How to split test your landing pages using Google Optimize

If you’re running a paid search campaign with Google Ads, you probably already consider a great conversion rate to be your one and only holy grail.

After all, if you’re placing a lot of resources into your account, you want to make sure that your investment is actually generating interest in your products and services (and, if you’re an ecommerce company, direct sales). In other words, you want your visitors to complete an action that’s valuable to your business – not just have a look around and leave!

We’ve talked about the importance of optimising your campaigns for conversions in many of our previous posts. We’ve covered how you can use better onsite design and UX practices to maximise opportunities, and we’ve suggested ways you can improve your ad copy to skyrocket your click-through rate.

But we haven’t yet mentioned the many benefits of split testing your landing pages within Google Ads.

What is split testing – and why is it important?

Split testing – sometimes referred to as AB or A/B testing – involves creating two versions of a web page; using both of them equally and at random within the same campaign; and monitoring which one generates the most conversions.

AB testing allows you to easily compare the performance of your original control page (the page without any amendments) with the performance of a page that contains slightly different elements.

To help you understand exactly which of your design and copy elements is proving to be more appealing to your audience, be sure to change only one or two parts of the variation design, then run the split test for a set amount of time to monitor the user response.

After several rounds of testing ‘experiments’, you should be left with a page that is well-optimised in terms of its copy and design, and that generates significantly more conversions than the original control version.

What is a ‘good’ conversion rate?

Looking at all data from all sectors, the average conversion rate for a website is 2.5%. This means that, for every 1,000 people who have visited your website, you can expect 25 of them to complete your desired action.

That said, paid search campaigns tend to fare a little better than advertising via other web-based channels. Recent data from Google Ads suggests that the standard conversion rate across its search network is 4.40%, which isn’t surprising, as users are presented with ads after using a particular search query, so are likely to already be in a research or buying mindset. By implementing best practice techniques – including split testing landing pages – our team regularly achieves conversion rates for our clients of 10% or more.

Here are some basic split testing ideas to inspire your landing page experiments:

Adjust the position of your call to action (CTA) button. Why? Because you might get more clicks or sign-ups if it is more prominently visible and located above the fold.

Tweak the copy on this same CTA button. Even making subtle changes to your language and tone could encourage a better response. If you’re stuck for ideas, Wishpond has put together a brilliant blog packed full of call to action messages that are proven to work.

Highlight the CTA button with an accent colour. Clever use of a contrasting shade will draw the eye to the most important part of the page and encourage higher conversions.

Change your page headline. There is some evidence that suggests titles that contain a key benefit of the product or service you’re trying to sell will bring you more conversions – but the key is to experiment with the wording and see what resonates with your audience.

Test long form copy vs short form copy. You may find that replacing wording with icons, images and infographics increases understanding and makes it easier for the user to digest your information, which in turn will lead to more take-ups.

Swap in-depth product or service information with client testimonials. Potential customers may respond better to social proof than wordy explanations.

Switch copy for video, or vice versa. Your audience could prefer a different medium.

Try bespoke graphics and illustrations in place of stock photos. Branded photography that’s unique to your business will create a stronger connection with potential customers.

Why use Google Optimize?

Instead of implementing all these experimental changes manually, you can use a tool like Google Optimize to manage your landing page changes and monitor which amendments have the most noticeable impact on your page performance.

Google Optimize is an online split-testing tool that enables you to experiment with different conversion rate optimisation (CRO) methods, then collect data on the results from each test. It’s designed to work seamlessly with Google Ads and Firebase, Google’s mobile app development platform – and you can link it up to your Google Analytics account so that all reporting data is fed straight back into a familiar system (and goals data is merged together for a complete picture of your conversion activity).

For those of you who are more familiar with Google’s wider range of products, Optimize also integrates well with BigQuery and Accelerated Mobile Pages (AMP).

The pros and cons of Google’s own split testing software

Optimize’s free plan allows you to:

Create and adjust your Google Ads landing pages straight from the Optimize dashboard

Edit raw HTML, JavaScript and visual code without any prior coding knowledge

Access previews so you can see your changes in situ before your page variation goes live

Schedule changes in advance

Carry out multivariate, split URL and server-side testing, in addition to more straightforward AB split tests

This plan is perfect for smaller businesses who are just starting to get to grips with split testing. But there are limits to the number of experiments you can run at the same time, and you can only assign three goals to each test.

If you want access to even more valuable insights, you need to sign up for Optimize 360, which offers sophisticated testing functionality, plus full multi-variate testing features and experimentation capabilities.

Yes, you will need to pay a monthly fee for these more advanced tools – but use it to its full potential, and the tool will pay for itself several times over!

How to run a successful split test using Google Optimize

Follow these this simple step-by-step guide and you’ll soon be comparing data from your split test experiments.

Set up an account

Head over to to get started.

If you are completely new to Google products, you’ll need to create an account from scratch. You’ll need to enter a username, set up containers for each Analytics property using the domain as the name field, and accept all the relevant terms and conditions

If you already have Google Analytics logins, use these to access Optimize. You can also get to the Optimize interface from Analytics by clicking on the Experiments tab in the Behaviour section.

Choose your experience

Click on ‘Let’s Go’ to start the process. Name your variant page, then enter the URL of your control page (otherwise referred to as your editor page). From here, decide whether you want to run a:

AB test

This is the standard test used to compare results from one or more variant landing pages. Multivariate test

This approach will test variants or two or more elements at the same time to find out which combination delivers the best results. Choose to run a multivariate test if you really want to drill down into which element is making your page a success.

Redirect test

This is a kind of AB test that allows you to split test different web pages instead of similar web pages that have simply been adjusted. Pick this option if you want to compare two totally different designs.

Link your Google Optimize account with your Google Analytics account

Hook up your Google Analytics property to its corresponding container (ie, the correct web domain) to make sure information can flow seamlessly and accurately between Analytics and Optimize.

(Make sure Google Analytics is installed correctly on your website, otherwise you’ll fall at the first hurdle).

Add Optimize to your website

Now it’s time to add a small snippet of code to your website that enables Optimize to track and store user information. To do this, you’ll need to be able to access the HTML code in the back end of your platform. You’ll also need to remove any old or outdated Optimize code before setting up these new snippets.

You can add the necessary code to each web page that you want to test by using the optimize.js snippet or, alternatively, Google Tag Manager. For full instructions, and a list of the pros and cons of each method, click here.

Create your control and variation landing pages

It’s time to get creative! You can use the editor palette to make changes to your variant page(s). Click on ‘Edit element’ to make changes to each element’s HTML, text, or JavaScript components, and control where it’s located, how big it is and how it looks.

If you prefer, you can edit your variation landing page using the CSS code editor, instead.

Check your work in Preview mode

Before you hit the live button, check that your page looks as it should in Preview mode. Everything here will be displayed exactly as your users will see it.

If you want to be extra thorough, download the Chrome browser extension for Optimize and use it to access preview modes for tablet and mobile devices.

Run or schedule your test

The most efficient way to manage your tests is to schedule them in advance. Doing this enables you to set a pre-confirmed start and end date for each experiment, let the tool work its magic, then check back in once the test has come to an end.

However, bear in mind that once you have scheduled an experiment, you won’t be able to end it early – you’ll only be able to edit or cancel it. You can monitor a live experiment by clicking the Reporting tab at the top of the Experiment Detail page.

Google recommends running any experiment for at least two weeks, as this should give you enough data to allow for any fluctuations in traffic. Similarly, if, at the end of the experiment, you don’t have enough data to conclude that at least one of your variation pages is 95% likely to deliver a better conversion rate than the original, keep it running for a little bit longer.

You may also want to end any inactive experiments to make sure you have enough slots in your quota to meet your testing requirements.

Once an experiment has finished, you can archive it or copy it for your own records.

Check your reports

The reports in Google Optimize enable you to analyse every aspect of your experiments, including the number of sessions they received, the dates when these sessions occurred, and how each of your variants has performed according to its objectives.

The report summary will tell you which of your pages is the ‘leader’ in terms of its probability to be best (PBB) – be it the original control page, or a variant – and the software will also tell you if it encountered any problems running your experiments due to incorrect tagging, rigid targeting rules or a lack of real data.

If you want to, you can also view your Google Optimize reports directly via Google Analytics. For more information on the wealth of information generated by Google Optimize, and the various terminology the tool uses to assess the effectiveness of your experiments, visit the Optimize Resource Hub.

When is it time to scale up?

Google Optimize’s free plan is designed to suit the needs of small to medium sized businesses that want to run a small number of tests at any one time, with limited elements.

If your company requires more experiments and more advanced features in order to make the most of the tool, you could benefit from setting up a paid account. Only do this if you’re serious about investing in CRO, though, as costs can easily start to escalate.

Other split testing tools are available

Google Optimize is by no means the only split testing tool available to data-hungry webmasters. Alternatives such as Optimizely, Convert, Unbounce and PageSense can also be used to implement landing page changes and monitor conversion results.

There’s no harm in testing each of these programmes to find one that you feel comfortable using and that offers your desired functionality.

If you need help, we’re only ever at the end of the phone.

At Multi Layer Media, we recognise that setting up and managing split test experiments can be time-consuming – especially if you have never used Google Optimize or a similar AB testing tool before.

If you’d rather outsource this vital CRO work to the experts, contact our team today.


Are you convinced that your Google Ads campaign isn’t converting as well as it could? Ask our expert PPC managers to carry out a FREE review of your account and they will quickly identify the areas that could be improved for a better ROI.

Posted By
Aaron Suleyman