In 2004, eROI and MarketingSherpa made waves by attacking conventional wisdom and suggesting that Monday might be a better day for some marketers to send email than the traditionally favored midweek hump.
This week eROI released its data for October-December 2004 and MarketingSherpa got an early peek.
The bottom line? Performance by day of week is a moving target that may be affected by seasonality as well as overall email send trends. See the link at the end of this article for fascinating charts comparing send volumes, open rates, and click rates by day of week across the three quarters -- from May to December 2004 .
The research is based on open/click and send data for more than 60 million emails sent by more than 7,000 different marketers using the eROI system.
Clicks are defined as percent of emails sent minus hard bounces. (Some people call this "emails delivered"; however, we don't because it doesn't take filters into account.) Unless specified as Pacific Time, times of day were generalized to reflect whatever time zone the recipient is in.
It's not a strictly scientific study but more of an indicator showing trends you should consider when planning 2005. Here are our quick notes on the results.Monday -- high opens, higher volume and so-so clicks
When we last looked at the data, we saw that Mondays had strong open and click rates but represented only 8% of weekly volume. By the fourth quarter, Monday's volume had grown to almost 15%. In part that's due to increased frequency overall as well as the impact of the initial study, which opened the door for eROI's customers to Monday as a viable day for email.
Monday had the highest open rate of the week at over 32%, but a 4% click rate that was third, behind Tuesday and Thursday. The high open rate may reflect that many people get ready for their week by clearing their inboxes and look through the weekend's email.
The low ratio of opens to clicks -- 12.5% -- seems to be related to two things. First, lots of weekend spam. Second, it may be that people are reading emails but don't have the time to act on them on Monday, leaving that for later in the week (especially when it comes to shopping at work, which seems to happen on Thursdays).Tuesday -- the heavy lifting
In the first half of last year, Tuesday was the most popular day to send email, with almost 29% of total volume. In the fourth quarter, Tuesday had dropped to the second most popular day with 19% of total volume. Its 23% open rate is third behind Monday and Thursday.
The CTR on Tuesday is 4.3%, second only to Thursdays. Interestingly, the click to open ratio on Tuesday is the highest of the week, with a higher percentage of people who opened the email going from open to click.Wednesday -- high volume and lackluster results
More than 21% of weekly email volume went out on Wednesday in the fourth quarter, the highest for the week but down from 25% earlier in the year.
The click rate was only 3.3%, fourth for the work week and just ahead of Fridays.
The expression "email fatigue" has often been attached to Wednesday, with an assumption that by Wednesday the sheer volume of emails has taken its toll on interest.
But the results may be more about Wednesday itself than its place in the trend of the week, since Thursday shows an increase in clicks. Thursday -- It Clicks
It's often assumed that greater volume decreases interaction, since recipients must choose among more messages. But the data for Thursdays suggests that may not be the case. It's the third highest day by volume, with almost 16% of the total, but has the highest click rate of the week at 4.7%, representing nearly 20% of all clicks. Mills believes that there is a natural tendency for emailers to want to put their messages in front of users as the weekend approaches, and most people do their "real world" shopping. Friday -- it doesn't click
Even though the volume numbers are only slightly lower than Thursdays, the open and click response is poor. Opens come in at 21%, and clicks at only 3.2%, both lows for the workweek. Fridays also tie Wednesdays for the lowest click to open ratio at 15%. In this last fourth quarter, there were three Friday holidays (or Fridays next to holidays, at least) so that may have something to do with it. The data is in aggregate, so we can't be sure.Weekends -- Gaining ground
In previous quarters, weekend email volume made the data difficult to analyze, with wide variations in click and open rates. But with the large increase in volume, the numbers seem to have normalized. Both days have low open rates around 16%, but the click rates are substantially different, with Saturdays getting the week's lowest at 2% and Sunday getting about 3%. In terms of click to open ratio, Sundays are second only to Tuesdays at 18%. So while fewer people are opening their emails, a high percentage of them act on the messages.How seasonality affects email response rates
We know from research and common sense that the volume of eretail emails skyrockets in the fourth quarter. Occasional emailers save their campaigns for the holiday season, and those who have regular email communications ramp up their frequency. Fortunately, consumer interest seems to rise as well. Fourth quarter open and click rates were lower than the rest of the year, but not by much, even though volume was approximately 29% higher than previous quarters.
Perhaps the most remarkable finding was the degree of change. Most email metrics move slowly into new trends, but in Q4, mailers and recipients behaved very differently than they had in previous quarters.
You can tell a lot about data from the shape of the curve formed by the numbers. In previous quarters, the graph of email volume looked like a witch's hat -- flat volume for ends of the week and a spike in the middle. However, in the fourth quarter that trend changed sharply as marketers seemed to explore the whole week.
"During the holiday season marketers are looking for every opportunity to connect with customers," said Jeff Mills, Senior Analyst for eROI. "Their frequency of mailing increases and they experiment with mailing on weekends, Mondays, and Fridays to differentiate themselves from the barrage of email marketing."
In fact, in the previous quarter the volume of email sent on Saturday and Sunday represented around 5% of the total weekly volume; in Q4, the two combined for over 15%. However, overall performance on weekends lags behind the work week with lower click and open rates. Interestingly, Sundays did have the highest click to open ratio, a metric that can describe how relevant the messaging is to the consumer.
Mills suspects that, during the holiday season, Saturdays are "action days" when people are out of the house, shopping in the real world. Sunday may show better results because people are at home and more likely to have the time to consider online shopping.
Fascinating side note: one fourth quarter characteristic is less attention to deliverability. eROI's VP of Marketing, Dylan Boyd, notes that in that heavily trafficked period, a huge number of eretailer messages ignored some deliverability basics and had red-flagged words like "Free" and "Act Now" screaming from the subject lines. Seems that when the competition is especially heated, and consumers are ready to open their pocketbooks, some retailers throw deliverability concerns to the wind so they can be aggressive in their messaging.MarketingSherpa's five lessons
Because of the nature of this research, there's no way to separate out what kinds of marketers and campaigns saw success trying new days for their emails. But we do know that the numbers varied substantially from quarter to quarter.
Lesson #1. Experiment. Take random segments of your list and try different days of the week and different hours of delivery. Make sure you are sending them the same email for comparison's sake and that the segment is random.
Lesson #2. Factors change over time, so your opportunities are changing as well. For example, people may be in shopping mode during more days of the week during the holiday season but get back to business in January. Set up a monthly or quarterly schedule to test which day of the week performs best for your list segments and make sure to regularly create new groups for testing.
Lesson #3. It's worth considering whether the flattening of the curve of email response across the week indicates that people at work are becoming more willing to engage in personal shopping and business. It's a trend we've seen in other research, and we suspect that it's especially true in the weeks leading up to the holidays, as the office environment is typically slower and more relaxed than at other times of the year.
Lesson #4. Consider altering day of week for sends based on seasonality. This may not be smart for beloved content newsletters that readers have come to expect on a certain day, but it's certainly doable for sales alerts.
Lesson #5. Forget the fiscal quarter and focus on the season. In future reports on this topic, we'll not lump months into quarters that don't make sense together -- such as August and September. If time of year really has such an impact on emailer and recipients' activities, we should examine the data that way.Useful links related to this article
Here's the link to the comparison charts we promised you at the start of this article: http://www.marketingsherpa.com/eroi/study.html