Based on my years of experience both in house and with an agency, I have come to the conclusion that there are three reasons why a large percentage of digital marketeers shy away from CRO:
- Cost – Visual Website Optimizer (VWO), Crazy Egg, Unbounce, Optimizely all have monthly costs, although they do start off fairly cheap. VWO for example starts at just £38 a month, so it’s not going to break the bank but is still another cost to add to the 400 tools you are already paying for.
- Effort – Setting it up, speaking to developers, code in the wrong place, code not found – you get the picture. Changing elements of a website without liaising with the developers can make CRO confusing. From personal experience “don’t know how to set it up” and “I’ll sort it, but later on” are common excuses, even among experienced marketeers.
- It’s Hard – I’ll let you into a little secret. CRO doesn’t have to involve months of planning and huge teams of data crunching specialists to extract top secret information, which can make or break of your website. You should always be able to say to your boss or client “let’s test a couple of versions of that page we have seen drop in performance” or “sure, let’s split test that new landing page for a few months” . Of course if you are working on a huge site, or for a multinational company, you’ll need a greater level of detail compared to working on a tiny ecommerce website.
This is where Google Optimize comes in, this CRO tool is free and can be set up in minutes. It is also incredibly intuitive and makes A/B and multivariate testing a complete breeze. If you are working in-house, you will be able to amaze your boss. If you are at an agency, get ready to blow your clients away with stunning in-depth reports to help blast conversion rates through the roof.
In this post, I am going to show you how to:
- Set up Google Optimize.
- Create your first experiment (A/B and Multivariate).
- Track your performance.
- Export your winning report.
By the end of this post, you will be able to go from dreaming about running CRO campaigns to running multiple experiments, testing hundreds of different elements and working towards improving your conversion rates in the long term.
How to set up Google Optimize
First of all, you will need to have Google Analytics (GA) set up for your website. As you are reading a post about CRO, I am going to assume you already have this in place. Don’t worry if you don’t, Hallam has you covered.
Once you have GA set up, you will need to head into your account and navigate over to the admin tab, then click on tracking info > tracking code and you will see the Website Tracking box with your GA code in there, like this:
Copy this code and put it somewhere safe, as you will need this later. I usually stick it in a notepad and save it to my incredibly organised desktop. Now you have your GA code ready to edit, you will need to set up your Google Optimize account.
Head over to the Google Optimize homepage and click the ‘Sign Up For Free‘ button to get started. Sign in with your Google Account and you will see the account home page. From here :
- Click the ‘Create Account‘ button
- Choose your company name
- Agree to all the Google spiel
- Acknowledge that you have read and agree to the T&C’s (once you have read them of course)
- Add your container name (this is your website URL, leave out http/https)
You should now be looking at the following screen:
Everything you need to do to get up and running is set out in five easy steps on the right hand side. Now, this is might sound confusing – jump straight to ③ ‘Link to Google Analytics‘. As long as you are logged into the same Google Account that your analytics is assigned to, you will be good to go. If not, you can request access from the Admin tab in GA.
Click Link Property, choose the correct property and you will see the ③ will now be highlighted in green, indicating that you have successfully linked the two accounts and are ready to install the snippet.
Installing the Google Optimize Snippet
Expand step ④ on the right hand side of the Container Information panel (on the right hand side), click on View Snippet and you will see the following window:
Now is the time to get your GA tracking code loaded back up again. We are going to need to insert a small piece of code into the GA tracking in between ga(‘create and ga(‘send. Copy the line of code in step 2, then paste this below the attribute that starts with ga(‘create, as seen in the example above.
Keep your updated GA code to one side and click Next. You will see the following window:
This snippet of code ensures you only see your variant content and minimises page flicker to ensure your CRO experiments don’t affect the performance of your website. This snippet of code is required to go before your updated GA tracking code. Grab your updated GA tracking code and add the page flicker snippet above. The new code will look something like this:
Now all that’s left to do it upload the amended GA tracking code (including the Google Optimize snippet and page flicker code), paste this code as high in the <head> of every page as possible and you are good to go.
One thing to note is that installing the Google Optimize snippet is a self verified process, so once you have added it to your pages you will not get a green ✓ like you would see in GA. However, a quick way of checking you have installed the code correctly is to create a test experiment and look within the edit feature.
Starting Your First Google Optimize Experiment
Head to the Google Optimize dashboard and click the blue Create Experiment button. You will see the following window:
Enter the name of your experiment and make this easily identifiable, as you may have multiple experiments running at certain times. Stick in the URL of the page you wish to test (editor page). Now comes the exciting part – choosing the type of experiment you want to run:
1. A/B Test
Optimizely define A/B testing as “A method of comparing two versions of a webpage or app against each other to determine which one performs better.” This is sometimes referred to as split testing or bucket testing. Let’s say you own a small international business consultancy and aren’t sure you are portraying a corporate image on your homepage – you could implement the following experiment:
Current homepage (A) – Use large image of your staff in the main banner with very personable language, such as “We are a great team who will help you transform your business”. You could use very soft calls to action and add a list of awards you have won.
New Variant (B) – Use large corporate stock image in the main banner and messaging that gets straight to the point like “Award winning corporate consults, take your business to the next level”. You could use multiple hard calls to action (eg. book a free consultation now) and a list of multinational companies that you have worked with on projects.
You could run an A/B test to show a certain portion of your traffic one page (A) and the other portion the other (B). This will allow you to put together a case for whether to change the language and imagery on your site, rather than simply taking a guess.
2. Multivariate Test
Multivariate testing is similar to A/B testing. VWO define the difference as follows: “In a multivariate test, you identify a few key areas/sections of a page and then create variations for those sections specifically (as opposed to creating variations of whole page in an A/B split test)” Say you own a ecommerce store selling t-shirts and want to see the effect of using different headlines and images on your main category page. You could set up the following experiment:
Variant 1 – Headline 1 & Image 1
Variant 2 – Headline 1 & Image 2
Variant 3 – Headline 2 & Image 1
Variant 4 – Headline 2 & Image 2
Google Optimize will then split the traffic between these variations for a set time period (defined by you – the longer the better!) and deliver you a report of the best performing combination.
3. Redirect Test
This experiment is used to split traffic between two different URLs. Say you are a health and fitness provider and have a BMI calculator on your website that requires the user to fill out their contact information before being given access to the tool. You could get the developer to remove the step where you request information and move this to the end of the process, so the user can find our their BMI data without commitment. You would then create a redirect test, splitting the traffic between the two URLs for example:
URL 1 – www.healthandfitnesscompany.co.uk/bmi-calculator
URL 2- www.healthandfitnesscompany.co.uk/bmi-calculator-2
Ensure you add a no-index tag on your new variation URL to ensure it is not indexed by search engines. Be sure to also keep it out of your sitemap, just to be safe – nobody wants duplicate content.
So, you now have Google Optimize set up and are ready to get your very first CRO experiment rolled out.
So, How Does It Actually Work?
This is more relevant for A/B and Multivariate Testing as Redirect Tests do not require any editing within the Google Optimize platform. Firstly, create a quick A/B test experiment that we can use for this run through . Here is one I made earlier:
Click Create and you will be directed to the experiment details page, here you will see the Original variant above a blue + New Variant button. Click this and name it something relevant to the changes you wish to make to the page.
You will see the new variant appear under the Original. Click this variant and you will be directed to the editing page – like this:
Now this is where it gets interesting. You can edit almost all features on the page with a single click. Say I want to change the headline to “We are the best digital marketing agency on planet earth” all that I need to do is right click on the headline, change the text and save. You don’t even need to worry about fonts, Google sorts that for you:
Want to change an image? Right click, select Edit HTML and change the link to your new image:
It really is that simple. Each change you make is logged in the header of the editor, so you can go through them one by one and add/remove as necessary. The magic thing about this free software is that you don’t have to speak to a single developer (🙈) you can literally click, replace, delete and add new items.
Once you have completed all the changes on your variant page, head back to the experiment overview. Click on Targeting and you will see the following:
The top two options relate to who you are going to be serving your experiment to. The percentage is relative to your site traffic, so if you wanted to run the test on every single website visitor keep this at 100%. If you want to target a quarter of your visitors, drop this to 25%. Similarly, the weighting determines the split of users who will see one page or the other, 50/50 is the middle ground. For whatever reason you may want to serve one variant more than the other – simply change the percentages and you will be good to go.
You will also need to determine when you want to serve your experiment. This example is set to fire when the URL matches the home page but there are many other options:
Don’t fret, there are lots of great targeting features available on the free version of Google Optimize like behaviour, technology, URL sequences and much more. You will have more than enough to use. You don’t need to invest in GA 360, it’s just not worth it unless you are dealing with big data.
Now you have everything set up, chosen who to target and when to target them – head to the main dashboard and start your experiment:
How Do You Track Your Experiments?
Linking GA and Optimize allows both platforms to pass information between each other, so your experiments will be waiting for you in Google Analytics. Head to Behaviour > Experiments, choose your experiment and you will be able to look at the performance on a day by day basis:
You don’t have to configure any settings, your experiments will automatically appear in GA and you will have a log of all past experiments, so you can keep track of all of your CRO activity in one place.
What Do the Google Optimize Reports Look Like?
You have two choices with Google Optimize:
- Run your CRO experiment for a certain amount of time based on your own objectives
- Leave your CRO experiment running indefinitely until Google tells you which variant is the winner
When it comes down to getting the best outcome, the more data the better. Websites receive different volumes of traffic. If you only receive a few hundred sessions per month don’t be disheartened. You can still use CRO to improve your site performance but it will just take a litter longer to get the definitive result you are looking for. If your sessions are in the hundreds of thousands, your CRO experiments will deliver results quickly and efficiently. Either way – CRO is for almost everyone (10 sessions a month? seek SEO help).
Whichever way you decide to go, ensure you have a decent enough sample size to make an informed decision. By waiting it out you will be delivered with some pretty cool reports:
Google Optimize will deliver you a clear report showing you exactly how each variant performed, along with a solid recommendation to help you make your decision. If you have set an experiment live and it really hasn’t offered any insights, I recommend setting a cut off date and then making the decision yourself based on your GA Experiment report.
What Are You Waiting For?
CRO is getting a whole lot easier. “I don’t have the time/money to implement CRO” is becoming a tired excuse. Thanks to Google Optimize you can now run CRO experiments quickly and easily.
One thing this post has not touched on is CRO strategy. Rather than testing for the sake of testing, I would highly recommend bringing in a digital agency/CRO expert to help give you direction and ensure you benefit from your experiments.
To conclude, if CRO has been pushed to the bottom of your digital pile, never made it into your 12 month strategy, or you simply chose to ignore it – now is the time to take action. You will be able to wow your clients, make your boss smile and take the first step to truly optimizing your website.