There’s no way to know if your content will be a hit or a flop until you publish and promote it.
Right?
Wrong.
You can use data to help you choose the right content idea to start working on.
You’re about to invest a significant about of time, energy and money into creating great content. Some of the pieces are smaller, so you’re not so worried if they don’t get much attention.
But what about the big content you have planned?
A single piece is a significant investment of time and money.
You NEED to do everything you can to get a good ROI.
This is the first article in a series that shows how to know if your content idea is a winner (or a loser) *Before* you hit publish.
We’re going to use Facebook ads to collect data and predict which content will perform best, and which blog topics and ideas you should invest your time in.
When you’re testing, there are some key issues to be aware of and pitfalls to avoid.
I call it the 2x2x2 test.
Get the screencast here.
What Content Should You Create? The 2x2x2 Test:
This test is a framework and tool for a/b testing your content ideas, and getting insights into what your audience wants to hear about.
Why 2x2x2?
Because you’re going to split test 2 ideas, using 2 ads, and hopefully get 2x the success.
What the Test Does for Content Creators:
- Gives you data to decide between topics to write about.
Works best for top of the funnel content. - Helps you determine the popularity of an idea / topic before you write.
- Helps de-risk creating “big” content.
What the 2x2x2 Test Doesn’t Do:
- Tell you what topic to write about.
- Guarantee a hit piece of content.
- Replace your insight and judgement to decide what to write about.
Step #1: Pick 2 Content Article Ideas to Test
(The Test Helps You Choose Between Topics and Ideas.)
You still need to put in the work to come up with quality ideas to test.
Narrow it down to two topics. The test will help you decide which article to write.
If you’re having trouble generating content ideas, check out:
- https://blog.kissmetrics.com/101-content-ideas/
Most idea-generating articles spout the same stuff. This one from Kissmetrics, however, has a lot of high-level ideas.
Or you can try generating a data-driven story – something I’m particularly interested in these days. Because when you do it right, an information-based story spreads organically after you give it a nudge with outreach. And it’s just plain fun and interesting to write about.
Resources for creating data driven content:
- https://builtvisible.com/ideas-that-work-content-marketing-data/
- Read anything from the Priceonomics data studio
- https://www.sitepoint.com/generate-content-ideas-data-journalism/
- https://www.thinkwithgoogle.com/articles/tell-meaningful-stories-with-data.html
- http://buzzsumo.com/blog/how-to-write-data-driven-stories-5-core-narratives/
For Example Here Are some Content Ideas I’m testing:
- How to Create Data-Driven Stories That Spread VS. Epic Content Marketing for B2B Businesses
- The United States of Nuts VS. The Almond Milk Revolution
- Ranking Content marketers vs Ranking Colleges’ Content Marketing
Step #2: Create a Test Content Blog
You don’t want to put crap in front of your customers and readers. And you don’t want to ruin your SEO.
So, buy another domain and host your test blog there.
It’s better not to use a subdomain on your main site.
Make it quick, easy and dirty.
Your new blog shouldn’t win any awards. The design doesn’t matter. The domain name doesn’t matter.
To keep it SUPER simple – throw up a WordPress site and keep the default theme.
Keep Google away.
Best practice here is to tell the search engines NOT to index your test blog.
Because you might publish the first paragraph, you don’t want Google to first see it on my test site rather than my main site.
More importantly: You’ll also use the test blog to split testing headlines. For that You’ll usually publish the full-length finished article on the blog. And you definitely don’t want Google to find it on the test blog first.
Step #3: Create 2 Blog Posts – 1 for Each Topic
The goal here is to quickly create your articles. Because you don’t want to waste time on a blog topic that won’t perform well.
Side note: You might be wondering – why take the time to create a post because we’re just using ads to tests?
Because the ad may not be approved by Facebook if you just used a headline and there wasn’t relevant text on the page.
So here’s how to do it:
How to Create Your Articles Without Writing a Single Word:
You’re going to use public domain text and free/public domain images to create your blog post.
When choosing public domain text, you’ll want to make sure that the text you choose is related to your topic.
Where can you find public domain text?
- http://ebookfriendly.com/free-public-domain-books-sources/
- https://search.creativecommons.org/
- Select the usage rights and search.
- Google “uncopyrighted” blogs.
You do have to be careful to make sure the works REALLY are public domain. You can get in trouble if they aren’t.
Most blogs that apply a creative commons license have it listed in the footer of their site.
You’ll likely have to link to the article and attribute it to the author. That’s fine for our purpose.
Alternative to public domain content:
Because public domain can be tricky, instead you can create a small excerpt of text and use that for your article. Just the first paragraph and a quick one-to-two-sentence conclusion is fine.
I’ll typically whip up a paragraph rather than search for something in public domain. But either method works for the test.
Where can you find images?
My one stop: http://allthefreestock.com/
It’s an aggregator of free stock image sites. Some images do require attribution. Most don’t.
Publish both articles
Once you’ve selected your text and free image, hit publish.
Step #4: Create Facebook Ads
The goal with the ads is to see which blog idea gets more clicks and “wins” the experiment.
All it takes is $10-20 .
Here’s what you do:
- Create a new campaign, with the title: blog [name] a/b tests.
- Create a clicks-to-website ad group.
- Focus your targeting.
- I like to use an audience size of 50,000 – 100,000.
- Your targeting DOES matter, so do your best to reach your target audience.
- Use the same targeting for both ad groups.
- Create one ad for each post idea.
- Set the daily budget to $10 or20.
- Have Facebook optimize the ad delivery for you.
- Select a single image ad. Use your blog image for the Facebook ad.
- Use your post title as the headline.
- Write a short description.
- Let ‘er rip.
There Are 2 Ways to Run the Experiment.
Option #1:
Test one idea vs. one idea – and create one ad for each idea.
Option #2:
Create two identical ads per idea (with the exception of the title).
That way, you end up with a total of four ads.
Why? Because titles matter.
Sometimes your title can be the difference between a stellar click-through rate and a piss poor rate.
To make sure your title doesn’t kill the experiment, create two ad groups; one ad group per idea that you’re going to test. Then make two identical ads per idea, but with an alternate title for the second ad. Everything else stays the same.
The easiest way to do this is to create one ad, duplicate it and then change the title for the second ad.
To be clear, you’ll end up with:
- Idea 1 ad group
- Ad 1 (with original title)
- Ad 2 (alternate title)
- Idea 2 ad group
- Ad 1 (with original title)
- Ad 2 (alternate title)
Spend $10 per ad group. See which ad group gets the most clicks.
And you have a built-in headline split test.
Check which ad got the most clicks out of the two.
Content Idea Test Examples
Example ad split test #1:
Can you guess which of these two ads got the most clicks?
Example A:
Example B:
Example ad split test #2:
Example A:
Example B:
I tell you which ad won below…
Step #5: Quick Analysis of the Results
For $10 you should be getting around 30 clicks.
A typical cost per click for the United States should be around 30 cents.
But it’s highly dependent on your audience targeting and the quality of your ad.
For $10 you should be getting around 30 clicks. If it’s significantly fewer, then it’s possible your ads didn’t work, which means your audience didn’t like the blog topic.
What’s a good click-through rate (CTR)?
According to a 2013 SalesForce study, the average CTR for an external website ad is .02%. But our ads have more in common with a sponsored page post, which has a CTR of just over 2%.
2% is in line with what Andrew McDermott, a Facebook preferred marketer, observed in his Quora answer.
Below 1% – not good.
Above 2.5% – awesome.
A Clear Winning Content Idea emerges?
Often, a clear winner will rise to the top.
For me it’s happened almost every time.
For ad test #1, Ad A was the clear winner.
It had 13 clicks, and the other ad had zero.
Admittedly, I ended the test early because the results were clear.
Use your best judgement if you want to end the experiment early. The test is only meant to guide you in making your choice. Because of that, we’re not looking for scientific accuracy (which would cost us a whole lot more than twenty bucks).
For ad test #2, Ad B was the clear winner.
Ad B had a CTR of 2.7%.
Ad A had a CTR of 1.06%
I tested the title for this post too…
Example B
Again we had a clear winner.
Both titles were shown to about 350 people.
Both had really good click through rates.
But the title: How to Know if Your Content Idea is a Winner (or a Loser) Before You Hit Publish.
Won big time.
Ad A: 14 clicks and 3.79% CTR
Ad B: 24 clicks and 5.38% CTR
Facebook can skew the results – and that’s ok
Part of the reason that my first ad had dramatically more clicks is that Facebook showed it to more people.
As Facebook figures out which ad will perform better, it optimizes your campaign to deliver that ad to more people.
Why doesn’t that ruin the experiment?
It actually makes the decision easier. Facebook is smart.
Their algorithm is going to serve the ad that will perform the best.
Take advantage of Facebook’s algorithm, and let it help you decide which topic / headline will be the most popular and perform the best.
If you’ve targeted the audience that reads your blog, then you’ve just leveraged Facebook’s billion-dollar algorithm to figure out what your audience likes best.
Go with it.
It’s okay if there’s no “winner”
If your results are split pretty evenly, don’t try to look for a winner where there isn’t one.
When both ad groups perform well, you know that you have two solid content ideas.
Caveat: The test works best with top of the funnel content. Content that will have a broad appeal is easiest to test. You can still split test headlines for other content.
Just because a topic performs poorly doesn’t mean you shouldn’t write the post.
For example, your product vs. a competitor’s product may not get a lot of clicks, but it may still be a key piece of content on your site that helps convert visitors into customers.
De-Risking Content Creation Is Key to a Winning Content Marketing Program
By testing your content ideas before you hit publish, you can reduce the chance that your content will flop, you’ll make sure your team focuses their efforts on creating content that will perform, and you’ll create hit pieces more often.
Get the screencast here.
Leave a Reply