On our first post Publishers: Increase your conversion rate, we described the importance, as a publisher, of improving your conversion rate (the percentage of your visitors who end up reaching a given goal or performing a certain action, such as visiting a certain page, reading a whole article, or signing for your newsletter). With an easy example, we also described a basic technique to analyze your website, define your goals, and test your website. There, we concluded that the goals in our website (and the testing elements related to those goals) were:
- Visiting and Reading our Published Posts
- Write different Titles?
- Use alternative Featured Images?
- Compose alternative Excerpts?
- Modify the number of visualized posts?
- Re-order news?
- Use a different WordPress theme?
- Clicking on our Ads
- Re-order Ads?
- Use different call-to-action elements?
- Sharing our Content in Social Networks
- Change the location of our social links?
- Detect (and remove) distracting elements?
- Remaining More than 2 mins in our Pages
- Do readers scroll until the end?
- Do I have the right categories or tags?
This post continues the previous one by illustrating one of the essential A/B tests we can perform to reach the first goal: make visitors read our posts.
Goal: Visiting (and Hopefully Reading) our Posts
In order to reach this goal, you have to test the “Headline” of a post; that is, the combination of its title, featured image, and excerpt.
Test of Headlines
It’s clear that headlines are extremely important. If your visitors don’t like the headline of a post, they won’t read any further. How to create this type of test?
- Click New A/B Test for Headlines.
As we previously stated, when testing the headline, you can be interested in testing any combination of:
- the title of the post,
- the featured image, or
- the excerpt of the post.
Let’s see the whole process with the following example of a company that sells solar panels. Suppose you want to publish a new post that describes why anyone can save money with the solar energy.
The original title of the post is: “Save money with the solar energy”. However, your team is not very sure about whether a different title would attract more readers. They propose an alternative title: “Save hundreds of dollars with the solar energy”.
Your team also discusses whether the image or the excerpt of the headline should also be changed but, to start with, you’ve decided that it’s enough with just changing the title.
Fill out the Basic Information
First, fill out the Basic Information of the Headline experiment under the General Tab.
Basically, you just have to define (1) a meaningful name for the experiment with an optional description, then (2) select the post you want to put under test, and finally (3) decide under which condition the running experiment should stop.
See our running example below:
Create Alternative Headline(s) for Spit Testing
Then, under the Alternatives tab, (1) click New Headline which will allow you to create the new alternative headline, where you (2) set up the new title (also new excerpt and/or featured image) and (3) click the Create button. Note that you can create as many headline alternatives as you want. However, don’t try to test a lot of alternatives if you don’t have much visitors in your site. In this case, it is better to start with just one alternative and see what happens.
The experiment has already been created and it’s already prepared for execution. In the Experiments page, you select the new experiment you’ve created and then click Start to launch it.
Automatically, half of your visitors will see the original headline and the other half the alternative one. In any other case with more alternatives, the visitors would be segmented in as many groups as different versions of headlines you’ve defined.
Now, you have to wait for a while and…
Find the Results of Your Headlines Test
When your test is running the most relevant information is always available in the Dashboard (
Nelio A/B Testing -> Dashboard). At some point, you may see the following information:
As you can observe in the image, even though the percentage of visitors that have clicked the original headline is higher than the alternative one (8,09% vs 1,12%), there is no confidence enough (see the red icon) to trust these results. If one alternative is actually better than the other, you’ll see it after a while.
If you want to see more information about the Progress of the Experiment, either click on the experiment’s card in the Dashboard or go to the Experiments page, hover over the experiment, and click View.
Moreover, you can also easily track the progress of the headlines tests with the following graphics:
Finally, you’ll find a Chart with the exact numbers and percentages used in the graphics as well as the information about the type of statistical test used to compute the experiment (along with a link to an explanation of the test displayed).
For How Long Should a Test Be Running?
Your test should be running until you get significative results. Remember that the confidence value tells you how “trustable” is the fact that one alternative is better than the original.
With a very low number of visits, you can not trust the results since they can be just due to pure chance. Imagine that only two people visit the variant of your tested page and both of them convert (100% conversion rate). Can you infer from this that everybody that will visit your site will convert? Clearly not! You need more visitors to see if this is the real trend or not. So, first of all, to get a minimum level of confidence in the results, they are not computed until you get at least 10 visits per variant.
On the other hand, the time that your A/B test should be running depends on the “complexity” of the test. The more complex the test is, the longer you will need to run the test to get a statistical sigificance out of your test. For instance, if you plan to try 10 different combinations of headlines, you will need to run the test long enough to identify a winning alternative. How long you should run the test also depends on your traffic on your traffic website.
And… What Happen When the Results are Confident enough?
At some point, when the results of your headline experiment reach a high level of confidence, and if the winner is the original one, just stop the experiment and now you have the certainty that your original headline is the best one (among those tested) for your post! But if the winner is the alternative one, you may use the Apply button that will overwrite the original headline with the alternative.
And that’s it! If you’ve arrived here, you’ve become an expert optimizing the goal of Visiting and Reading Your Posts!
In following posts, we will continue explaining other type of experiments you should create to reach your optimization goals in your WordPress site. Stay tuned!
So, publisher or blogger, what are you waiting for optimizing your conversion rate? See more information in Nelio A/B Testing for Publishers