by
Courtney Eckerle, Reporter
CHALLENGE
"We strive to always have our finger on the button of pop culture," said Jonny Cottone, Interactive Marketing Director, BustedTees, an Internet-based retailer of humorous T-shirts and novelty items.
Started in 2004 by employees from CollegeHumor.com, the pop culture part of BustedTees' business was flourishing with quirky and unique T-shirts.
However, its email user database was suffering from a generic send time for its product.
While email marketing has always been a top revenue generator for BustedTees, according to Cottone, he and his team knew they weren't maximizing the value of the channel. How to best serve a global audience was a growing challenge.
"BustedTees is a global business with customers located throughout the world," he said.
Historically, BustedTees delivered a single email blast to all email subscribers, no matter what time zone they lived in.
"Which translated to customers in L.A. receiving it in the early morning, [New York] mid-morning, London afternoon and Sydney the following morning," he said.
By sending out its daily email at a time that was best suited to an American audience, BustedTees was "underserving our international customers," Cottone added, and in order to achieve its goal of entering new overseas markets, it was necessary to "develop a more personalized email strategy if we were to achieve our overall goals of … growing the international side of the business."
CAMPAIGN
To solve this issue, as well as better serve its international customers, the BustedTees marketing team developed a plan to move towards personalization in its email program.
Working with a vendor, BustedTees started out by segmenting its sends by time zone, delivering emails at a reasonable time for all subscribers.
It decided to take personalization farther, and use its past data on individual open times to develop a personalized send time for each subscriber. This would allow the daily BustedTees email to reach customers at the time that was best for them — no matter the time zone.
After working with the sales department and determining a six-hour time period within all emails would be sent out, BustedTees had implemented the practice and seen an opportunity to grow and apply personalization along every step in the funnel.
Step #1. Segment subscriber list
BustedTees began by segmenting the subscriber list by time zone, which allowed delivery of emails at a fairly consistent time in each local market.
Adam Schwartz, General Manager, BustedTees, said for this switch-over from a generic send time to segmenting into time zones, "We send it 10 a.m. Eastern, and then it's just going out at 10 a.m. Eastern for everybody. Then … we started sending it at 10 a.m. for your time zone. So we would segment Eastern, Central, Pacific, a couple of international time zones."
"Despite this achievement, we recognized that we were still falling short of our personalization goal … we were still treating every subscriber in a given time zone the exact same," Cottone said.
Sending the emails by time zone gave BustedTees "a pretty small lift in open range, nothing too significant, but it seemed helpful," Schwartz said.
Seeing that small bump encouraged the team to pursue email send time personalization even further.
Step #2. Test for optimal send duration
In order to determine both the optimal duration and start times for personalized email blasts, Cottone and his team had to test both variables.
"First, we tested various durations ranging between six and 24 hours to see which generated the highest open rate based on each communication blast starting at the same start time. We then tested different start times, each with the same duration as learned from the results of the first test," Cottone said.
Doing those A/B tests determining best time of day, Schwartz said the team discovered between "10 [a.m.] to 1 [p.m.] was a pretty good time."
Narrowing those times down with a large group of people, he believed, wouldn't produce any significant results.
"People talk about running that test and getting good results, it's pretty shoddy data in terms of what hour is better than another. I think that's a bit overstated in terms of a useful thing, but that's sort of all we had so we got to there," Schwartz said.
With its vendor, BustedTees was able to set up that specific period of time that worked best for them, around six hours, wherein subscribers would receive their email. Once the optimum duration and send time was determined, the information was mapped against the promotional calendar.
Coordinating between Marketing and Sales
Schwartz explained a large part of his role in this campaign was coordinating the details of the campaign between Marketing and Sales. Sales and promotions for BustedTees only run for a limited time period, which made choosing a time tricky.
"Most of our sales are timed, let's say 24 or 48 hours, to create a sense of urgency. So, it was figuring out when, even though it's a personalized send time, we should be hitting someone so that they're getting the most amount of the sale possible," he said.
Schwartz said the team sat down and "started with the sun, where it's earliest, so Australia is ahead."
By organizing the sends that way, they were able to accommodate sales so "those people in Australia will get it the same day, personalized, as the people in the United States, more or less, so that the sale will always be relevant by day," he said.
Step #3. Implement a personalized send time
It wasn't until the team leveraged their vendor's ability to analyze their subscriber's open history that they implemented a personalized send time.
"We were actually able to achieve our personalization goal, as we're now able to send an email to each subscriber at the exact time they're most likely to open, which is the first phase of our email conversion funnel," Schwartz said.
They may have reached their objective of implementing the personalized send time, but their goal for metrics turned out to be off the mark.
"The original goal with personalized send time was definitely open rate," Schwartz said. However, results only showed a minimal increase — usually under a 5%. The results they did achieve were certainly a surprise, according to Schwartz, but a welcome one.
"Everything else was a constant so we weren't expecting to see a change in clickthrough rate or change in conversion rate from clicks, but we were trying to move the needle on the top of the funnel … we sort of expected the other KPIs to be a constant," Schwartz said.
For this campaign, the data collected by BustedTees' vendor analyzed a subscriber's previous open history, as well as the sends to the subscriber from the previous step in this case study, determining the designated time period.
Digital marketers have access to "an endless supply of actionable data which presents an opportunity for continuous improvement both within marketing and throughout all departments of an organization," Schwartz said about the team's ability to implement and build upon this campaign.
This personalization allowed his marketing team to increase the value of their already valuable email marketing channel by "allowing us to target individual subscribers at the specific time that they're most likely to engage with BustedTees," he said.
Dangers of personalizing too highly
The problem with personalization, Schwartz said, is most of the time, "you don't have enough data to make an educated personalized decision."
A change like personalized send time, or gender-oriented emails, is small enough to make a difference and still be mostly accurate.
"If you've been on our email list for a while and you've opened them a number of times, we can get what I would consider to be pretty solid data about when you like to open your emails," he said.
Deeper than that, he said, "I think it becomes really hard to have good data on, let's say, should we be sending you T-shirts that are related to TV shows like 'Breaking Bad' as opposed to TV shows like 'Futurama.'"
That would be difficult for them to do, he said, but with personalized send time, "you're able to get aggregated enough data to make a good decision."
RESULTS
Implementing a personal send time has allowed Schwartz and his team to make operational improvements "as everyone on the team now understands the importance of meeting deliverable deadlines to ensure on-time email delivery," he said.
The results they achieved from this personalization campaign were:
- 8% lift in email revenue overnight from personalized send time
- 17% increase in total email response rate
- 11% higher clickthrough rate
- 7.6% increase in post-click site engagement
Cottone is encouraged by the success they have seen so far, and said, "our work is far from over."
"Now we've figured out how to bring more subscribers to the site through individualized email delivery, we have the opportunity to apply personalization learnings to each and every phase of the conversion funnel ranging from email design, to on-site experience checkout, and post-purchase communications," he added.
Sources
BustedTees Sailthru — BustedTees' Personalized Send Time Vendor
Related Resources
MarketingSherpa Email Summit 2013 Wrap-up: Top 5 takeaways for email marketersEmail Marketing Segmentation: Clothing brand uses social behavioral data to drive a 141% increase in revenue Email Marketing: What I’ve learned from writing almost 1,000 emails for MarketingSherpaEmail Personalization: 750% higher CTR and more revenue for e-commerce siteEmail Marketing: Reactivation campaign for performing arts center sees 738% ROIEmail Marketing: Weekly newsletter sees 400% lift in reactivation of inactive subscribers with personalization
