6 common reasons for email delivery problems

Avoiding the accidental reasons why your marketing emails may not get through

The surprising thing about email deliverability is there are some common causes that are often innocent mistakes and can be avoided. The first two posts in this series covered how ISPs filter and how to monitor your deliverability.

In this post I'm covering some of the common issues I've seen happen, although they're not all obvious. I'll also give some tips on how they can be avoided.

Reason 1 Lack of transparency

High spam complaint levels occur when subscriber expectation is not managed correctly. This is either due to a lack of transparency about emails that will be sent or by setting an expectation and not adhering to it.

Typically this happens when an email address is provided for a specific purpose such as to get a quote, make a purchase, obtain a whitepaper, enter a competition or to create a logon account. The customer will expect the email address to be used for that purpose but will not expect to get other emails unless this it’s clearly signposted during the process of providing an email address.

Clearly signposting means being up front about the further emails that will be sent and giving the option for the customer to choose how their email address will be used. Relying on a statement deeply embedded within a privacy policy may make the emails legal but is a long way from acceptable practice in the eyes of customers.

Reason 2 Bought data

Bought data always carries a high risk and is generally considered a 'no no'.

Even expensive data from suppliers who consider themselves to have the best data quality assurance processes have bounce rates of 20% in my experience. This is well over the safe threshold. At the other end I've seen purchased data used from dubious sources with 70% bounce rate. As they say on children’s programmes “don’t try this at home”.

If you really must consider using third party permission data then look at data rental not purchase.

Reason 3 Additional data

Problems can occur when a significant amount of data is added ad-hoc from sources not normally used. This might be the result of a particular marketing effort to collect email addresses or from partner collaboration. Whatever the source it’s most often data that has been collected outside of the normal CRM system and subscription processes.

This means the care and due diligence in place for the day to day data collection has been bypassed. All of the good practices put in place are forgotten or may be unknown to the team involved in the ad-hoc data collection.

A variation of this issue is data re-discovery. Finding previous segmentation or data management errors has meant not all emailable customers have been emailed. The usual response is to simply add the email addresses back without a second thought.

The solution to reducing the risk in these cases is to test a few thousand email addresses coming from the sources not typically used to ensure open, click, bounce and complaint rates are at safe levels and in-line with usual values.

Reason 4 Segmentation errors

By this I mean an error in data segmentation that causes emails to be sent to the wrong people.

This can be a simple human error when creating a new segment, accidently wrongly defining a database query, or forgetting to apply the correct suppressions such as unsubscribed, hard bounced or spam complaint addresses. It can also result from changes to databases or upgrades to IT systems, which have a secondary effect of changing how selection criteria work.

To protect against segmentation errors do check your segment counts. If they look too low or too high, stop and investigate. Pick sample email addresses for those that should or shouldn't have been included in a segment and double check they are or aren't within the segment.

Reason 5 Automated data synchronisation complexity

All companies have data in many systems. For example, email address data held in spreadsheets, CRM solution, email solution, web system backend, transactional database, analytics tool etc.

The problem comes in managing data synchronisation. New email addresses may arrive from multiple sources; purchases, recommend a friend, subscribe forms, telesales. The same goes for unsubscribes; unsubscribe forms, telesales requests or due to hard bounce and spam complaints.

This means the theoretical simplicity of a single master data source with single direction data synchronisation is not possible. Data synchronisation needs to be bi-directional. This makes good data and permission management surprisingly complex. This becomes even more complex when several different types of permission are being managed.

If all scenarios haven't been considered and how data moves been systems not well defined, then failure to add subscribers or re-subscribing unsubcribers and hard bounced addresses results.

Reason 6 Email volume send peaks

A sudden and big change in the volume of emails sent each day. This tends to be more likely to impact larger list sizes where daily volumes are in the millions. Its best to spread the volume out over time to avoid peaks in these cases.

Take care, be safe and hopefully you won’t fall into any of these traps.

Tim, Email Crimewatch