January 22, 2008
Case Study
SUMMARY:
Don't take every fact and figure you read as gospel. Marketers need to rely on their own tests to determine important email elements, including when's the best time to send your blast.
An eretailer wondered how much of a difference they could make by testing the time of day to send their blasts. Turns out a lot -- clickthroughs increased 15.63% and revenue jumped, too.
|
|
CHALLENGE
Matt O’Laughlin, Marketing Business Analyst, Second Act LLC, was brought aboard the electronics eretailer last year to modernize email marketing. At the time, their program amounted to a monthly newsletter blast.
O’Laughlin and his team started applying best practices to their email right away. They revamped their email signup page to include 16 preference selections and began segmenting their database accordingly. Then, they ran a series of best-day email tests.
When the results came back, they were surprised to find that Tuesday beat Monday and Wednesday for opens and clickthroughs. But O’Laughlin didn’t want to stop there. He wondered if time of day would make a difference, too.
CAMPAIGN
O’Laughlin knew a big portion of Second Act’s orders were being placed during regular business hours. He and his team decided to zero in on their at-work consumer audience.
Here are the four steps they followed:
-> Step #1. Use analytics to determine busiest time
To get started, they checked for their website’s busiest windows of time and verified what they suspected: The site was most active during the 9 a.m.-5 p.m. workday.
“Morning and noon were peak times for unique visitors,” O’Laughlin says. “We also wanted to see how another time that saw fewer uniques would affect clickthroughs.”
The difference in traffic wasn’t tremendously different between 9 a.m. and 30 or 60 minutes later, for instance, so they didn’t test slight variations. Instead, they used analytics to boil down the test to three different send times (in Central standard time). Their rationale for each:
- 9 a.m.: People are just getting settled at their desks and haven’t locked in on work yet.
- 12 p.m.: When most recipients begin their lunch hours and often make orders online.
- 4 p.m.: The time O’Laughlin wanted to test to determine whether email would perform better than the website at this slot -- when workers begin winding down their days and start thinking about home and their family needs there. The team speculated that a little call-to-action email might pick up sales.
-> Step #2. Segment the list
The team made use of recently instituted preference segmentation. O’Laughlin took a file of consumers who signed up specifically for weekly coupon alerts and split it evenly in an A/B/C fashion. This file made sense because it was filled with names mostly from the Central, Eastern and Mountain time zones (in that order).
This was important because O’Laughlin and his team weren’t ready to segment separately to all four U.S. time zones; all of their campaigns in the near future were going to be sent to an entire file at the same time.
Hence, it seemed smart to concentrate on the three time zones that were not only most prominent in their database, but also “closest together” on the world clock (e.g., 9 a.m. Central is closer in the work-day sense to both 10 a.m. Eastern and 8 a.m. Mountain when compared to 7 a.m. Pacific).
-> Step #3. Keep variables consistent
Because they wanted time of day to be the only variable, each file received the same subject line with the same message and the same offer (Apple iPods sale).
-> Step #4. Pay attention to what matters
O’Laughlin also focused only on clickthroughs. Opens were inconsequential, he says, because the subject line was the same for all three splits. Because the sophistication of their landing pages was still in their infancy, “what was most important to us was to get people from the email message to the site.”
RESULTS
Test returns produced a clear winner -- 9 a.m. The clickthrough rates for each:
o 9 a.m. performed 15.63% better than 4 p.m.
o 9 a.m. performed 9.4% better than 12 p.m.
o 12 p.m. performed 6.9% better than 4 p.m.
“As for email-driven revenue, the test was sent on Aug. 30th. Then, 85% of email-driven revenue for 2007 occurred between September and the end of December,” O’Laughlin says. “So, yes, this definitely increased revenue.”
Because the only variable was time of day, the results were conclusive enough for O’Laughlin and his team to implement the 9 a.m. mailing tactic as an important strategy from this point forward, although they plan to continually monitor with more tests down the road. In the end, his hunch that people at work most often read their emails first thing in the morning turned out to be correct.
“A lot of people right after they get into work, they check their email, log onto ESPN.com or their favorite news source and take care of those things before digging into their professional tasks.”
Useful links related to this article
Creative samples from Second Act’s email test:
http://www.marketingsherpa.com/cs/secondact/study.html
Past Sherpa article - Organic Dish’s day-of-the-week and subject line tests:
http://www.marketingsherpa.com/article.html?ident=30280
Silverpop - Second Act’s email service provider:
http://www.silverpop.com
Omniture Inc. - Second Act’s analytics service provider:
http://www.omniture.com
Second Act LLC:
http://www.secondact.com
