How Wide Should You Make Your Email Newsletter? (Test Results)

by Derek Halpern · 46 comments

A couple of weeks ago, I showed you how simple text emails are more preferred than fancy html emails.

Now, the question is, how wide should you make your simple text newsletter?

Hardcore internet marketers often send a thin email, often no wider than 300 pixels, or 45-50 characters.

I always believed that was too thin, and on Social Triggers, I’ve often sent emails that were 380 pixels wide, or around 65 characters.

However, here on DIYthemes, I decided to run some tests, and here are the preliminary results:

First, Here Was The Setup

On AWeber, my personal favorite email provider, they make it extremely easy to run a split-test broadcast.

After clicking “create split test broadcast,” you’ll see a screen just like this:

How to Split Test Emails in AWeber

As you can see, I just simply added 25% to each section. Then, when you press create split test, AWeber automatically generates 4 separate emails, each of which goes to 1/4 of a randomly selected portion of the email list.

Cool, right?

Now What Were the Email Split Test Variables?

As I said, I created 4 separate emails, and here’s how I broke each one down.

  • Email 1: This email was limited to 380 pixels
  • Email 2: This email was limited to 450 pixels
  • Email 3: This email had no width, meaning the text would expand to the browser window.

And then, for an added test, this last email was a different type of email.

  • Email 4: This email had no width, but the font was larger than then the other 3 emails.

So, for the purpose of this test, Email 1, Email 2, and Email 3 were competing with each other.

And then, Email 3 and Email 4 were also competing.

Now let’s look at the results.

If you’re wondering how I limit an email to a certain width, I simply add a table into the broadcast section with a specific width. In AWeber, it’s as easy as pressing a button

Preliminary Results: How Wide Should You Make Your Email Newsletter?

So, here’s what happened:

Email Split Test Results

Now, as I said, these are preliminary email split test results.

But as you can see, it appears that no width with a larger font is the best solution.

Let’s break it down…

The first 3 emails were using the same font, and in that scenario, 380 pixels was the best for clicks, by a large margin.

To put that in perspective, on 10,000 emails, Email 1 would generate 380 clicks, whereas Email 2 would only generate 320 clicks, which means the results of Email 1 are ~18% better than email 2.

However, when you look at these test results, you can get the same number of clicks with less unsubscribes by simply using no width + larger font.

Now, I’m not one to complain about unsubscribes. If people are going to unsubscribe, they’re going to unsubscribe. That’s just the way it is.

But that doesn’t mean you shouldn’t seek to minimize them where possible, which is why I like Email 4.

Further Testing Needed

The above test was based on one email broadcast.

And while the sample size was more than adequate, I won’t be happy until I replicate these same results over the course of several different broadcasts.

So, you can bet I’ll be doing that over the coming weeks.

Now, if you want more on email marketing, here are some articles you should check out:

1. Why Email Marketing Crushes Social Media Marketing (and How to Get Started)

2. Here’s How You Can Build Your Email List Fast (Three Articles on Effective Email Marketing)

3. How The Feature Box Increased Our Email Subscription Rate by 51.7%

And now, what do you think of these test results?

How wide is your email newsletter?

About the Author: Derek Halpern ran marketing at DIYthemes, and is the founder of Social Triggers. To get more tips on how to be confident, sign up to his list here.

If you enjoyed this article, enter your email below to get free updates!

{ 46 comments }

Josh Sarz

No fluff indeed. Straight, cold, hard data.

Josh Sarz

Personally, I prefer the no-width, larger font template. Looks more email-like.

Jay Venka

I think its Jason Fladlient and John Rhodes who have their emails always formatted in short widths, usually in 10 words. I like the way it is presented, its like reading a newspaper article and i love that..

Thanks for split test im gonna use this for my campaigns..

Jon McCulloch

Interesting as they are, these stats aren’t as important as the sales you made. Open rates and click-throughs are good for the ego, but I’d rather have sales.

You could argue the more click-throughs you get, the more sales you make, but that’s a non-sequitur, since there may be subtle influences in the email format which negatively affect buying rates. Dunno. Test, test, and then test some more, I guess.

Warmly,

Jon

Derek Halpern

Obviously ;-p.

And we’ll be testing everything like that.

Rob @ Atlanta Real Estate

Super interesting. I was just about to ask how you limit the width of a simple text email, but you anticipated that one nicely.

Looks like there’s ~0.6% up for grabs here. Makes a real difference on huge mail outs.

Alexa

What size fonts did you use?

Kathleen Richwell

My eyeballs say it is 14. Derek will let us know if I am right.

Brenda Griffin

Derek – Thanks so much for sharing your results. I certainly had width on my radar, but not font size. Just reading the email with the link to this post paints a great picture. Again, thanks for sharing – and have a great day :)

jared

Again, great actionable stuff backed up by data. Thanks.

Drew

Derek – I may be reading it wrong but the descriptions of emails 1, 2, 3, and 4 don’t seem to match the preliminary results chart.

Derek Halpern

I don’t know what you’re talking about. :-)

(thanks. I fixed it)

Jeff

Great article. We were just having a go-around about this as one of our clients has a list that skews young demographically, and looks to be receiving (or at least opening) a great deal of their on their mobile devices. Your method gives us a running start at saner split testing setup. Thanks!

Annie Sisk (Pajama Productivity)

Very interesting – what I’ve been doing instinctively proves the better option via your data. I think that’s the second time that’s happened — the prior results on “plain-ish text” vs. heavily designed HTML favored the former, which I’ve also been using consistently for PJP.

AC

I’ve always preferred plain-text emails because they feel more like a “regular” conversational email between acquaintances. Over-designed newsletters feel too much like, well, a newsletter.

Denise Hamlin

Interesting. I love stats. I’m visually challenged, so large font is always good for me. I guess I’m a little surprised with the result of this test. Apparently it’s not just us older folks that like the larger font.

Homegrownart

Great Derek,

This is helpful to know. Looking forward to read what conclusions you make after future testings.

Cheers!

Ruan

Still building the list here on my side Derek, so I guess I’ll be in time for when your final results arrive ;)

Best of success with those!

Mark Hare

Good timing, Derek, from my point of view. Just about to kick off a campaign in the Down’s syndrome niche!

Did you split test your ‘What’s Your Reason Why?’ email? I received it in ‘No Width’ with larger font format. Was that responsive? It was a long email, with lots of line breaks—double returns—between paragraphs of just one or two sentences. Or do you find shorter emails more responsive? Obviously that’s another variable, hence you saying you want to test further.

Also, are you aware of any benefits if your simple text email font matches the font of your blog/website, for brand continuity?

Daniel M. Clark

I’m not at all surprised, and I’ll give you my theory why. Having that thin column of text looks like what you said – the work of a hardcore marketer. Average people, in my experience, *hate* hardcore marketers. They don’t mind a little, but hardcore isn’t looked upon favorably. So the regularly formatted emails performed better. *But* because of the type of list you have, a lot of your subscribers are hardcore marketers themselves, so they naturally react more favorably to the thin column. That’s why you get 3.8% on the thin column and 3.2/3.3% (no significant difference, IMO) on the wider column and no width.

The portion of your list made up of marketers reacted well to the emails formatted the way they format emails, and the portion made up of non-marketers (or “light” marketers for lack of a better description) reacted well to the regularly formatted emails.

Just a theory. I could be wrong.

Randy Cantrell

I agree with Daniel. The IM crowd seems to follow the same thin column email format. I feel I could easily substitute one email from one IM guru with another and I’d be unable to discern the author. I prefer to read emails the way “traditional” email looks and feels.

Kellie Brooks

I wonder if people clicked through more often with 380 pixels because it was less “read-able?” What’s your take on the why behind it?

Bob

What size was the “larger font” in email 4?

Laura Upcott

Thanks so much for sharing this Derek. I’m working on an email marketing campaign right now. I’m going to try “no-width larger font” to see how it works.

Terry McNeal

Good stuff! I’m old school and like the larger font. I also prefer no width or a larger width (not thin).

I would also like to campaign that we go back to 2 spaces after the period rather than 1. We should not have to make our readers struggle when they want to connect with us. White space is our friend and a friendly way to do business. Just putting it out there.

GaryJ

The only reason two-spaces ever came about was due to the use of monospace fonts. Unless you’re making a point of sending your fonts using a monospace font, stick to single spaces.

Ian

@Daniel M. Clark I think you hit it right on the head. I can tell when I’m being “sold” to in an email. More specifically, when I see that narrow plain text I am almost sure there’s a link coming for an IM or related product.

On another note, Derek, thanks so much for this and the last email on plain text. I’m re-shaping a new campaign I’m launching now.

Looking forward to what other split tests you do, because as @Jon McCulloch said above, sales is the bottom line. So depending on what the subject/niche of your email test was could obviously skew the results.

Ahmad Wali

I am with the old format of the email. I didn’t like the plain text as I never read complete text now. Well! Majority is the Authority…

Jacques

Nope, what you do is isolate a segment of people with your behaviour and then test again.

Segmenting your list is something every competent email marketer should be doing.

Ivete

Good info, thank you! I’m curious whether the Complaint number changed at all between the tests? I hate complaints way more than unsubscribes =)

Jesse Petersen

Absolutely need to do this over n=4+ broadcasts. This is the first one in 5-6 that I clicked because I’ve been so busy, but I took time to look at your data because I love hard data, which is not the subject of every broadcast. I can assure you that the e-mail’s formatting had nothing to do with my click – it was the subject (not subject line).

Richard

Hi. I’d like to see you test the 300 width since that is what the internet marketers do.

The unsubscribe rates seems non-substantial relative to this test.

THanks for the test.

Neville

What if you combined the 380 width + larger font and call that email 5, wonder if it beats email 1 and 4….

Marion Ryan

Interesting initial results, Derek. I assume the email you sent announcing this info, title “How wide should you make your email newsletter? (test results)” was what you call ‘no width, larger font, or maybe this too was a split test? It looked odd ie unusual but definitely came across as more the sort of email I’d receive from a friend, so I felt more obliged to read it.

I saw your earlier email about plain text but I am resisting your findings a bit. A lot of my website clients are female service providers (coaches, consultants, alternative practitioners) and their target market seems to prefer the pretty stuff.

Love this sort of experiment, thanks for testing.

Chris

Thanks for this. It would be helpful to hear a rough count of sample sizes for each segment, and also to see screenshots of the different variations.

Also, today is Jan 24, 2012 in case someone stumbles across this article in the future and wonders if this technical advice is from 5 years ago or last week. Articles about web best practices have a short shelf-life and should always have timestamps IMO (that goes for comment sections too).

Daniel M. Clark

I agree. I strongly dislike and disagree with the new movement (and it’s not just here) to remove timestamps from everything in an attempt to trick Google into thinking your content (or comments) are evergreen. Users come first, and if your users don’t know when something was written, they will distrust the advice or content – as well they should.

Tim

You’ve changed multiple control variables (testing for different things). Doesn’t this kind of render the test irrelevant insofar as the results go?

I mean, oughtn’t you have tested email widths ONLY? Introducing the font size variable is a separate test in itself.

Jacques

He ran two tests.

Emails 1-3 were testing for width only.

Email 3 and 4 tested font size.

You can run both of the same tests at the same time as you’ve always got a control for both, which in this case is email 3.

Ana Hoffman

I got email #4 and hated it.

The only reason I clicked through was because it was from you.

Shajjad

Another great post and tips . Thanks Derek.

But i don’t like the font of diythemes newsletter :(. That is very big for me. Don’t know what the expression of other member.

Shajjad

Sreejesh @techgyo

What was the font in mail #4, I think that was less readable.

Stephen

What’s the test sample size?

Eg, this difference may be statistically meaningless unless the sample size is quite large.

thank you,
Stephen

Derek Halpern

Each email was sent to more than 4,000 people.

Seppo

Thanks for sharing some real data. I’m not a statistician (and have blissfully forgotten the courses I took in the Uni), but the differences seem so small that this might just be due to random variation. Did you check if this is statistically meaningful result?

Kat M Bee

Hey guys,

I unsubscribed to this newsletter. Not because I didn’t like the layout. I actually MUCH prefer it to the narrow layout in your SocialTriggers newsletter.

I unsubscribed because I’m receiving the same information from DIYThemes, SocialTriggers, and The Thesis Statement. :)

Derek Halpern

What do you mean?

We never run the same articles on any site.