A/B Testing Your Way to Better Titles

Which would you rather drink? Simple A/B testing.

The online world is filled with content — some of it very good. Amidst the glut, having the right title can make the difference between finding an audience and languishing in oblivion.

For me, the right title is the hardest part. And for something relatively permanent, like a book, the stakes are high.

What Makes An Effective Title?

First, the headline or title must meet physical or length requirements for the medium.

There’s an optimal length for a Facebook or LinkedIn post headline, for example. Email subject lines are truncated in many mail clients. Book titles should fit on the cover and be legible in a thumbnail.

Beyond those guidelines, you can find numerous articles offering suggestions about crafting titles. Use certain words, but avoid others. Ask a question. Say something surprising.

Alas, there is no single correct answer for writing the killer headline. The best measure of a title is how it works for your audience in the context of the work.

An effective title is the perfect alignment of two variables: the ideal reader (your target market) and the content (the product).

As the writer, you understand the content— the second variable. You already know what the book or post says, and may be tempted to write a title that sums everything up.

Your reader doesn’t share your knowledge, not yet. Authors suffer from the Curse of Knowledge — the cognitive bias that makes it difficult to understand the perspective of someone who does not share your knowledge. Although you understand the product, your insight into the reader is limited.

To get a better understanding of how your titles work, you need outside perspective. That’s what testing is for.

The ABCs of A/B Testing

As the simplest form of polling, A/B testing is rather like a basketball tournament. (Can you tell I wrote this in March?)

  1. Identify a small set of candidates for testing. Keep the group fairly small if you don’t want testing to last as long as the American Presidential elections.
  2. Choose two options to compare. If you’re testing a landing page, compare versions that vary in only one dimension (button location, image) so you can discover what’s important to your audience. For a book cover, change the words but not the layout, or the image but not the words.
  3. Ask a clear question: Which do you prefer, or which would you click on?
  4. Keep testing the winners until you identify a good candidate or two.

The more people respond, the more faith you can have in the results.

Unlike a basketball tournament, you don’t have to assign third and fourth places for book titles. If two are pretty close, go with the one you like best — you’re the author.

A/B testing capabilities are built into many tools you may already use.

  • Many blogging platforms enable A/B testing for different versions of the same page.
  • Email marketing software typically lets you test subject lines.

You can make informal assessments without testing frameworks. Send an email to your colleagues or a mailing list. (This works best if those people match your target audience.)

Or, experiment with different titles on tweets pointing to a post. Make sure to run those tweets at the same general time of day. You can also use simple polling on Twitter to test variations.

But the problem with these simple approaches is that you only see the first response to the title — the click — rather than what happens after.

Testing Beyond the Click

You need more data to determine if the title attracts the right people — the ones who will find value in what you’ve written or will read and enjoy your book.

A great title focused on the wrong audience is merely click-bait, from the reader’s perspective.

If you have the time and budget, hire someone to research for you. But there are other, less costly options.

I’m in the process of finalizing a title/subtitle for my next book, which applies basic principles of cognitive science to optimize the writing process.

To examine different titles, I’ve used a service called PickFu that crowdsources A/B testing polls, delivering prompt and detailed responses for a small fee. You can choose how many votes you want, as well as basic demographics for the respondents.

PickFu reports the overall results quickly, depending on how narrowly you have focused the demographics.

Beyond reporting the numbers, each respondent leaves a short reason for their selections. That feedback helped me understand why and how people reacted to each title, and to find those people who sounded the most like my ideal reader.

Here are a few of the comments:

“I would be curious to find out how the brain and writing skills are related”

“I feel like most books have a title and subtitle, so that’s why I gravitated towards choice B.”

“The writing process beginning provides a much clearer context for what the book encompasses, which is what you must ultimately convey”

“I like the idea of preparing my brain to write…”

If you’re interested in following the progress of the book, sign up for my writing practices email list.

[Image: Jan Vasek on StockSnapio.com]

Originally published at annejanzer.com on March 10, 2016.

Author, Writing Coach: Writing about Marketing, Technology, and Writing Itself (very meta). AnneJanzer.com

Author, Writing Coach: Writing about Marketing, Technology, and Writing Itself (very meta). AnneJanzer.com