Email subscriber engagement is increasingly important as a determinant of inbox placement and positioning. For email marketers, the challenge is to identify drivers of positive engagement, and applying that knowledge to create campaigns that their subscribers consider informational, educational, or amusing. Success is being measured by important new metrics such as “retrieved from spam”, “deleted after reading”, and “marked as not spam” as these metrics increase, so will program performance.
Subject line testing has always been a simple yet highly effective way of identifying positive behavioural drivers. The principle is easy – prior to the main broadcast, create two or more test cells, create different subject lines for a common email creative, identify which one performs best (opens, clicks, conversions), then apply the winning subject line to the remaining subscribers. I’ve regularly seen this approach driving positive uplifts in campaign responsiveness of 33% or more, when benchmarked against the worst-performing alternative.
However, even subject line testing requires a degree of organization that isn’t always available to email marketers, who are notoriously time poor. I’ve often found myself wishing for a tool that would answer those magical questions (“What’s the best subject line length?”, “Which words can’t I use?”, “Are special characters a bad idea?”) without running a load of tests first. Now, to my unalloyed glee, I have such a tool at my fingertips, and I spent a few happy hours last week using it to test (and deconstruct!) some conventional email wisdom.
To provide some context – I have recently worked with email marketing programs in the “Restaurants” sector, so I pulled a 30 day snapshot from our reporting network, comprising 826 discrete campaigns, from a range of casual dining vendors. The key metrics that I evaluated against were “Read Rates” (a highly positive subscriber engagement metric) and “Spam Complaint Rates” (highly negative).
The first test focused on campaign size. I categorized each broadcast as small, medium, or large, with a hypothesis that “smaller” equated with more targeted audiences while “larger” corresponded with more generic audiences. This was spot-on – read rates for smaller campaigns were higher than the sector average, but declined as campaign size increased. Spam complaints for smaller campaigns were below the sector average, but increased as campaign size increased.
I then used the subject lines to categorise emails by message type. A well-understood element of email best practice is to avoid bombarding subscribers with a steady stream of offers, and identifying other touch points to create a more varied subscriber experience. Again, the numbers backed up the rationale – on average, welcome emails generated read rates that were almost 3 times higher than the benchmark, with thank you emails, birthday emails, and anniversary emails all generally twice as effective. Birthday emails also carry an important lesson about relevance – complaint rates for this message type are twice the benchmark, suggesting that for emails where date of send is crucial to success, getting it wrong will be punished.
Using ostensibly “spam trigger” words (“Special”, “Offer”, “Free”) all delivered read rates 25% to 50% higher than the sector average. However, there was also a corresponding trend of increased spam complaint rates. Clearly, the use of these words generates something of a love-hate response from subscribers, and marketers need to evaluate whether the upside outweighs the downside for their programs.
Subject line length also demonstrated some clear variations in subscriber responsiveness. I categorised them as: short< = 25 characters; medium <= 50 characters; long > 50 characters. Short subject lines performed best, with a 10% uplift in read rates against the benchmark, while medium subject lines under-performed by a similar measure. Interestingly, responsiveness then started swinging back toward long subject lines. While a short, punchy subject line that doesn’t get truncated would seem the preferred approach, longer subject lines that carry more detail about the email’s contents are also effective.
Lastly, I evaluated the effect of commonly-used special characters on subscriber responsiveness. Top-level results were that using financial amounts is good (read rates up 10%), as is exclamation marks (read rates up 30%), but subscribers clearly don’t like being too challenged – read rates went down by 20% when the subject line posed a question!
All of these findings are a bit of fun, and I wouldn’t want readers to start quoting these numbers as the new industry gold standard. However, there is a serious point to be made – email subject lines represent the point of departure for any journey that leads toward improved subscriber engagement. Email marketing is highly competitive, and smart practitioners are investing time and resource to identify the approaches to which their subscribers will respond most positively. And, as with all competitions, to the victor go the spoils – in the form of improved deliverability, greater subscriber responsiveness, and increased campaign ROI.