As a member of the Greater Frederick Advertising Federation board for the past four years I’ve had the opportunity to engage in a number of debates about various topics related to email and I thought it might be a good exercise to go back into our email data and answer a few of those questions Starting out, I knew our current system did not offer the kind of robust reporting I was looking for to help analyze and answer those questions, so, with a quick dump from MySQL and some help from excel I was able to crunch the numbers and I began to see some of the answers I was looking for.
Our system during that time was a popular email messaging system called PHPLIST. It’s free and lightweight, which makes it perfect for starting non-profits. But for what you trade in price, you make up for with labor because the system requires a significant investment in time for each email sent. In addition, with email, you can only track what the user gives you. In other words, we need the user to “do” something before we can collect any data. Links must be clicked, images approved, or emails forwarded, before the system can track any metric of success for each email. This limitation defines the current state of email metrics, and at this point seems to be enough information to produce some basic assumptions about our users.
In this exercise, I wanted to focus my questions on some of this basic information and really lay to rest some of the mysteries pondered by the board.
- Which days of the week are more successful that others?
- Which time of day, morning of evening has better results?
- What type of content yields the greatest success?
In this exercise, success is going to be counted in views. Views, by the definition of the system, are each time the user opens the email in the email application, pending this person has approved the use of images in their email client. Many other events you can track such as links, or clicks, and views all produce different insight into our users, but views would give me a general idea of the usefulness of an email to our members.
Here’s a look at the results based on views:
Our overall view rate was 36.339% for 171 messages sent. So, the following numbers float around this number with slight variation to present the anomalies or possible trends in the message type or timing used in sending the message.
% of viewed messages by day:
From the data above, we see that the later in the week, the lower the overall view rate. This gives a small indication that Mondays or Tuesdays might be the best days to send email to the membership. All though this seems surprising, this may also be affected by emails being retained in a recipient’s inbox in the beginning of the week, which would explain for additional views throughout the week.
% of viewed messages by hour:
In the results by hour, I see a very interesting pattern that defied my expectations. It seems that the morning does present a good opportunity to reach our readers, but not nearly as much as the late afternoon, with a peak being reached at 2pm followed by a decline of 7% between 3 and 4, with a reemergence of the apex at 5pm. This points to a higher view rate in the later part of the day overall, than in the morning hours.
By Content Type:
This last filter was intended to see if certain elements in the messages could produce a higher email view rate, with the hopes of conversion. With just about every email having a link, the link view rate was almost the same as the overall average. The drop in views for emails with “?” in the subject lines, either points to SPAM protection, or a general disinterest in questions in subject lines. Emails with images did receive a slightly higher view rate than those without. Overall, I do not see a notable trend in any case except those with the question mark.
The results of this exercise defied my assumptions about our users, and I’m sure will enlighten the board to the best way to carry out email campaigns in the future.
- Overall, a 36.339 percent view rate for 171 emails over 25 months
- Monday had the highest view rate, followed by Tuesday
- 2pm in the afternoon presented a sharp peak in the view rate
- Emails with images and direct subject lines presented a slight gain in views