Testing - How To Boost Your Response Rates

Testing can be an important part of your email marketing strategy. A simple tweak in your email campaign could squeeze in some additional clicks, opens or even added revenue!

By starting with simple rules you can be a testing genius.  An old direct marketing rule of thumb is the 40/40/20 rule. Simply put, 40% of your response is driven from your list, 40% from your offer and 20% from your creative.  With email it can be a bit different, since your list and your relationship with your recipients will dictate if they even see your offer due to the potential of them filtering you. For this reason we like to put more emphasis on the list so for email we'll go with the 60/20/20 rule. How do you put this to work? Here are some ideas for each part of this rule which can hopefully help you garner better results.

Develop a “Control” Email

What is a "control"? It's you putting a stake in the ground as far as what you think will work for your audience. You can’t really determine if any changes you make to your campaign in your test are having an effect unless you have a comparison baseline. That’s where the control message comes in; it’s a handy reference point that allows you to evaluate those "tweaks" you made in your campaign.

So take your best stab at developing your creative, copy, offer and messaging and always test against that control email. Once you’ve beaten your control, that is, garnered a better result, then the new email becomes the control.

Test Your List

A good way to test your list is to simply divide your mailing list into different segments, send each group a different email, and determine which one got the better response. This usually works best with larger lists (over 1000); smaller lists may have a sample size too small to be statistically valid. Either way it doesn't hurt to try.

You also might want to try separating your list by "customer" vs. "lead". This way you can try moving the leads along to one day be customers and you can have different offers go to your customers.

Test Different Variables Within Your Email

When you try a new idea, make sure it contrasts sharply with the wording, theme or style of your control – it’s the only way to really determine what works and what doesn’t. With so many different variables you can test, here are just a few:

Offer – Your offer is 20% of your response so let’s get it right. Once you have your list segments in order try some different offers to your list.

• Test a hard sell (Buy now for only $29.95) vs. soft sell (include content around your offer)
• Test pricing – 29.95 vs. 39.95
• Test varying discount levels – 20% vs. 30% off
• Test a free offer - a free product with purchase vs. no free product
• Test free shipping vs. discount

Look and Feel – You’ve built your control based on your brand and what you are trying to accomplish. Now you can learn by testing specific variables within your email as it relates to your general look.

• Colors and Fonts – Try not to stray TOO far from your branded image but if you’re looking to boost response test it out.
• Mail Format and Layout  - Test Text vs. HTML, long vs. short form or columns vs. vertical layout.
• Length of Message – Include the entire message in the email or just a small teaser with a link to the rest of the information.
• Degree of Personalization  - Test no personalization vs. personalized name.

Subject Line – Your subject line IS one of the most important parts of your email, if it doesn’t get opened you don’t get your desired results so go ahead and test within the subject line to get your email opened. If you can get access your list of those who did not open your email, you can send another one with a different subject line.

• Test your offer wording – 20% Discount and Complimentary shipping vs. Complimentary shipping and 20% discount, remember to stay away from excessively using the word “free”.
• Test different article headlines from your newsletter too see what gets opened more.

Test Your Mailing Days

The day of the week and even the time of day when an email is received can have a significant effect on the response. Does your audience consist largely of people with desk jobs who check email throughout the day, or do they check personal accounts in the evening hours?

The impact of timing will vary according to the target market for each company, so there’s no hard-and-fast rule. The only way you can determine what’s best for your audience is to supplement gut feel with trial-and-error.

Test Your Frequency

The volume of email you send, as well as the frequency with which you mail your list affects the response. Does regular mean twice a week or once every two weeks? It really depends on what you told your recipients when they registered, so if you do test, don’t stray too far from what they expect.

Also be aware of your frequency as it relates to bounces and unsubscribes. If you mail too sporadically, you’re likely to lose your recipient’s interest or they may have changed email addresses in the meantime and forget to inform you.

Got any insights on testing? Tell us!

Comments

[Darla Dixon]

Testing the list...that is brilliant! Michelle Cubas is right, every one of your blog entries has a golden nugget of info in it :) Thank you!

[Michelle Cubas]

Dear Janine,

One reason I open your email blog is that I "trust" there is a golden nugget in there somewhere. You've established your credibility with me by consistently delivering information I can use and share with my clients.

Thank you. It is refreshing.

Regards,
Michelle Cubas, Positive Potentials LLC
Enterprise Brand Strategist & Business Coach
Business Influences http://coachcubas.blogspot.com

[Dr.M.Grant]

I had use some different techniques, before I manage to work out Best working one. It’s start with invitation. Invitation is your first contact with the respondents and may be the only opportunity to persuade them to participate in your study. A catchy subject line may get people to open your email but what really gets them to read it is their name in the preview pane. It must contain all the important information in the first few lines. State who you are, why you’ve contacted them, the purpose of the study, what you will do with the results, and if an honorarium is on offer. Give people some sort of time frame to work within. Set an appropriate deadline. Whether conducting business-to-business or business-to-consumer research, assuring confidentiality is essential to increasing response rates. An honorarium is a way to thank respondents and show them that their time is appreciated. Send at least one reminder three days prior to the deadline, to those who haven’t yet participated in the study. I have pointed that in my Marketing Strategy dissertation, which you may obtain in here( http://www.coursework4you.co.uk/sprtmrk14.htm). Yet, your article give some performance ,if I may say. Would recommend to some of my partners to read .

[Janine]

Hi lrm - Not knowing your open rate or conversion rates I'll take a stab at this. For your open rate you'll probably want to get more than the usual statistically valid sample of 100 becuase I'm guessing you'll want them to then click through. So in this case let's define a "response" as an open since you want to test your subject line. If you know your open rate will be 20%, I'd suggest you mail roughly 5,000 to get 1000 opens. (1000/.20=5000). Now if you know you have a 3% click through rate you can compare this at the same time. Email 5000 at that rate and you'll get a statistically valid sample too (5000 x .03 = 150) since I'm assuming that what you really want them to do it click. So here, either way, you're covered.

[lrm]

Say I have a list of 50,000 names and want to test three subject lines. What percentage of the list should I send the test to for it to be statistically valid?

I'm positing that I will test the three subject lines on a small sample of the list (well three small samples) and then send the best performing one to the remainder of the list. So I want to keep the sample as small as possible to so as to be able to send the best performing subject line to the greatest number of addresses in my list.

Thanks.

[Janine]

Hi Andi - Great question. If you have two separate campaigns and someone appears on your list twice, then yes they would recieve it twice. One way to prevent this is to download each list to an excel file and sort by email address in Excel. Find the like email addresses on the list then suppress them from your mailing so they don't recieve the campaign for this test at all. Not the most optimal way but it works. Another way to prevent this is if the two lists are "alike" meaning both are customer or both are prospect you can re-upload one of them, then append the second list to the first. This way all duplicates are removed. Then download the list, split it in two in Excel then re-upload two lists for your test. Again, a little bit of work, but it might be an answer for you.

[Andi]

I have two lists set up on my account: one with about 5000 users and one with about 10000 users. If I send a campaign to each of these lists separately, to track each list separately, will users who happen to be included in both lists receive the email twice?

[Janine]

Hi Rich - Not yet, however, our engineers have been hard at work reworking our systems to be able to handle this. Believe me, it's something I'd like in our system and it's something the team here is striving for. Stay tuned...

[Rich Majorek]

I would love to be able to test creative, offer, subject lines, etc., but I don't think vertical response really offers me the necessary tools to do this without setting up multiple campaigns and a lot of manual segmentation of lists. Are there tools available that I'm missing?

[Janine]

Hi Alon

You can download your list into MS Excel and divide your list into a few parts. Then you can name them and upload them as separate lists.

Hope this helps!

[Alon Raskin]

How does vertical response facilitate in dividing my mailing list into various groups? For example, if I have 10,000 names in a mailing list, I may want to be able to compare different subjects lines. How does vertical response allow me to do this?