How to Implement AB Testing on Google Analytics

Google analytics AB testing is desirable if you care to effectively hypothesize your marketing campaigns.

Everyday as digital marketers, we seek to create successful marketing solutions for business of different sizes.

Some of which includes; emails, social media ads, PPC campaigns and leadpages.

But sometimes the best optimized ad copy does not convert optimally.

When a low marketing conversion rate happens, it could be very frustrating if one considers the time and energy invested.

Hence, to ensure you obtain consistent success, you need to deployed a multivariate testing that is trackable for some of your online marketing implementations.

To do this, one digital marketing tool that is required is the Google Webmasters Tool.

In this post, you will discover A-Z how to implement Google analytics AB testing.

Don’t worry if you’re a beginner in the online marketing industry, this article is a walk-through regardless of your level of internet marketing experience.

So, relax and enjoy this post.

What is AB testing in Digital Marketing Parlance?

AB testing is an hypothesized experiment between two variables or testing samples.

It is often deployed by online marketing professionals in landing page optimization, email marketing, PPC campaigns and social media ads.

The goal of doing AB testing is to obtain optimal conversions from deployed marketing tactics. 

AB testing is otherwise referred to as bucket tests or split-tests.

Why is AB Test Important in Marketing?

The importance of AB tests in digital marketing cannot be over-emphasised.

Infact, to work smart as a digital marketer, you should really understand why you should frequently run AB tests in most of your marketing campaigns.

That said, let’s flip through why AB tests is very important to digital marketers.

To Understand Which Marketing Strategy Works Best.

See, experienced digital marketers understand that using a single web conversion strategy seemed like casting all of one’s oranges in one basket.

That’s not a smart strategic move, right?

Therefore, running AB bucket tests is important to gain an objective insight into the most performing marketing strategy.

Seriously, there are times a marketer will feel a particular test should deliver the best result but could realise the opposite after testing it with a variance.

This goes to show you that even the seemingly best marketing moves might not deliver the desired result or expectation.

Tentatively, the outcome will always be different based on the judgement of your leads.

This judgement could be determined by their tastes, needs, emotions, demographics or other sentiments.

To Consistently Achieve Optimal Conversions.

Believe me, using a single format to run all your marketing campaigns will not help you achieve an optimal result.

This is because the preference of a lead is subject to change over a period of time.

Hence, to meet-up with the changing preferences of your leads, you need to implement AB tests.

Doing this helps digital marketers plan ahead to make marketing decisions that meet the needs of their leads.

To Improve Website Visitors’ Engagement.

Running AB tests on your website also help to reduce the bounce-rate.

Which invariably means that web visitors are spending more time on your website.

Normally, setting up random AB test help to guide new visitors to navigate, act or relate with a website in a particular way as premeditated by the webmaster.

Using different marketing aesthetics for version A and B, website visitors could be guided to submit their personal information, subscribe to a newsletter, download a pdf or e-book, view a sales page or make a purchase.

To Recapture Leads For Effective Re-Targeting.

Oftentimes, a web visitor could bounce out of your website without finalising a process which has been initiated.

Yet, you need to find a way of re-capturing their attention to the abandoned process.

In your move to convert this category of web visitors, you might need to reach out to them again using an entirely different marketing template even though the deliverables hold same.

This process often involves adjusting some lead page aesthetics, opt-in forms, headlines, featured images or CTAs to a new one.

The ultimate goal here is to re-ignite visitors’ interest by offering them a strong incentive to click again on the CTA or finalise existing process.

Having considered the importance of running AB tests in online marketing, let us now proceed to see how you can set-up A and B test experiments using Google Analytics.

To Implement New Marketing Ideas.

Needless to say, the internet marketing industry is dynamic and always subject to new innovations, ideas and algorithm changes.

This shows that the best strategy that delivers optimal performance today might fail to achieve desired result tomorrow.

Hence, there is the need for marketers to upgrade their marketing tactics to accommodate new digital marketing innovations.

So, AB testing the old and new marketing practices will help decide whether to jettison the old marketing strategies or proceed to implement the new ones.

How Does AB Tests Work in Online Marketing?

Just like statisticians engage the use of two-sample hypothesis to check if there are any significant statistical difference between two independently obtained data, so also is the AB testing.

It is often engaged to help marketers make better marketing decisions based on the data obtained from two sample tests, usually termed A and B.

Let’s say we two landing pages named; LP A (control) and LP B (variation).

The former is targeted to PC users only while the latter variant pops up on mobile devices.

Taking a glance look at the test outcomes; it is obvious that LP B generated optimal conversion rate (37%) compared to LP A that captured only 23% of leads.

Usually, when a growth optimizer sets up AB tests, the two experiments could see some variations in the CTA button, headline, snippets, images, lead magnets or demographics used.

The ultimate objective of this AB testing strategy is to see what works best between the two versions.

Once this optimal conversion rate has been obtained between A and B variants, the most effective one will be used when making subsequent marketing decisions.

However, to build a solid foundation for your AB split-tests, you need to have a blueprint that defines your AB variants testing processes.

Follow these AB test guidelines to run an effective AB testing process.

Gather Relevant Data.

Let’s assume that you’ve installed Google analytics on your CMS before now (importantly you should anyway), just check-in to your analytics dashboard.

Therein you can obtain the raw metrics that have been collected by Google based on the tracked activities and visitors interaction with your website.

You should know that one of Google analytics stored data displays a traffic overview of pages on your website driving the highest number of views or traffic.

This is where you can acquire insights into the exact page(s) that is likely to deliver the best AB test results.

The reason for this is quite logical, isn’t it?

Absolutely, right?

Yea, areas on your website with high number of visits are more likely to generate high conversion rates compared to pages that get few visits.

Pages with high-traffic could be one that features a cornerstone content or one that visitors find more helpful.

So, once you have filtered the page with low traffic metrics on your website (you can do this conveniently using the ‘filter’ feature on your CSV excel sheet), proceed to set-up your AB test goals.

Set-Up AB Test Goals.

Setting up goals for your A and B tests is important because it helps you to effectively measure the performance of your tests.

Hence, having a clear AB tests conversion goals is important in determining which variable record more success.

But, how do you design AB test goals?

Well, you can create AB split test goals based on a particular course of action you want a visitor to take while engaging with your website.

It could be anything like clicking a CTA button, email subscription, newsletter sign-up or buying a digital product.

Generate AB Test Hypothesis

You should start gathering hypothesis once you've established your conversion goal.

To do this, you should brainstorm on possible AB test differentials that are likely to generate optimal change in your conversion indices.

For example, here are 3 simple AB test hypotheses that could be formulated;

·        Thumbnail images will improve sales funnel CTRs.

·        Discount offers on leadpages will improve conversions.

·        Red CTA button will increase Clicks and Pageviews.

Develop Your Split Test Variations.

Proceed to develop AB test variations by making deliberate changes to selected elements on your leadpages, sales funnel or mobile apps.

This could also involve you changing some custom aesthetics on these selected elements.

It could be the CTA button colour, headline fonts, page navigation or images.

Likewise, you can opt to use an AB testing software which offers a custom editor that will assist you to seamlessly make these changes.

Once you have done that, it’s time to split-test it against your pre-determined AB control experiment.

Another thing you should fix at this junction is to select your AB tests duration.

That is, the time frame you’ll want this test to run.

So, fixing AB tests scope could cover a 7 days period, 30 days period or even more, depending on your prerogative.

Set-Up Your Bucket-test experiments on Google Analytics.

Having pre-set your test variations, you should set it up on your Google analytics account.

Doing this will help you to accurately monitor your AB test outcomes.

As you continue reading this piece, you will discover how to set-up AB tests on Google analytics.

Analyze Results

Once your test experiment gets completed within the time allotted, you can proceed to obtain the AB test feedback.

You can obtain this from your Google analytics dashboard or consider deploying A/B testing software.

From the metrics presented from using either of these AB testing tools, you can then check the significant statistical differences between any of these two test buckets.

Normally, the variation or B test should generate better outcome than the controlled experiment-A.

If this AB test result holds, then it is successful and if otherwise, you have to adjust your AB tests.

What Do Online Marketers AB Test?

For many inbound marketers, applying AB tests is most desirable because it helps to maximize their web conversions.

So, there are many aspects of a website that can undergo AB test experiments.

These includes;

·        Landing page.

·        Headings.

·        Visuals and Images.

·        Software and Apps.

·        Links.

·        Testimonials.

·        Social Proofs.

·        Call-to-Action text copy.

·        Call-to-Action button.

·        Sub-headings.

·        Paragraph Texts.

·        PPC Campaigns.

·        Social Media Ads.

·        Leadpages.

·        E-mails.

How to Do AB Testing on Google Analytics Experiments.

Here, let’s see how you can set-up Google Analytics AB testing.

Follow this sequence to implement your split-tests for A and B.

Go to your Google Analytics Webmasters Tool.

Punch into your search browser Google analytics URL link where you will be directed to enter details.

Normally, many folks login to Google analytics account using their Gmail login details.

Whichever way you decide to go, just get yourself into the Google analytics dashboard. 

Click On The Behaviour Icon

Once you get into your analytics dashboard, locate the behaviour icon at the left-hand side of the Google Analytics user interface.

Click on the Behaviour option to view the drop-down menu where you can locate the ‘Experiments’ option.

Click on Create Experiment.

Ok now, let’s perform some experiment abracadabra.

Choose a name for your experiment, it could be anything that comes to your mind.

But I no you wouldn’t like to go weird in naming your AB test experiment anyways.A

Likewise, select the Objective that you want the experiment to achieve.

The options are;

·        Bounces

·        Pageviews

·        Session Duration

Finally, set the traffic range for your split tests.

Select Your Preferred Experiment Metric Options.

This Google analytics AB testing phase is where you are expected to input the page URL link of both A and B.

Usually, the Original page which also serve as the control page is normally termed A, while the Variant 1 equally serve as the experiment page and is termed B.

Check out Advanced Options.

The advance option is needed in situations whereby you are running multivariate experiments.

That is, your variance is more than one.

This will help a webmaster to evenly distribute traffic across all the B experiments.

Set-Up Your Experiment Code

Grab the experiment code displayed in the box provided and paste it into the HTML head of the pages that are set to undergo AB testing.Spli

Review and Start Your AB Bucket Tests.

This is the final step.

It requires you to double-check all the experiment options you have chosen.

In case you aren’t satisfied with any of the pre-set options, you can edit it to meet your goal.

Once this is done, it’s time to start experimenting A and B tests.

Conclusion.

For a webmaster to constantly generate high conversions, low bounce-rates and increase page views; applying AB techniques is very key.

The reason is because the digital marketing landscape is ever-evolving and new marketing ideas are constantly being birth.

Which by implication demands that a marketer test-run these new ideas against existing ones before opting to overhaul or adjust a performing marketing strategy.

So, if you haven’t applied A and B tests to your product page, ecommerce or online store website, just know that you have been surcharging yourself.

But why continue to ignore a marketing strategy that is proven to make you increase your online scalability?

C’mon, grab your Google Analytics and get down to implement all that you have learn’t about AB testing in this article.

Lastly, if you’ve got some clarifications or questions to ask, feel free to punch it in the comment box provided below.

I’ll be glad to answer your comments and queries.

Cheers!