-
Manage your services:
By Luc Vezina, Director of Product Management - GOT Corporation
Email marketers often shy away from using techniques like A/B testing because of the misconception that such tests are "scientific" and therefore complex and only accessible to marketing organizations with huge budgets. Indeed there is some "science" behind the technique and large corporations like Dell have put A/B testing to great use on their retail website, but there is no reason why any marketer cannot enjoy the same great benefits.
A/B tests are an easy and highly effective way for all marketers to increase their open and click-through rates above a pre-defined benchmark. What is required is that the marketer be willing and able to devote time, to be patient and to have a keen eye for detail-and what marketer among us isn't willing and able when the result is improved performance on key email campaign metrics?
A Science and an Art
A/B split testing with email marketing is as much a science as an art - requiring patience and a scientific touch in order to optimize the message mixture gradually and effectively over time. Effective testing requires diligent analysis, modification, experimentation and re-analysis to maximize the efficiency of the message over time. Marketers expecting to send two versions of a campaign to two subsets of their list - and see a marked difference in the results immediately - will be disappointed. Why? Because achieving significant improvements will require several small iterations over time in order to ensure that changes in results can be directly attributed to a particular change in the campaign.
While there are the obvious and easily changeable variables apparent in an email marketing campaign such as subject lines and "from" fields, the numbers of variables that can be changed in an email message are endless. A/B split tests primarily fall under two main categories - changing the offer and changing the creative.
- Offer changes often include the messaging related to the offer itself: 15% off vs free shipping, subject lines, copy/content, and hard vs soft sells.
- Creative changes reflect the 'packaging' in which the offer is sent, including: color choice, images and image placement, and overall mailer design.
Here are some great steps to follow when implementing your own A/B test
1. Establish a Benchmark campaign
Testing is a waste of time without anything to compare future results against. Establish your benchmark with your best-performing campaign, and tag this as your control message. Measure the success of this control message over a set period of time, and use that same time period for all subsequent incremental changes.
2. Split your lists.
Split your list into 3 random sets - test list A, test list B, and a third list making up the remainder. Randomizing your lists is crucial to getting the best exposure of your testing campaigns to your entire audience, and not a specific type of subscriber, reader or buyer. Once split, your test lists should remain as the test lists - if possible they should not be modified until the A/B test is completed.
4. List and record the variables that you want to change.
Define the changes you want to make to your campaigns. Don't limit yourself to simple adjustments like subject lines, from fields or blue vs. red buttons (although always include these basics). Imagine any and every conceivable change or modification you can make. Keep your variables specific - 'change design' vs. defining what changing your design entails - and record each change as another variable.
5. Set your goals, but remember to be patient.
If you are implementing an A/B test to increase click-through or open rates, set realistic goals for your metrics. Click-through rates will never increase by 90%, no matter how effective your campaigns are, but an increase in 5% over a defined period (always depending on your knowledge of your audience and their behavior) is practical. Your goals however are set by you - therefore can be as ambitious or pragmatic as you like.
6. Make incremental, gradual progress... and measure.
The cornerstone of successful A/B split testing. Always make your changes incrementally, one at a time. Changing multiple variables within the same campaign will not pinpoint the precise change resulting in improved performance. Instead of moving your logo to the right and moving copy to the top of the campaign during one send, space these changes out over two separate campaigns. Although painstaking and time consuming, proper A/B testing requires that each and every variable be isolated and its results recorded. Although it's not a quick process, the results speak for themselves, and all changes, however small, can actually have a dramatic effect in your overall click-through and/or conversion rates.
About GOT Corporation
GOT Corporation™ offers Customer Relationship Marketing technology that allows organizations to move beyond traditional product-centric marketing to create relevant communications with customers, resulting in long term, profitable relationships.
About the Author:
Luc Vezina is Director of Product Management at GOT Corp. He monitors industry trends closely and has been instrumental in helping shape email marketing best practices for thousands of GOT's clients. GOT is a founding member of the Email Service Provider Coalition.
Find Out More About Got's Products