquisition and sending practices, as well as a vetting of senders’ e-mail infrastructure. Participants in the program must use dedicated IPs for whitelisted outbound mail, and those IPs must have and maintain good reputation scores – senders can be suspended from the program if their reputation tanks after they’ve been approved.

Senders on the Safe list typically get delivery to the inbox, but with links and graphics off; senders on the Certified list get inbox with links and graphics enabled. Because they’re so widely used by receivers, ReturnPath commands a premium for inclusion on their whitelists – and that’s certainly a consideration for senders of any volume.

Another big player in the space is the CertifiedEmail whitelist offered by Goodmail Systems. This program offers services similar to ReturnPath, but with a slightly different angle. Partner ISPs who agree to use Goodmail also agree to allow certified mail a free pass through all of the ISPs other filtering mechanisms, and to deliver it with links and graphics enabled. However, only opt-in and transactional e-mail is eligible for the program. Prospecting or acquisition mail and opt-out mail will not qualify.

Goodmail is an attractive proposition for senders, but it also comes with a big price tag – so big, in fact, that the Goodmail web site warns its prospective customers that it may not be worth their while if they’re sending to fewer than 50,000 recipients per month, and with at least 15% of that volume to recipients at Goodmail partner ISPs. The network of partner ISPs includes some very big e-mail inbox providers, like AOL, Verizon, and it’s affiliated domains, and others, but no longer includes Yahoo! after a falling-out this past winter.

SuretyMail, a relative newcomer to the space, positions itself as a low-cost alternative to the first two. However, SuretyMail is not a whitelist per se; rather, it certifies as many different attributes – good, bad and indifferent – of the sender’s e-mail as it can verify. For example, if a sender uses only opt-in, SuretyMail will certify that in the form of a response to an automated query by the receiving ISP. If mail from an IP is coming through a social networking service, SuretyMail will certify that, too. The idea is to provide the querying ISP with enough hard data about the mail to make its own automated delivery decisions. While SuretyMail can’t guarantee preferential delivery, it does promise significant improvement in deliverability metrics, and a money-back guarantee.

Unlike Goodmail and ReturnPath, SuretyMail doesn’t perform an advance audit (though it will do a background check of historical sending practices). SuretyMail can also monitor senders’ feedback loops to keep tabs on any changes in sending practices and reputation.

A seal of approval program called “I Don’t Spam” launched this spring, offering senders the opportunity place the company’s Spam Free Seal on their web sites. Prospective subscribers who click on the seal will see a count of the number of verified spam violations the sender has had, though the violations “age off” the tally on a rolling six-month basis.

The company has done some split testing to show significant increases in the number of subscribers and conversions for seal-bearing sites versus a control group; however, they don’t mention what gains in deliverability participating senders enjoy. I think senders shouldn’t expect any, as the program is only subscriber-facing. ISPs can’t know whether mail is coming from a seal-bearing sender – and that’s probably a good thing, because the program’s definition of spam falls quite a bit short of most ISPs’ operational definition. Permission is never mentioned in the program’s participant requirements, which instead seem to focus on CAN SPAM compliance. On the other hand, costs are low, with a monthly fee that’s dependant on the number of sites on which the sender displays the seal.

Last and most recently comes an announcement from CRM-provider and ESP RatePoint, who are planning to port the venerable VeriSign Trust Seal over to their own offerings, and re-brand it as “SafeSender”. It’s the first time VeriSign has partnered with an e-mail provider to use the familiar red checkmark seal in the actual e-mail creative. But it wasn’t clear from the announcement exactly what RatePoint and VersiSign will be certifying – are they asserting some level of compliance with best practices, or merely authenticating the sender?

I put the question directly to a RatePoint pre-sales engineer on the day of the announcement, and after putting me on hold for a few minutes, I was advised that RatePoint “would not be talking about that” until product release next quarter. Since my call, though, other industry publications have written that the seal will indicate to recipients that the sender has been authenticated, and that the message has passed a VeriSign malware scan. This makes sense, since VeriSign earned it’s original fame and fortune in the SSL certificate business (which it recently sold off to Symantec).

So how do senders decide whether and which program they should participate in? Price will certainly play a big part in any decision, insofar as cost of preferred delivery offsets any gain in ROI. Seal programs seem like a cheap alternative, but they are not true deliverability solutions.

The conclusion I draw is that there’s just no shortcut around good sender practices: send the mail your customers want, and only to those who asked for it. If you can do that well enough, you may find you don’t need any of them.

The Untethering of Reputation

If you’re a loyal reader of Steve and Laura Atkins at the Word to the Wise blog, you may have read Laura’s recent posts about sender reputation.  The posts detail what reputation is, how it’s used, and why senders need to manage it carefully. ISPs monitor the reputation of IP addresses from where an e-mail originates in order to make decisions about deliverability, and whether it belongs in the Inbox or the Junk folder. As Laura points out, reputation is a simple but extremely important concept to understand.

Over the last few years, both ISPs and senders have been adopting the use of what are known as sender authentication protocols, like DKIM (and, perhaps to a lesser extent, SPF and SenderID). This trend has the potential to make important changes to the way sending reputation is created by senders and assessed by receivers.

Today, sender protocols are used to make verifiable assertions about sender identities, or their authority to use a particular IP to send e-mail. While ISPs tend to rely heavily on the reputation of the e-mail’s originating IP address when they make delivery decisions, sender protocols can also be used to untether reputation from the sending IP, and tie it instead to a domain (like

This is what we in the e-mail space technically refer to as “A Big Deal”. Why? Think of what it used to be like to change mobile phone carriers before Congress mandated cell phone number portability. If you wanted to switch networks before the winter of 2003, you had to switch phone numbers. Coworkers, friends, relatives, and vendors all had to be notified of the change if you wanted them to stay in touch with you. It was a royal pain — painful enough in many instances to keep folks from switching carriers in the first place.

Now fast forward to a future when ISPs give domain reputation as much or more weight as IP reputation. If you’re a sender getting ready to ramp up your small in-house program and migrate it to an Email Service Provider (ESP), you get to keep the great sender reputation you’ve built thus far, even though you’ll certainly be sending from different IPs. It’s like taking your number with you when you switch networks.

There are a few more benefits worth mentioning, for both legitimate senders as well as for ESPs. Broader adoption of sender authentication protocols should make it easier for ESPs to use precious IP space much more efficiently. If you’re a sender who’s sending from IP space shared with other senders, their mistakes should have much less of an impact on the deliverability of your e-mail.

So, what should you be doing right now about reputation? Take Laura’s advice: keep doing the right things in terms of relevant, engaging content and best sending practices, and your reputation can only shine. Domain reputation will play an increasingly important role in delivery and spam filtering as its adoption rate continues to ramp up. It gives senders more and better reasons to adhere to best practices, and has the potential to give ISPs more accurate information for making better delivery decisions.