In today's post we bring you a tool that we are starting to work with and that promises to be of great help in advising our customers about what changes to make on their website. Many of you will already know that we are talking about Google Optimize, but for those of you who don't know it yet, I'll give you a brief summary of what it is:
It is a web site optimization tool that helps to improve the overall satisfaction of the users, thanks to the fact that continuous tests can be carried out with different aspects of the web site. Experiments are created based on the hypotheses we have previously formulated and we can check which version of the website gets better results. Sounds good, doesn't it? Let's see what we need to put all this into action:
- Account Google Analytics with its respective tracking code implemented on our website.
- For the experiment setup we need to have installed Google Chrome and its Google Optimize extension for this browser.
- Its own tracking code for Google Optimizewhich we can obtain directly from the tool once we have created our account.
Once these simple requirements are met, we are ready to start using the tool. These are the next steps we will have to take:
- Go to optimize.google.com and click on Create an account.
- Enter a name for the account and, within the account, name the container in which you will create your first experiment.
- Accepts the Terms of Service Optimize and click on create.
It is time to link it with Analytics. For this purpose:
- Go to section 3 and click on link with property:
- Select property Analytics that corresponds to the web domain on which you are going to run the experiments and click on guard.
Now let's take a look at what to install and where for the tool to execute the changes we want to reflect on the web:
- Go to the Experiments y localizes the panel Information of the container, located on the right side of the page.
- Click on Install the Optimize fragment and then on the blue button view excerpt. That fragment will have to be installed right after the code we have on the Analytics website.
- It is also advisable to implement the anti-flicker fragment, which will temporarily hide the page while the Optimize container is loaded. We will install this code just before the Analytics code.
Et voilà, we have everything ready to start setting up our experiment. There are 4 types of experiments:
- A/B testing: We show users two or more variants of the same page.
- Multivariate: This type of test consists of showing the user different variants of two or more sections on the same page.
- Redirect test: This option is used to test independent web pages, which are identified by different urls or paths. As a reminder, we must enter a "noindex" to the test page.
- Personalization: Allows you to customize the page for targeted visitors.
Once we have decided which option best suits the changes we want to make, we can begin to shape our hypothesis:
- We access the interface of our experiment and click on add variant. Optimize distributes the traffic to 50% between the two alternatives, although we can modify it.
- Click on edit and the web site referenced above will be opened to make the changes that we consider appropriate.
- The editor interface is very easy to use and very intuitive, we can try to make simple changes like this one:
- In the section Segmentation we will be able to decide to whom and when we are going to show the variant we have just created.
- In order to launch our experiment, we will choose one of the objectives that we have configured in Analytics, through which we want to measure the results of our variant.
And that's it! Now we just have to let both versions collect data and, when we have enough information, we can decide if our initial hypothesis was correct or not and if it is worth making the changes permanently on our website.
We hope you found this post interesting! If so, stay tuned, in a few weeks we will share with you the data we have collected with our experiments and our conclusions 🙂