36 essential A/B testing best practices to boost your conversions

Kyle’s Note: Increasing conversions on your site is one of the most effective ways to increase profitability in your business. It’s also an area where many entrepreneurs struggle with where to start. An easy way to get started is by using A/B tests.This post discusses generalized best practices as well as ideas specific to landing pages or blogs.

Most online businesses have one thing in common: they need to monetize their website visitors. Usually, they do so by turning them into coaching clients, selling digital or physical products, or having their visitors click on ads on a website.

Whatever the monetization strategy is, the website visitor has to perform a certain action for the business owner to earn money. The ratio of visitors who perform that action compared to those who don’t is called “conversion rate.”

You already guessed it, the higher the conversion rate; the better the business performs.

I’ll show you how to increase your conversion rate with a strategy called A/B testing.

A/B testing means that you’re creating several versions of your website who differ in layout, copy, colors or whatever (we’ll cover things to test later in this post).

You can run A/B tests with free tools like Optimizely, or you can choose paid tools like Nelio A/B Testing or Visual Website Optimizer. It’s up to you; the principle of A/B testing is the same for each tool.

Let me explain 36 A/B testing best practices you can start leveraging today:

General best practices for A/B testing

The first post section is about general split-testing advice

1. Focus on user experience

You want to make using your website as seamless and enjoyable as possible for your readers. Think about why people come to your website and what they expect from your site.

It’s the emotions of your users that often decides about whether they sign up or purchase your product. And you can influence that emotion to a certain degree.

As Neil Patel states:

“Any emotion can drive conversions. A user might convert because they’re happy, sad, jealous, or downright furious. But the really important thing is why they are feeling that way.” (Click to tweet)

Neil Patel on customer emotions

Related: How Emotions Affect The Decision-Making Process In Online Sales.

To create a great user experience, you need to know the emotional state of your website visitor. Not assume, know.

A/B testing is the perfect way to collect data on the emotions of your user. Test various color schemes to invoke positive emotions. Remove anger by rephrasing added charges. Test which image performs best.

Here are some questions to ask yourself when setting up A/B tests that focus on the user experience

  • What do your readers want?
  • What contents do your readers need to see or expect to see?
  • How did the user get to your page? (Campaign, social, SEO, etc.)
  • Why did the user get to your page?
  • What does the buyer’s journey look like?

Once you put yourself in the shoes of your customers, you can then form a hypothesis for your A/B test. That hypothesis serves as the foundation for your experiments and will show you exactly, what things to test.

Related: How to create content that follows the buyer’s journey.

2.Manage tracking pixels properly

Tracking pixels are tools to create custom audiences, e.g. on Facebook. They are little code snippets you add to your website that keep track of how far a visitor goes into your funnel.

By using tracking pixels, you can run ads specific for visitors and leads at all stages:

  • People who just visited your site but didn’t perform the desired action
  • People who subscribed to your list but didn’t buy
  • People who bought your entry offer, but didn’t get your upsell
  • And so on.

Though this tip is not necessarily specific to A/B testing, it’s crucial in the context of A/B testing.

While this post covers split testing strategies specifically for your website, you can also run split tests on the ads that you’re using in your paid traffic campaigns.

This case study shows how Design Pickle increased their MRR by $5.8k using Facebook Ads.

Make sure that, when setting up A/B tests on your site, you also pay attention to each tracking pixel you run.

Each tracking pixel helps you create a market segment, and you want to run ad campaigns and build landing pages that have no friction in them.

Your ads and your landing pages should use similar (if not the same) wording and images, but more on that later in this post.

Related: The Practical Retargeting Guide.

3. Split-test your funnel top down

The first pages of a funnel get the most traffic, that’s the nature of almost any funnel. You drive traffic to a landing pages using ads, guest posts, or SEO. From that landing page, people can opt-in and get their lead magnet.

When you’re running a sales funnel, you often present the first paid product after a lead downloaded the lead magnet. Then you might have up-sells and down-sells depending on the behavior of your lead.

Notice something?

The only traffic that is not depending on the user behavior is the traffic that comes to the landing page. ALL other pages will see lower traffic, based on the conversion rate of each page.

This allocation of traffic is why you should test your funnel top-down, meaning you should start running A/B tests on your landing page. Once you increase the conversion rate on the landing page, you’ll have more people inside your funnel and will be able to split-test the pages deeper within your funnel.

Remember that A/B tests need a certain amount of traffic to produce reliable results. Getting these valid results for the last offer in your funnel often will take at least 5x – 50x longer (depending on your funnel) than getting results for the landing page. And once you get more people into your funnel, you’ll automatically see more engagement.

Related Case Study: [Split Test Results] Top Funnel Copy Change Lifted Engagement By 15.2%.

4. Split-test specifically for each of your campaigns

I already mentioned in #2 that you need to remove friction between your ads and the pages in your funnel.

That’s why I (and many other CRO experts) recommend using an individual landing page for each campaign.

Tyson Quick, founder of Instapage, states:

“People are in quite different mindsets and settings when coming from different traffic sources. A Google AdWords click is typically coming from a visitor who is directly looking for a solution to a specific problem they’re already aware of, while a visitor from a Facebook or Twitter advertisement is typically still in the problem/solution discovery phase. The post click experience should reflect this mindset.”

Tyson Quick on the traffic source specifics

Related: How Different Traffic Sources Impact Your Conversion Rates.

Say you’re using Facebook ads to market to two different market segments. My recommendation is to create an individual landing page for both market segments that target their specific needs.

Those landing pages don’t need to be totally different from each other. You can just start with adjusting the copy and call-to-actions to meet the situation of your visitor and match their context.

5. Select the right audience to run your tests with

Depending on the size of your audience, you might want to run A/B tests with only a portion of it. If you have only 2,000 monthly visitors this in no option, but if you have 50,000 monthly visitors, it totally is.

Why?

Well, split tests need a certain amount of traffic to bring valid results. Jon Loomer says he expects at least 1,000 visitors to see an experiment before it analyzes the results.

If you get that critical mass in a reasonable time with just a portion of your traffic, go ahead and segment your audience. If not, use your full audience.

Another reason to divide your audience for split-testing is to test:

  • Mobile traffic vs. desktop traffic
  • Search traffic vs. paid traffic vs. social traffic
  • Traffic for each campaign (see #3).

Related: How Secret Escapes Increased LTV and Doubled Signup on Mobile

A/B Tests For Landing Pages

20 A/B testing tips for landing pages

6. Test completely different layouts

Testing “apples vs. oranges” is a good way to start your split-testing your site.

You create two rather different variations of the same website and see which one performs best. You then use that version to conduct further split-tests based on the better-performing layout.

I did such an A/B test for WPRSSAggregator.com. It’s a site that sells add-ons for the WP RSS Aggregator plugin by Jean Galea, which has 300k downloads and counting.

The goal of the test was to increase the conversion rate on the sales page, which essentially lists all add-ons Jean is selling.

Our first step was to simplify the sidebar and thus removing distractions. We ended up removing three widgets from the sidebar and giving the site a cleaner layout.

The results?

For some plugins, we increased the conversion rate by up to 241.6%.

You can adapt that test almost directly to any site that uses a sidebar, and remove all elements besides the primary call-to-action.

A common application for this best practice also is the test of long-form landing pages vs. short-form landing pages.

Related: Crazy Minimal Homepage Increased Leads for an SEO Service Company from 1.39% to 13.13%

7. Test your main headlines

Landing pages usually have three types of headlines:

  1. The main headline
  2. The supporting headline
  3. Reinforcing headlines

These headlines are what people should notice when skimming over your site (which is what most visitors will do). You want to make them consistent with your overall copy and compelling.

The job of headlines is to get your visitors excited about what you have to offer and to keep them on the page. If they fail, you’re losing leads and sales.

Copywriting experts like Jeff Goins, Megan Marrs or Neil Patel have given actionable advice on writing great headlines. You just need to test which type of headline converts best for your specific landing page.

Headlines are always amongst the most important A/B tests and are proven to have major impacts on the conversion rates. If you’re optimizing your landing pages, there’s no way around split-testing your headlines.

Related Case Study: How We Improved Landing Page Conversions by 79.3%

8. Test your reinforcing headlines that come throughout your page

As mentioned in #7, reinforcing headlines are important to keep people excited about what you have to offer.

They will remind people who read the full content about the benefits of your offer.

Visitors who just skim through the site will only read your reinforcing headlines and make a decision based on those alone.

That’s why you better pay attention to them and write them as effective as possible. And here’s how you do that:

  • Do your research. Before applying any magic headline formula, ask your existing customers what they like most about your offer and why they bought it. Use the exact words they used in your copy, the chances are that those words will resonate with your prospects better than your own.
  • Heat maps. Use heat maps on your site to find out which sections of the page get the most attention. Optimize the headlines in those sections first, and don’t waste time on areas that don’t get any attention.
  • Systems. Use success-proven formulas to write headlines.

Related: 3 Landing Page Headline Formulas to Target Your Prospects

9. Use hero shots on your landing pages

Asknative has a great hero-shot on their landing page

“Hero shot” is a term I first heard from Oli Gardner, co-founder of Unbounce. It describes an image or video that shows your offer in the context of use. The goal is to showcase the benefits your customers will experience from buying your offer.

Since we have so many different kinds of landing pages, there are many variations of hero shots you can test as well. Videos are often combined with short-form landing pages, but in some cases, images might even outperform them.

In 2014, we had two major trends in hero shots, which is the full-screen image and the video hero shot. You’ve probably seen many landing pages with either of those elements.

As usual in building landing pages, there is no “right format” for a hero shot. You never know what works best for your audience until you’ve split-tested it.

When testing your hero shot, though, it’s important to ONLY testing the hero shot and leave the rest of the page as is. It doesn’t make sense to compare a video sales letter on a short-form page with an image on a long-form page, as there are too many differences that also affect the conversion rate.

Related: 7 secrets of hero shot images that drive epic conversions

10. The benefit statement

The benefit statement explains to your website visitors in written form, what benefits they’ll get from purchasing your offer. Usually, the benefits are presented in bullet point format, which makes them easy to read and skim through.

They’re often placed in the context of the hero shot, to make sure your visitors pay attention to them. And most importantly, they have a major impact on your conversion rate.

Here’s what makes benefit statements perform badly:

  • Their wording doesn’t match the copy of the rest of the page
  • They use a standard “bullet-point” icon
  • The bullet points are too long
  • They don’t stand out visually

If you pay attention to these four factors, you’re in a good starting point to write a well-performing benefit statement. But of course, you should run A/B tests to find which copy & styling works best.

Related Case Study: 18.59% Increase in Downloads by Tweaking one Bullet Point

11. Adding (or removing) social proof

You’ve probably heard that adding social proof to your landing page is a surefire way to increase your conversion rates, right?

What if I told you that removing social proof in some cases increases the conversion rate even more?

By having testimonials on your landing pages or sales pages you proof that your customers actually like what they bought. You hope to entice your website visitors to opt-in or buy using these testimonials.

In some cases, however, testimonials make the page too long, and people abandon the page without taking action. And you don’t want that to happen, do you?

Related: Debunking the Social Proof myth

Related: The Science of Social Proof: 5 Types and the Psychology Behind Why They Work

12. Test the color scheme of your landing page

Countless discussions about A/B testing involve the coloring of elements like buttons or text color. So-called experts might have told you myths like green buttons always outperform red buttons or similar nonsense.

The truth is that there’s no “best color”. Only three factors should influence the coloring on your landing page:

  1. Color psychology
  2. The context of the call-to-action
  3. Branding

Knowing at least the basics of color psychology is important, as it helps you set the frame for your visitors. Remember that in #1 we emphasized the emotions of your visitors? Coloring is a major influence in this regard.

The context of the call-to-action also is important, as you want your CTA to stand out visually on the website. Using a light-green button on a dark-green background most likely won’t perform well – but orange on green most likely will.

Last but not least your color scheme needs to follow your branding or corporate identity. Branding is least important to me because (in my opinion) a high conversion rate is more important than using a color only because you use it somewhere else in a different context.

“A rule of thumb is that things work well if they either extremely blend in or extremely blend out” – Jeff Wenberg, LeadPages

Jeff Wenberg on colors of elements

Related Case Study: [Split Test Results] Changing Website Colors = 50% Bump In Conversions

13. The call-to-action

Landing pages should have only one call-to-action. Not two, not three – one.

If you’re going for opt-ins, focus the landing page on that and remove unnecessary navigation, links or sidebars. If you’re A/B testing your sales page, remove any distracting elements besides the ones necessary for your visitors to purchase your offer.

The call-to-action should be presented prominently on the website, using colors that stand out from the context (see #12).

A pro-tip from Oli Gardner is that you should remove any elements around your CTA that might have a negative connotation. Even removing a money-back guarantee is worth testing, as that makes your visitors think “what if this doesn’t work for me?”.

What we see recently, is that two-step opt-in processes in most cases outperform one-step signups. On the WP Summit, LeadPages educator Tim Paige explained this as follows:

“When you put an opt-in form on a website, people see the website as a ‘taking’ page. Whereas when you only have a button on the website that opens a popup with an opt-in form, people are more enticed to subscribe. They already took the first step (clicking on the button) and now feel the urge of following through (subscribing).”

Besides testing a two-step subscription process, you can of course experiment with:

  • Button copy, colors, and sizes
  • The context of your CTA (guarantees, badges, etc.)
  • The positioning of your CTA (above vs. below the fold)

Related: 10 Call-to-Action Case Studies w/ Takeaways & Examples from Real Button Tests

14. Test “my” vs. “your” copy

UX professionals and conversion rate experts know that A/B testing copy is crucial, especially when it comes to testing “my” vs. “your” copy.

An example for testing button copy would be:

“Download My Free eBook” against “Download Your Free eBook.”

Using “my” copy extends your website as being a part of the visitor, you imitate the view of the user. Your visitors will start seeing your landing page as an extension of themselves, as a tool they can use.

Instead, using “your” in your copy will create kind of a break between the landing page and the visitor, establishing a conversation between two sides.

Dustin Curtis wrote an interesting article on this topic.

Case studies in A/B testing have shown, that in some cases you not even want to use either “my” or “your” in your call-to-actions. These findings prove only one point: you need to run A/B tests to find whether “my”, “your”, or general copy works best.

Related Case Study: 31.03% Increase in Sales by Tweaking the Call-to-action Copy on a Payment Page

15. Fine tune your closing argument (P.S.)

Julie Boswell from Rebel Enterprise says:

“Your copy will be skimmed… plan for it.”

Many if not most readers will skim over your landing page from top to bottom and then decide whether they read it or not. They’ll briefly read through your headlines (see #7 and #8) and end up at the CTA you placed on the bottom.

Skimming is exactly the reason you should have a closing argument, as a “P.S.” below that CTA. In that closing argument, you should recap the benefits of your offer, who that offer is for and why people should act now and not later.

Running A/B tests against the closing offer is overlooked by some marketers, yet it yields high potential. It makes communicating to page skimmers more efficient, visitors, you probably wouldn’t reach without an effective closing argument.

Related: How to Increase Sales Letter Conversions (Without changing a single word)

16. Test mobile vs. desktop traffic

Making your landing page responsive should be mandatory and self-explanatory by now, but still there are many WordPress landing page themes or other frameworks that don’t adapt to mobile devices.

Do you know how much traffic on your landing page is coming from mobile devices? If not, go to your Google Analytics and open “Audience” -> “Mobile” -> “Overview.”

Mobile traffic statistics in Google Analytics

Tools like Optimizely give you the option to run split-tests specifically for mobile or desktop traffic, which is a great idea.

Here’s why:

  • You want to test short headlines on mobile devices compared to longer headlines on desktop screens.
  • You want to hide certain elements on mobiles to highlight the most important contents
  • You want to make sure that your mobile visitors can easily skim through the contents and still get the main message

While running A/B tests on any page, you need to keep the context of your visitors in mind. They use mobile devices while on the bus or train, while walking and sometimes even while driving their car.

Mobile users simply can’t pay as much attention to your landing pages as desktop users. Once you fully grasped what that means, chances are your mobile landing pages will look different than your desktop versions.

Related Case Study: B2B Web Optimization: 140% surge in mobile transactions through responsive design effort

17. Remove friction between campaigns and landing pages

“One of the most important lessons is that friction kills conversion. The good news is that reducing friction is one of the most effective ways of increasing conversion. The bad news is that it can be difficult to spot sources of friction if you don’t know what to look for.” – Michael Aagaard (Unbounce)

Friction is anything that makes people think too much about what they’re doing on your landing page; that’s slowing down their decision process in your sales cycle.

Friction could be that the ad you’re using doesn’t match the color scheme or wording of the landing page you drive traffic to. In this case, your visitors are wondering whether they’re on the right website (and in most cases they’ll leave instantly).

Since we’re bombarded with marketing these days, we pay less attention and become less patient with everything we do online. Removing the friction can be quite a complex process, so be prepared to run some A/B tests in this context.

Related: The web entrepreneur’s guide to creating and using landing pages

18. Hide what’s not essential

Landing pages have one purpose only: get visitors to act on your call-to-action. That’s it.

And here’s what that means for you:

You need to remove any element on your landing pages that don’t make your visitors act on your CTA.

Here are some of the most common elements I remove for my clients:

  • Menu bars
  • Outgoing links
  • Search forms
  • General content in the header or footer

The more focused your landing page structure is, the easier it’ll be for your readers to understand what you want them to do.

And that’s what it’s all about. You want your readers to know that they should opt-in to download your freebie. Wouldn’t the hours creating it be wasted if they didn’t?

Related: Minimalist design: an offer you can’t refuse

19. Personalize your landing pages

You can add a little flavor to your website by adding dynamically personalized elements to them.

For example, Neil Patel says that he wants to a business in my city successful:

Screenshot of Neil Patel's landing page

You could even get more specific and use software that detects the location of your visitor and displays a city name near them.

Or you can display personalized landing pages for a specific source of traffic like Barack Obama did in his campaign.

Redditors for Obama

Image from KISSmetrics

We see online marketers use personalized landing pages very often when they got interviewed. Smart marketers use an interview to drive traffic to a specific landing page where they give away a bonus resource, which expands on the interview.

This way they’re generating leads from an interview that would have otherwise only driven attention to the show host.

Related: How to Add Personalized, Triggered Changes to Your Website with inSite

20. Test various money back guarantees

By show of hands, who offers a money back guarantee to their customers? Ok, most of you will have raised your hands. Next question.

By show of hands, who tested 5 or more different money back guarantees to see what works best?

I’d bet only very few of you raised their hands. And that’s a pity because you might be leaving money on the table.

The truth is that we want to feel secure when purchasing something, especially in the online world. Money back guarantees give us that feeling of security, giving us the option to return our purchase if we don’t like it.

I bet you already ordered a shoe in two different sizes to see which one fits best, didn’t you? We can do that only because of the money back guarantee.

As case studies have shown, there’s more than our average “30-Day Money Back Guarantee”. In fact, there are at least 39 scripts for guarantees you can use.

Split-testing those scripts that make sense for your business can be a true game changer.

“By replacing all of our money back guarantee badges with free trial badges and by placing “30-day free trial” offer on every page of the Crazy Egg website, we were able to boost signups by 116%” – Neil Patel

21. Test various price points

Example pricing table

Pricing is a tricky topic. Most entrepreneurs I talk to either price their offers depending on the competition, or they go with their gut. Neither is the ideal choice if you ask me.

Let’s go through an example:

Say you’re selling an online course for $497, and you get a 10% conversion rate on your sales page. For every 100 visitors, you’re making $4.970.

Now you’re running a split-test and get a 15% conversion rate at $397. That means for every 100 visitors you’re now making $5.955.

Even though you decreased your price by $100, you effectively increased your revenue by $985 per 100 visitors. That’s a good deal if you ask me.

Or you might find that higher prices convert almost as good as your original price.

Maybe a second A/B test shows, that a $697 price tag leads to a 8% conversion rate on your sales page. That would earn you $5.576 per hundred visitors – increasing your revenue by $606 per 100 visits.

While this all sounds good, you’ve got to be careful when split-testing your pricing strategy:

You might upset existing customers when offering a lower price and will have to justify that price. Also, it’s hard to achieve statistical significance if you can’t run the test for a very specific market segment (e.g. based on a targeted traffic campaign).

There’s a lot of content to read about split-testing your pricing strategies, so before you take action on this point, please do your research.

Related: When A/B Testing Prices, Proceed With Caution

22. Test payment gateways

Did you ever hear that people trust some payment gateways more than others? Or that some countries require security layers like 3-D Secure while others don’t?

Having the right or wrong payment gateway can be a major factor influencing the rate of abandoned shopping carts. Some of your prospects might prefer Stripe while others love PayPal or any other payment system.

Also, certain payment providers require additional information on the sales page or shopping cart. Clickbank, for example, requires you to add text to the bottom of the sales page, telling your prospects you’re selling through Clickbank.

While marketers might debate whether it’s ethical to split-test payment gateways (as they might charge different fees), you should at least know about this option and have it in the back of your head.

23. Two step processes for opt-ins or checkouts

Research has shown, that two-step processes for opt-ins convert better compared to placing the opt-in form directly on the page.

“A website that has an opt-in form embedded directly feels like it’s a ‘taking’ page, taking information from the visitor.” – Tim Paige, LeadPages

Don't make your website a "taking" page - Tim Paige

LeadPages introduced this functionality with their LeadBoxes, but you can use tools like OptinMonster, OptimizePress or any popup script that opens on a button click.

Stripe often comes with a two-step checkout process as well, just like WP Curve uses it on their Signup page.

The beauty of this two-step process is as follows:

The first step (clicking on the button) is a low commitment. It’s easy for people to click on that button because they don’t have to decide what to do afterward (yet). The click itself doesn’t bring any commitment to your visitors.

However, after clicking on the button and making this first step, your visitors feel more compelled to follow through and finish the process by taking the second step as well. When humans start something, they want to bring it to an end. And that’s exactly the principle we’re leveraging with a two-step opt-in process.

Related: LeadBox vs. Regular Opt-in Box: See Which Won in the Big LeadBoxes Split Test

24. Pay attention to your heat map

Heatmap example of the WP Summit

If you don’t know where to start your A/B testing, having a look at your heat map will help.

Heat maps track the scrolling and clicking behavior of your visitors. Areas that get a lot of attention show up as red (read: hot) and sections on your site that don’t get a lot of attention show up as blue (read: cold).

You can use tools like CrazyEgg, Nelio A/B Testing, or SumoMe to run heat maps on your site – and I highly recommend you do so. These tools track the mouse movement of your visitors.

Another variation of heat maps is eye-tracking. That’s when users are paid to use a website, and their eye movement is tracked by cameras.

Tracking the mouse movement often is cheaper than leveraging eye-tracking heat maps, yet both bring important insights that’ll help you increase your conversion rates.

Related: What’s in a Heatmap? Predictive Eye-Tracking versus Click-Tracking

25. Run split-tests based on day/time of the visit

Tools like Optimizely give you the option to run A/B tests based on the day and time your visitors come to your page.

Why is this important?

Let’s go back to the context of your users. Do you think about the same things in the evening as you did in the morning?

The chances are that you experienced something throughout the day that shifted your focus, sparked new ideas, or raised new questions. This difference of focus is even more obvious when we compare your focus on Mondays to the focus on Fridays.

When the week is ahead, most people think like this “only five days until the weekend.” When Friday has come, they start celebrating life and enjoying their time. This, of course, doesn’t appeal to the lucky business owners who spend their time the way they want to.

So, let me circle back to the split-testing topic.

Say you’re running an online portal selling trips and events. On Monday-Thursday, you could test putting relaxing events on the front page. Advertise those events that help people forget about their work, reduce their stress levels, and so on.

Then, from Friday to Sunday, you could promote adventurous and exciting trips, perfect for a great weekend.

Makes sense?

This is just a made-up example, though. But it illustrates how you can run time-based split-tests to increase conversion rates.

26. Run A/B tests based on browser language

This best practice takes a new spin on the personalization approach we covered in #19. Showing your landing page in the right language most often gives you a huge boost in trust and credibility – and you can test that with split-testing.

Sure, translating long-form sales pages can require hours and hours of work (if done properly). Yet, it’s a great way to refine the copy on your site and make it more compelling.

Speaking to people in their native languages makes it easier to build up a connection with them, to spark emotions. And that’s exactly what you want to do on your landing page.

Running A/B tests on blogs

9 A/B testing best practices for blogs

Now, we’ve covered a lot of best practices and A/B testing ideas for landing pages, but what about blogs?

Most blogs nowadays serve the purpose of list building, meaning to turn your website traffic into email subscribers.

But even if you’re using ads to monetize your blog, focus on engagement, or sell directly (especially when you sell) – running A/B tests on your blog probably is a good idea.

Let me explain:

27. Define the goal of the split-test

Usually, blogs work as the top of the funnel. They attract people via SEO or social marketing – recently we also start seeing ads directing to blog posts with CTAs.

Now, what is the main CTA on a blog post?

Is it to read the post and to leave a comment?

Is it to subscribe to your list using the opt-in form below the post or the one in the sidebar?

Or is it to share your content on social networks?

Nobody can answer that question but you. And it’s important that you take the time to answer it before you dive head-first into A/B testing.

Once you know what goal you want to achieve with your blog (or an individual post) you can start phrasing a hypothesis for your test.

Let me say this the other way around: without knowing the goal, how do you know you’re going in the right direction?

Related: The WP Curve content strategy, and how to build your own

28. Use your Google Analytics to find the pages to split-test

A landing page is a single page, and even funnels usually consist of a handful of pages. A blog, however, can have hundreds of posts. Which one should you focus on for A/B testing?

If you don’t have a plan on testing a specific post that you’re running ads for or using as the top of a funnel, here’s what to do:

Open your Google Analytics account and open the report “Behavior” -> “Site Content” -> “All Pages”.

Traffic allocation for jkoch.me

This report shows you, which pages on your blog get the most traffic. I’d recommend starting on the top page and then expand the split-testing to maybe the top 3 or 5 pages on your blog.

29. Test various page layouts

You always want to test various page layouts on your blog.

Does your blog have a sidebar? What about removing all widgets that aren’t necessary? Doing so got me a 255.9% increase in conversion.

Or you might want to remove the sidebar completely. Bryan Harris from Videofruit reported a 26% increase in conversions after removing the sidebar completely.

The sidebar is only one example to test. Removing featured images, comment sections, social sharing buttons, or other elements is worth testing as well.

30. Test headlines on posts and pages that get the most traffic

Sounds familiar? We covered headlines in #7 and #8 already, but now we’re talking about a different purpose.

A good blog post headline keeps the reader on your site and gets him excited to read the next line of your content. That’s especially tough when you get cold traffic from Google.

Experts like Joanna Wiebe or Jeff Goins have created various formulas to write great headlines but do you know which one works best for your content?

I recommend setting up at least one A/B test, testing three different headlines for your best-performing post. You’ll see over time which headline keeps the most readers on your site.

Related: How to A/B Split Test WordPress Post Titles to Get More Clicks

31. Test CTAs below posts

Call-to-actions below posts are known to perform quite well. The reason is, that people already read a part your content when they see the post (or at least skimmed through it).

If the content got their attention, they’ll be eager to get a content upgrade or get updates on the topic by subscribing to your list.

As with all CTAs, you can test a variety of things in this context:

  • The coloring of CTA, fonts, imagery
  • The font sizes and font faces
  • The copy of the headline and CTA
  • The button size, color, position
  • Adding social proof or hiding social proof.

32. Include CTAs in your content

You can test including CTAs directly in your content and not just in the sidebar or below the post.

Especially when you’re using a blog post as the top of a funnel, to direct visitors to a squeeze page, having CTAs in your content can be a good idea.

When you readers see those CTAs, they’re directly involved in the topic. They’re just reading a post about it. If that post gets them excited, there’s a much higher chance of converting them into leads right then and there – compared to having an opt-in form at the bottom of the post, when they maybe lost the excitement again.

Related: 6 Proven Ways to Boost the Conversion Rates of Your Call-to-Action Buttons

33. Test removing non-essential elements

Blog templates usually come with all sorts of elements that distract your visitors from the essential: your content and your CTA.

Those could be icons to like your post, breadcrumbs, the date of posts, featured images, or similar things.

We learned in this post that focus is key for landing pages, that we need to make it as easy as possible for your readers to understand what we want them to do.

The same is true for A/B testing blog layouts. In #27 you defined a goal for your blog. Anything that distracts your readers from taking the desired action needs to be removed.

Using split-testing, you can find the best converting layout step-by-step.

34. Split-test featured images in blog posts

Featured images can be very beautiful, and spark engagement on your site, or they can kill it.

If you decide to use them on your blog, I highly recommend split-testing to find the one image that converts best.

The Titans of Marketing, for example, used an amazing image in their blog post about webinar hacks.

This featured image is just stunningly creative

The goal of the post, is to turn visitors into email subscribers, especially those who want to learn about sales strategies on webinars.

See how their featured image helps them emphasize the benefit of the reader? If you’re using featured images, this is how to do it (similar to the hero shot in #9).

A/B testing your featured images is also helpful when you split-test certain images in ad campaigns. As we said earlier, you need to remove the friction between ads and the page (or in this case blog post) the ad directs people to.

So you could set up an individual URL for each ad image, repeating the image as featured image in the post.

Related: 12 ways to keep visitors on your site longer

35. Two-step opt-in forms

This tip is the last recommendation for running A/B tests on your blog, and it’s a powerful one.

I already said that, for landing pages, two-step opt-in forms can heavily outperform a simple opt-in form. And the same is true for blogs.

In WordPress, you can use plugins like LeadBoxes, OptinMonster or Popup Maker (needs a bit of coding) to install popups that open on click of a button.

Run A/B tests that put those buttons directly inside your content, in your sidebar, or below the content. Test different copy, button colors, and all the rest of it.

In most cases, you’ll see an increase in conversions.

Conclusion

A/B testing landing pages and blogs is a powerful marketing instrument, especially when you’re optimizing a funnel for lead generation or sales.

Increasing your conversion rates helps you do more with the traffic you already have, and you can get started for free.

But A/B testing can also become overwhelming.

Make sure to test only one thing at a time, and to pay close attention to the audience segments you’re using.

For more ideas on where you could improve conversions on our site, check out our 18 point conversion review. Download it for free below.

Grab the free website review template

We will email you the full website review template (Google Doc). You will also join 20,000 people and get my weekly emails with business and online marketing advice. 

 

About

Jan is the co-founder of ABCD Impact. He runs A/B tests, and helps business owners increase their conversion rates. When he’s not working on ABCD Impact, he hosts the global WP Summit, blogs, listens to loud house music, and enjoys experimenting with fitness and health.

18 responses to “36 essential A/B testing best practices to boost your conversions”

  1. Nicole H says:

    BOOM!!! That’s a helluva’ lotta’ knowledge you’ve dropped here, Jan! Once again, you over deliver! THANK YOU!

  2. That’s a fantastic piece Jan. Looking forward to use some of these techniques as we are about to a/b test some new landing pages at Cloudways.

  3. This is a seriously fantastic resource!!! Thanks so much for mentioning https://instapage.com – I would love to chat some more, do you mind emailing me at kieran@instapage.com? Thanks so much!! 🙂

  4. Jan Koch says:

    Hey Kieran,
    Thanks a lot for the kind words!
    I’ll shoot over an email right tomorrow morning, looking forward to get in touch with you!

    Cheers

  5. Jan Koch says:

    So glad you like it Muhammad!
    Cloudways has been a huge help for me in the past, looking forward to hear more about your results 🙂

  6. Jan Koch says:

    Your ongoing support means a lot Nicole,
    thank you!

  7. Good stuff @jan_koch:disqus, glad to see you contributing here.

  8. Jan Koch says:

    Thanks @FollowJames:disqus!

  9. Joy Foster says:

    Point 34 is interesting… how in the world do you split test two images using wordpress though?

  10. Joy Foster says:

    ha… and I’d also like to know how you isolated the 3rd widget on the sidebar… on this post.

  11. Kyle Gray says:

    Hey Joy,

    I believe you’re talking about the “Tools to scale your content marketing” widget.

    We use the Q2W3 fixed widget. It’s easy to use and install.
    https://wordpress.org/plugins/q2w3-fixed-widget/

    Hope that helps, let me know if you have any other questions!

  12. Joy Foster says:

    Oh oh! Yes yes! Thanks!!

  13. Jan Koch says:

    Hey Joy,
    sorry for the late reply, I’m attending a conference currently 🙂

    You can use the tools mentioned in the post to set up variants of a post / page that include two separate features. Usually, one image (the control version) would be added in WP itself, while the other images are created through the split-testing tool.

    Hope that helps!

  14. Jeff Wenberg says:

    Hey Jan! Thanks for featuring LeadPages, Tim, and I in this fantastic article. Great work in fleshing out so many great points!

  15. SocialQuant says:

    This was wicked awesome.

    Just started playing with our signup page now to see what converts better and so much to test. Trying to make it as frictionless as possible yet people still are spending 2.5-3 minutes on the page. Definitely have to try all the above.

    Killer post. ~ Mike

  16. Kyle Gray says:

    Thanks for the kind words Jeff! We have been using Lead Pages with WP Curve and it has had a big impact on our opt-ins.

  17. Kyle Gray says:

    Hey Mike!

    Please follow back and let us know what you tried and what results you saw. Friction is a really interesting element to experiment with.

Leave a Reply

Your email address will not be published.

WordPress problems?

Our WordPress experts have you covered.

Hyper-responsive 24/7/365 WordPress support, maintenance and small fixes.