Email Marketing Testing and Optimization

1. Render Testing
Typically, the first tests occur upon completion of the html coding, render testing in various email clients to ensure it is consistent in appearance for all your recipients. Many email service providers include this testing feature with their product, if your tool does not, you might want to try the services below:
- Return Path
- Litmus

2. Link Testing
The second testing stage most marketers will go through before a deployment is link testing. Have you ever sent an email to someone forgetting to add the attachement? When you are sending to your list you 'll want to avoid having to send an apology. Make sure you click every link in your email, including function links such as the unsubscribe, the hosted version or the preference center link. An unsubscribe link that does not work is like not having one, which could mean your email no longer fits local regulations.

3. Split Testing
Finally, to make the most of every email you send it is necessary to test your template's performance, your subject line's performance, you might even want to consider testing a change in your from name and address... When you are split testing you are testing against your own list, which makes it highly relevant and accurate. If you could have a 5% raise in conversions by having a red call to action button instead of a green one, wouldn't you like to know?

This article will concentrate on this last form of testing, which is just as essential when optimising your emails for the best results.

Email split testing allows you to tailor your communications based on your subscribers' response. It enables marketers to discover key information about a given company's list to ensure the emails deployed are as effective as possible. Let's face it, the initial design of an email lies with the creative designer and the feedback he is given; we've all asked for a button to be a little rounder, or a title slightly larger.

At this stage the email is the result of your personal tastes, its absolutely impossible for any marketer, regardless of his or her experience to know how well it will perform.

Split testing allows you to find out information that your subscribers could not even have reported themselves. In other words, self reported behaviours are less reliable than observed ones, making split testing more effective than having every subscriber answer a questionnaire! What we found is that very often results are counter-intuitive, a main image on the left might feel more natural with your design but would you keep it there if having it on the left meant an increase in clickthroughs?

Effective email marketing testing done consistently on every send will result in a constant evolution and adaptation of your templates and emails. So how much difference can split testing have on your results, the exact answer is, just about as much as you are not the average person on your list.

Template Testing
Here there are no limits, anything you can think of can be tried and tested! The use of a navigation bar, and its format; the placement of images and their formatting; the color scheme and background variations, fonts and styling; text to image ratio; test and button links; cross and up-sell features' placement; calls to action buttons and function link placement and wording.

Subject Line Testing
Maybe you've observed great engagement when using "Great promotion!" at the beginning of the subject line. Challenge it! Try something else! Email marketing is a way into your subscribers' inboxes, where they receive emails from friend and family, treat it as a conversation. Try: "We've got a new promotion you might like!", try removing that introduction sentence altogether.

From Name and Address Testing
Playing around with your from name can be done quite easily with most tools, and can have a great impact on your results. From address testing is often discouraged but could be a solution for some. For example it can play a key role in ensuring the success of a re-activation campaign, which could have been easily ignored if it had been sent from the same address as the newsletter.

Send Time Testing
If you are sending a regular newsletter, it is likely you send on the same day every week or month. The beginning of the month and close to pay-day is a good time for many online retailers, some will find that the end of the week or even early in the week-end is the right time for them. Think about time of the month, week of the month, day of the week, time of the day, test it on your list and adapt.

Focus on where your emails do not perform well enough. Perhaps you are seeing a very high unsubscribe rate, perhaps its time to reassure your subscribers by adding a permission reminder, or making the preference center link more prominent than the unsubscribe. Perhaps your clickthrough rate is low and test your calls to action, people might feel more comfortable clicking a "More" button than a "Buy" one.

Partial A/B Spllit Testing
When a split test is done first on a portion of the list, say 20%, and the most successful version is sent to the remaining 80%. This allows you to ensure only 10% of the list will receive the less efficient mailing; however it requires having a large list so the result is statistically accurate.

Full A/B Spllit Testing
This is when you send your current mailing unchanged to 50% of your list that were randomly selected, and show the remaining 50% your new mailing. This solution is ideal with smaller lists, that would not have a statistically accurate representation of the full list with a test on 30%.

If your email marketing service provider does not facilitate split testing schedule the two sends yourself to ensure they are comparable. Ensure you are always happy with the results, if the impact is insignificant on a partial A/B split, consider doing a full test for your remaining send.

- Render, and link test your emails, don't send an apology
- Never assume, always test your assumptions against your list
- Prioritise your testing based on your weaknesses
- Use a statistically valid sample for A/B split tests
- Ensure your emails are constantly evolving through constant testing

Act like a child and question everything: Why this word? this link? this button? this image? why here? why this way? and never be satisfied with a "why not?".

And remember, if you need a little push to optimise your emails, ImpartialMail is always happy to advise on all matters relating to email marketing, not only will we point you in the right direction we will also give you the tools to continue yourself.

ImpartialMail is a limited company that was registered late in 2011, by Greg Fournier. The company specialises in advising small and medium sized businesses in the UK on email marketing.

The company was created on the basias of 2 simple principles:
- Empowering businesses to do it alone
- Helping businesses start email marketing

We are not just another email marketing agency, ImpartialMail strives to empower businesses over email marketing service providers through impartial expert advice and services.

ImpartialMail Ltd.
Tel: 08009788759
Email: info@impartialmail.com



Starting Email Marketing by Greg Fournier is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License.