Palmer Web Marketing Blog - eCommerce Blog PWM Products Justin Palmer Contact Justin Palmer at Palmer Web Marketing Free Internet Business e-books Palmer Web Marketing & E-Commerce Blog

Why Your Email Marketing Can’t Survive Without A/B Testing

by Justin Palmer - September 14th, 2007

As marketers, we all like to think we know all the answers for our clients. Unfortunately we don’t, which is why we need testing. Especially in regards to email marketing, testing campaigns can be the difference between a moderately successful program and a breakout success. Let me share some variables I’ve tested with my clients in order to determine the best paths to success.

Text vs. Image: Should emails be text or image heavy? In my experience with one particular client, a text heavy email received a slightly better delivery rate than an image heavy email. However, both received an almost equal amount of clicks and conversions.

Subject Lines: Should they be short or long? Is ALL CAPS ok? What about using words like “Free?” In my experience, the subject line doesn’t matter as much as it used to since ISPs seems to be leaning towards reputation based spam filtering. While I still avoid using all caps, I have been able to use the word free and acheive good delivery rates.

Calls to Action: What works better, “click here” or “learn more?” Depending on your type of business, and what point of the buying process your customer is at, the answer to this will vary greatly. The best you can do is try different CTA’s and measure the results.

Time Sent: Every email list has a sweet spot in the week when subscribers are most apt to open and engage your email. The only way you can find this sweet spot is to test different timeframes. Traditionally, Tuesday seems to be the favorite day to send emails.

There are an endless number of variables you can test. While many of the lower cost email service provider do not provide a built in a/b test functionality, I’ve found that you can simply split your list in two and create two different versions of an email. As long as your provider offers basic analytics, you can then track open rates, clickthroughs, and conversions.

In my experience, you can’t trust the delivery rates shown by your ESP’s report admin. Recently I tested an email delivery monitoring service by Delivery Monitor for one of my clients.  The service allows you to see which ISP’s (most of the major one’s except gmail) are sending your emails to junk or not delivering them at all. I’ve this service is extremely helpful with a/b testing content.
Happy testing…


Related Posts:

One Response to “Why Your Email Marketing Can’t Survive Without A/B Testing” by Justin Palmer

  1. Nakul Mehra Says:

    Permission Based Email Marketing has a new meaning – Not only does the recipient want to receive your email only if he opted for it but only he/she wants it in a particular pattern. The term “User Friendly-ness” is often used for web sites but even commercial email content needs to be:
    1. Visually Appealing
    2. Precise & Crisp
    3. Easy To Find Contact Information
    and most importantly…
    4. The purpose of the email content should get absorbed by the reader without him trying to struggle with it.

    A/B testing is available with a few providers and can be explored in unlimited combinations to get good results.


Blog Updates
RSS Twitter