I have seen a steady rise in both the use and the variety of special characters in email subject lines in my inbox over the past few months.  The first few times I saw a special character in a subject line, I thought, “wow, that was neat and it rendered!”  During my years in the trenches running day-to-day operations for email marketing, I had run into problems rendering special characters, especially across platforms.  It seemed that while most email clients supported some special characters none of them supports all of the characters.  I also worried that special characters may trip spam filters and result in lower inbox placement or non-delivery of messages.


Recently I started using EDS Analyst from eDataSource, which is a great tool for collecting data on email programs.  EDS Analyst provides subject lines, read or open rates along with data on inbox placement overall and split out by major webmail providers.  It truly is a treasure trove of data for email geeks like me.  After working in their system for a while, I thought now I can finally answer whether special characters truly help or hinder open rates as well as impact deliverability.  I began pulling data on major brands that I knew used special characters in their subject lines, primarily in retail or travel industries, both heavy users of special characters.


I looked at more than 7,000 campaigns sent between November 2012 and May 2013 and began to tag which campaigns had special characters in the subject line.  I ended up with a sample of the 7,263 campaigns; 889, or 12.2%, of which used a special character in the subject line.  The campaigns from the travel industry were the heaviest users of special characters, primarily using the airplane symbol, followed by star.  Retailers were more cautious in their use of special characters and primarily used Hearts, followed by various stars and starbursts.


I ran several tests to see how special characters affected subject lines and deliverability.  To start, I just looked at some trend lines for the deliverability scores and read rates.  These two resulting graphs came out of that analysis shown below.

One interesting point I see in these graphs is that while read and overall deliverability rates have remained constant across the sample for subject lines without symbols; there has been steady decline since the holiday season for subject lines with symbols.  I thought quite a bit about this observation to formulate an opinion on why this has happened.  I believe that as the holiday season ramped up, more and more marketers tried using symbols in their subject lines.  As symbols became more commonplace and in some cases, perhaps over used, consumers either noticed them less or began to regard them as a gimmick instead of being creative.  Hence, their positive effect of producing engagement was reduced.  As engagement waned, the negative effects start to outweigh the positive effects of using symbols in subject lines producing the downward trend for the overall deliverability and read rates.


Next, I looked to see if there were any correlations between the presence of symbols in subject lines vs. read rate and overall deliverability.  I found negative correlations between the variables.  Finally, I looked at the range and variance in read and overall deliverability rates between the two groups.  The range for subject lines without symbols was higher overall, and the variance in the range was lower than subject lines with symbols.  Also included in SPSS output was the mean (average) for both overall deliverability and read rate (which was lower for subject lines with symbols) but I disregarded this result because averaging several rates rather calculating the Average Rate is not mathematically correct.  Finally, I ran some tests to see if there was a statistical difference between the rates for the two main metrics reviewed and I did find a statistical difference between the two using a two-tail at a confidence level of 95%.


Based upon all of this data my conclusion seems clear-cut; that overall, the use of symbols in subject lines has decreased overall deliverability and read rates.  While this may seem obvious based upon the sample and analysis shown above, it may not be so straightforward.  I suspect the frequency that symbols are used and the type/tone of the brand’s communications also play a role in the effectiveness of symbols in subject lines.  In my next post, I will look at the differences in email performance within specific campaigns rather than across the campaigns to analyze on a deeper level the effects of symbols within subject lines.


Comments are closed.