Email

Email has become an accepted means of communicating with your clients and your clients are happy to communicate with you via email. This means that any email addresses on your website need to work effectively to ensure good customer service.

Email problems

There are several ways that problems with your email system can impact your business. There can be problems with failing email addresses or problems with your email servers that impact your users.

The most common problem is failing email addresses. These can arise by email addresses being removed from the email system when someone leaves or somebody decides to remove a generic email address (say, inquiries@sitemorse.com). Or perhaps the email address was misspelt. Either way emails from your customers won't get through.

Your email servers may also cause problems. Generally organisations have multiple email servers, e.g. a primary server and a backup server. Emails will always be sent to the primary server unless there's a problem in which case email is automatically routed to the backup email server.  Because temporary blips in service rarely cause a problem, as email servers will try to resend emails for some time before giving up, you may not be aware of problems. This is particularly true of your back-up servers that normally only receive emails when there is a problem with your primary email servers.

Sitemorse email testing

What do we test?

We test all email addresses included in "mailto:" links we find on the website (including any in PDFs that we find). We test all email servers that are defined within your email domain.

What sort of problems do we detect?

Amongst the common problems that we find are: typing mistakes in email addresses, email addresses that no longer exist, and mis-configured backup mail servers.  Some broken "mailto:" links can easily be detected - for example, "mailto:bill.sitemorse.com" is obviously incorrect because there is no "@" in the address. 

Mis-configured servers can be particularly hard to diagnose without Sitemorse's email testing, since the likely effect is that email will get through fine except when the main mail server is down when suddenly all email will start bouncing!

How do we test?

We look up each email address's domain in the DNS ("domain name system") to obtain the list of email servers for the domain. Each of these is the name of a server which will accept mail for the domain (e.g. "mail.sitemorse.com").

We connect to each of the mail servers and go through the "SMTP" conversation which computers go through when sending an email - stopping just before the point when an actual email would be sent. This establishes whether the email addresses are valid and whether all the email servers are capable of receiving emails.

Any error messages or problems seen are recorded.

Sitemorse email testing gives you the confidence that your email service is effective in serving your customers needs.

View larger "eMail Diagnostic" screen shot

1.1 - Summary of our findings

1.2 - The list of email address results

For each email address found, a status is provided - "satisfactory"," impaired" or "compromised" along with any error messages and where they appear on the site.

1.3 - The mail server detail

Shows status of each email server with any error messages and timings in graphical format

1.4 - "Test an email address now:"

Simple means fo testing email addresses to make sure they work OK.

1.5 - email syntax failures

1.6 - Other email domains found.

Clicking on a domain shows full details in right hand panel

A little background to the problem

Anyone that uses the MessageLabs service will know that we have been attempting to have some level of reasonable dialogue with them about changes they made to their infrastructure some time back in order to handle the deluge of Spam that was hitting their servers.

Unlike most of their competitors that either added more capacity or added more sophisticated front-end filtering MessageLabs placed a blanket block on any email directed to any server other than the primary server.  They also appeared to have removed the actual physical servers that were defined (and still are defined) as the fail-over email servers within a cluster of servers.

So when Sitemorse attempted to check that all the defined email servers were present, defined correctly and configured correctly we would get bounced.  Which resulted in us reporting errors for those email servers.  This resulted in some organisation's scores being impacted which upset several people as it was a problem they could do nothing to affect.

We have given up on being able to engage in any sensible dialogue with MessageLabs and so, with the advent of the updated email service that we introduced last week, we have changed the way that we designate these types of errors.

The idea behind our scoring regime is to penalise sites that have errors which WILL impact users of the site (broken links, failing email addresses etc) more severely than than sites with errors that MAY impact users (non-compliant HTML, missing Metadata etc).  We have reassessed some of our tests and recategorised them.  Some have a higher severity level and some have been lowered.  See my posting on the subject here http://blog.sitemorse.com/2009/01/some-tests-have-moved-between.html

The solution

We have reclassified the email server errors to have a lower severity level - because if a secondary or fail-over email server gives a problem email will still be being received by the primary email server so there is no immediate impact on a user of the site, this is just a warning of the potential problem. This means that the impact on a site's score is greatly reduced.

Hopefully this will resolve the issue that some of our users had with this problem.

If you have any questions on this please don't hesitate to contact me on gevans@sitemorse.com

We have now launched the new email service to all existing customers.

The service continues to check email addresses found in Webpages and PDF files to ensure that they are formatted correctly and that they will receive emails.  Failures are reported.

The new features of the service look at the email server infrastructure behind your email addresses to make sure that they are configured correctly and that the servers will receive emails.  We have come across numerous email server infrastructures that have either been incorrectly defined to DNS or are mis-configured and wont actually receive emails.  Email servers can have different roles, e.g. they may be primary servers, secondary servers, load balance servers or fail-over servers.  Each would be defined and configured differently.

This adds another layer of sophistication to our audits and ensures that you are confident that your email infrastructure is defined correctly and configured correctly.  We'll also report any availability issues or intermittent type problems if they occur during one of your audits.

We have decided that before gaining access to the reports at least one person from your organisation must attend a short (10-15 mins) on-line training session.  When you click on the email icon on the Summary Page of an audit you will, on the first instance only, be prompted to book on a session and provide your primary email domain.  This ensures that this domain is the first one we report on in the right hand panel of the page with other domains we find being summarised in the left hand column.

Here are the prompt screens you see the first time you click on the email icon.

If you have any questions on this please contact me on gevans@sitemorse.com

There are always lots of "scandals" reported about the level of service that organisations provide to their prospects and customers.  One of the measures used in the surveys is responsiveness to enquiries and one of the detailed measures is how long it takes organisations to respond to emails. 

Many sites have email mailto: links dotted around their sites.  Some of these are generic email addresses like sales@xxxxx.com and some are the email addresses of specific individuals.  The generic ones tend to work OK as they tend to change very infrequently but the specific ones need to be kept up to date especially if they are an important means of your prospects and customers to make contact with you.  It becomes even more important if this is a means for them to raise a concern or complaint with your organisation.  You will turn a concern into a complaint and a complaint into a rant if the email bounces and, as the surveys reveal, if the email disappears into a black hole and no one responds that raises the temperature still further.

Email is becoming a key element of most companies communications.  And it's not just the mailto: links on your own websites that are important.  There are  many on-line and off-line mediums that are used to communicate with prospects, customers, partners or suppliers and some of these may have generic or specific email addresses that are important to ensure that they are still valid.

But there can be other problems with email addresses other than they have been removed from the mail server's directory.  Many organisations have a limit on the size of the in-box associated with the email address.  Once this is full you have to start deleting or archiving emails to free up space before you can receive any more emails.  Incoming emails can be bounced under these circumstances and will certainly be delayed.

It's important, then to ensure that all of your "important" email addresses are up to date and working.  Sitemorse will currently alert you to any problems with your mailto: links on your websites and in the near future will be able to check other email addresses that you wish us to test.

Beyond pure email address checking there is also your email infrastructure.  This is often made up of load balanced primary servers, secondary servers and fail-over servers.  All of these are defined in the MX records of your DNS.  We currently test all the servers defined in a MX record for the sites we are testing and report any errors we find.  It's remarkably easy to mess up the MX definitions and render your mail server infrastructure inefficient.  We often highlight problems at this level for organisations. Mail server infrastructures often perform well under medium levels of load but with badly defined MX they can rapidly deteriorate when under load or failure situations.

This is another area where your customer service can suffer and needs to be checked on a regular basis.