DM News' Essential Guide to E-Mail Marketing: How Triggered Messaging, Testing Can Boost Conversion

Share this article:
Mortgage applications are notorious for being long and complicated. E-Loan has worked to change this perception by developing a simple online application to guide borrowers through the process.


Even so, many borrowers abandon the application before completing it. E-Loan's challenge was to create a cost-effective process to re-engage these customers and get them to finish their application. Because borrowers need to create an account before they apply, we have captured useful data (name, e-mail address and telephone), even for customers who later abandon.


Outbound telemarketing turned out not to be cost effective to generate completed applications, so we tried e-mail. The strategy was to contact customers in two waves: shortly after they abandoned and one week later.


With help from our e-mail services provider, Responsys, we created an e-mail campaign that waits 30 minutes after abandonment, then automatically delivers a soft reminder to complete the process, with two ways for customers to finish, online or by phone. The follow-up e-mail gives more information about our value proposition and again encourages the customer to complete the application.


We used an A/B/control testing platform at Responsys that assigned customers to one of three segments. The A/B segments got one of two versions of the e-mails (30 minute and one week) while the control segment was suppressed from receiving either version.


We used a 4 percent control group, with the idea that we would maximize the number of customers who got the messaging while building enough of a sample size over time. After several months, we had enough control samples to properly read the incremental value of the campaign.


Results have been encouraging, with the champion version outperforming control by 28 percent. Our A/B testing so far has been limited to subject lines, as this is the top of the funnel (opens) and has the greatest effect on our overall conversion rates.


We soon will test copy, creative and promotional offers as the next level in the funnel (clicks). Our final phase of testing will look at different landing pages (conversion). It makes no sense to test creative or landing pages until you have optimized the number of people who open your message.


Key learnings:


Automation is critical. We used to send this type of campaign once a week on a manual basis. It wasn't scalable and lacked the immediacy of the 30-minute trigger, which operates 24/7.


Build intellectual capital. The automated programs don't rely on one employee who might jump ship along with his accumulated knowledge. Don't be left high and dry when a skilled employee leaves.


Continual testing. Our testing platform lets us continually test and learn with very little incremental investment.


Not everything is incremental. Measuring campaigns against a holdout cell is critical to understanding whether the campaign has value. This is especially important for multichannel marketers who generate awareness and interest through channels other than e-mail.


This material may not be published, broadcast, rewritten or redistributed in any form without prior authorization. Your use of this website constitutes acceptance of Haymarket Media's Privacy Policy and Terms & Conditions