First, let’s lay to rest the myth of hard and soft bounces. There really are no such things, though the terms are convenient to describe what happens to e-mails that don’t go through.
Technically, there are permanent and transient e-mail failures.
A permanent failure is one whereby the receiving server immediately recognizes that delivery will be impossible for that specific message. This might be because a mailbox is full, or that e-mail address no longer exists (or the address has been incorrectly spelled). It also might be because the e-mail server is (perhaps temporarily) misconfigured.
Temporary failures are those whereby the destination server cannot process the delivery request at that moment, but will continue to attempt delivery for a given period. Only when this delivery timeout occurs will the server send an unsuccessful-delivery notification. Examples might be a server that is temporarily overloaded, misconfigured or where the connection to that server is temporarily unavailable.
Hard and soft bounces are more of a marketing description of the longer-term availability of an e-mail address. For example, an out-of-office notification is usually called a soft bounce because the user will return to availability soon. However, technically this is simply one of several e-mail replies that can be sent by a recipient – in this case automatically. It is neither a permanent nor a transient failure.
Remember that e-mail was never designed to perform the role we ask of it today. It was designed to deliver simple text-based messages to recipients. It wasn’t intended to provide a means of clearly and reliably identifying the sender and permitting detailed receipt notification and reporting.
Wouldn’t it be great if you were able to tell what had happened to every package that was delivered in your latest promotional print mailing? You’d know not only about the successful aspects of the campaign, like who responded and whether they showed the offer to a friend, but also what happened to the packages that didn’t make it.
With e-mail marketing you’d better prepare to get your hands dirty. Unlike postal direct marketing, where you are unlikely to learn the fate of mail pieces in detail, e-mail lets you learn more than you probably wanted to know about the unsuccessful aspects of delivery.
You will hear from customers who got duplicate e-mail messages; from those who think you are spamming them; from those who cannot view the message correctly; and from those who failed to receive it at all. Your customer service staff needs to be prepared to respond to these customer inquiries.
Let’s look at the main areas glitches occur: message delivery and message creation.
The first category includes everything that happens from the point you hit the “send” button and your message wings its way across the Internet. The second involves the idiosyncrasies of creating and displaying content that needs to be compatible with the huge variety of e-mail browsers, and in particular the use of HTML-formatted messages.
E-mail is not like postal mail, where we have a well-established, simple and consistent process that hopefully ensures that you can create a package and reliably deliver it to the recipient. The Internet is a still-evolving collection of hardware and software technologies, operating over a Web of different data communications systems and protocols. As a result, any number of combinations of technology-related issues can conspire to screw up delivery of a message.
E-mail uses a “store and forward” process. Messages do not magically get handed over directly to the recipient’s inbox, but undertake a sometimes-tortuous journey between computers. En route a message is delivered from your PC to your ISP’s e-mail server, and from there via several other e-mail servers to the recipient’s e-mail server.
Sometimes one or more servers in the chain is temporarily unavailable – perhaps one is overloaded for a while; perhaps it is offline for servicing; there may be Internet network congestion; or perhaps a data cable has been cut by a backhoe! For many reasons, e-mail can take awhile to reach its destination, and it is impossible to predict how long it will take to deliver every message.
Here are some other reasons message delivery might fail, or be delayed:
o The intended recipient might have changed e-mail addresses and forgotten to notify you.
o If you send messages often enough, or in a high enough volume, the recipient’s ISP might automatically invoke spam blocking, thus preventing the message from reaching the recipient’s inbox. Or, it might be diverted automatically to a bulk e-mail folder, which the user has to search separately.
o In some cases, virus-scanning software on your customer’s PC may erroneously block or discard HTML messages, thinking they contain viruses.
o Occasionally, messages simply get lost on the Internet because of a system or network failure somewhere along the delivery chain. More often, however, it will simply be delivered later after a number of attempts by the system, once the temporary network glitch is fixed.
A recipient might get multiple copies of the same message for these reasons:
o Failure to de-dupe your customer e-mail address list prior to delivery.
o If you use more than one vendor to deliver your e-mail, each might have a record for the same customer on their unrelated lists.
o The most common cause is a misconfigured e-mail gateway at the destination ISP. Typically, the recipient will see all his e-mail duplicated within a given timeframe.
o A rare timing error whereby a message was successfully passed along by a mail server, but received no delivery confirmation from the receiving mail server, and so re-sent the message. This confirmation process occasionally fails if either the sending or receiving server has a temporary technical glitch during the send and response cycle.
How can content creation and formatting lead to problems? When it was conceived, e-mail was capable only of handling simple text messages. In due course, extensions were added to enable the transport and delivery of more advanced content including attached documents and HTML-formatted messages.
These extensions (known as multi-purpose Internet mail extensions, or MIME) are, as far as possible, backward compatible with the old text-only system.
However, this transition has not been perfect. There are still e-mail browsers that are incapable of handling MIME messages at all, while others may be capable of handling attachments but cannot display HTML. This history leads to the majority of issues with the presentation of HTML messages.
These issues can be minimized either by sending text-only messages or by delivering a multi-part/alternative MIME message.
This involves sending a message that includes a text version and HTML version. The receiving e-mail client displays only the HTML version, if it is capable, or the plain text version if it is not. There are exceptions to the rule, but this method ensures a high success rate and optimizes delivery format for the vast majority of recipients.
Most of your customers are likely to use Microsoft Internet Explorer or Netscape Navigator, and both platforms offer predictable, reliable presentation of HTML content.
AOL is a different matter. Until the latest version, 6.0, it used a unique version of e-mail formatting, which has required a separate AOL-specific version for each message. AOL 6.0, however, fully supports HTML. The current versions of Microsoft Outlook and Outlook Express display HTML correctly, but older versions of Outlook displayed only the plain text portion of a multi-part message. Finally, there are e-mail clients such as Pine, and older versions of Outlook, Lotus Notes and Eudora that incorrectly display HTML or cannot handle it at all.
Much also depends on the bandwidth and versions of the e-mail browsers used by your customers. Those with a cable or DSL modem are likely to be able to receive rich media enhanced e-mails complete with streaming audio and video, or Flash animations. Those with dial-up modems are less likely to appreciate the degraded quality of their low-bandwidth versions of rich content.
Other issues arise when you try to deliver messages in foreign languages or include 8-bit characters such as © (the copyright symbol) within the message. Such 8-bit characters require an added layer (and subsequent unpredictability) of message encoding and decoding by the recipient’s e-mail client software.
Use of JavaScript in an HTML message can cause problems, too, because many e-mail servers are configured automatically to reject messages containing JavaScript (which is a common method for propagating e-mail-borne viruses). Others might try to “de-fang” the JavaScript, resulting in the message not being delivered as you intended.
Your customers might complain of links within the message that do not work. This can happen for several reasons:
o The URL link within the message might be wrong, pointing to the wrong page or a broken one.
o The linked Web page might not yet have been uploaded to the Web server, or might have been moved.
o If a URL within a plain text message was too long, it might have been wrapped onto a second line by the e-mail client software and therefore doesn’t work correctly.
o Sometimes, because of encoding that occurred during message relay and improper decoding on receipt, your links do not work.
You can never expect 100 percent delivery of any e-mail campaign. But you will get a very accurate picture of where the errors occurred, and you can forearm your customer service staff to handle some of the questions that will arise.
Though it can cause an unexpected level of novel, and unpredictable, customer inquiries, you should view this extra accountability as an asset. After all, when was the last time the U.S. Postal Service provided you with a blow-by-blow account of its failed deliveries and gave you the chance to improve the customer experience?
