In this article, I will show you to launch an A/B test experiment via Google Optimize.
1. Overview
1.1 Tools and Materials
- Blog article on Blogger.com
1.2 Workflow
1.3 Hypothesis
In the implementation, I built a blog on Blogger.com which is a blog-publishing service that allows multi-user blogs with time-stamped entries and I want to improve the search volume of it on the Internet. Before creating your first experiment you need to identify a problem, then create a hypothesis about what you can change to improve it.
To do so, I need to clarify a feature that can be added to my site to improve exposure on the search engine. In this example, I hypothesize that adding alternative text for figure element can improve the rank of the blog in the search list. Next, I will show how to design an A/B test by google optimize to test this hypothesis.
2. Blogger Article Preparation
Blogger
is a blog-publishing service that allows multi-user blogs with time-stamped
entries. It was developed by Pyra Labs, which was bought by Google in 2003. The
blogs are hosted by Google and generally accessed from a subdomain of
blogspot.com. Blogs can also be served from a custom domain owned by the user
(like www.example.com) by using DNS facilities to direct a domain to Google's
servers. A user can have up to 100 blogs per account. Next, I will briefly introduce
the process of creating a blog.
1. Go to the
Blogger.com and click either create your blog or sign in. at the top right-hand corner of the page.
If you already have an account, then, you can sign in directly. If not, you will be prompted to sign in using your Google account or other accounts. I recommend you to use your Google accounts.
2. Build your first blog by clicking click the Down arrow icon at the top left-hand corner of page and New blog.
Then, choose a name for your blog and click NEXT.
Give a
URL for your blog and click SAVE.
3. Manage your blog.
4. Create posts for your blog by clicking click “+NEW POST”. Then, you can go into the editor of posts to write your articles. The content can be forms of text, image, video, link.
5. To save your post, click the Down arrow icon at the top right-hand corner of page and New blog, next to Preview.
6. You can also publish your blog by clicking publish shown as the above figure.
As of now, we have been done with our website (Blogger). Next, we need to link our blog to Google Analytics, i.e., installing Google Analytics on your Blogger.com blog.
3. Google Analytics
Refer to my another
post.
4. Google Optimize
By now, we have installed Google Analytics on our website through which your GA property has been an agent of your blog website so that GA will have the data of your website. Therefore, you can monitor and analyze your website now. Next, we will introduce Google Optimize to facilitate you’re A/B test.
Google Optimize is a platform that allows users to create personalized experiences and run website tests, such as A/B test, redirect tests, and multivariate tests.
4.1 Sign Up for Google Optimize
1. Go to the Google Optimize web page.
2. Sign in using your Google account email as Google Optimize and Blogger.
3. Sign up for an Optimize account:
- If you don’t have a Google Optimize account:
A. Go to https://marketingplatform.google.com/about/optimize/.
B. Click on Start for Free.
C. Click “Create Account” and give your account a name. This can either be your company’s name, or your own name. Accept the terms of service agreement for your country. Click next.
D. Set up the container’s name as your website’s name to keep things simple, click Create.
- If you do have a Google Optimize
account:
A. Go
to https://marketingplatform.google.com/about/optimize/.
B. Click
on Sign in to Optimize and go to the Dashboard.
C. Click
“+” button to create a new container. As the previous case’s step D, set up the container’s name as your website’s name to keep things simple, click Create.
4.2 Link to Google Analytics
Having signed up to Google Optimize, to have your Optimize account (container) fully set up, you need to link it to Google Analytics. Linking Optimize and Analytics allows you to run experiments in Optimize and access data about running experiments in Analytics. Once you’ve linked your Optimize container to an Analytics property, you’ll be able to start experiments, target experiments to Analytics Audiences (Optimize 360 customers only), and use Analytics goals and metrics as experiment objectives.
4.3 Install Google Optimize on Your Website
Refer to my another
post.
4.4 Create an Experiment on Google Optimize
Next, it’s time to create your first experiment on Optimize.
1. Go to your Optimize Account (Main menu > Accounts).
2. Click on your Container name to get to the Experiments page.
3. Click the “Let’s go” button, enter an Experiment name (up to 255 characters) and the URL of the page you want to test, and select the type of experiment you want to run (A/B test). Click Create.
Note: Use a fully resolved URL in the editor page. Redirects aren't supported here.
4.
Create a variant.
A. To get started, click Create variant to alter that page for your first variation if it requires. In order to do this, you’ll be prompted to install Google’s Optimize extension for Chrome. The Chrome extension is required to use the visual editor and installation diagnostics, but not to create redirect experiments or view reports.
Once you complete the installation process, the page you want to test will open with Google Optimize’s visual editor.
You can check whether the Chrome extension has been installed or not at your Optimize Account> Container>Settings
B. Click Add variant (blue button), enter a variant name, then click Finished.
a) click Edit button in the variant row (which will say “0 changes”). This will launch the Optimize visual editor – an overlay on top of your editor page consisting of two components: the app bar (at the top of the page) and the editor panel (floating in the lower right).
b) Click on any web page element you wish to edit (e.g. a button). Use the editor panel to make a change (e.g. change the button color).
In my case, I will make changes based on my hypothesis, i.e., adding alternative text for all of the figures in my blog.
D. Optional: multiple pages in an experiment
In some situations, you may want to include more than one webpage in one experiment. Fortunately, Google Optimize provides such a function for users to add multiple pages. To do this,
a) Scroll down the experiment edit page to find button Add page and click on it.
b)
Give the new page a name and enter its URL. Then click save.
c)
Edit the invariants of those pages.
5. Set up weight distribution over original page and the variant(s).
The weight is the probability of which a visitor who is included in your experiment will see the variant. By default, all variants are weighted equally in Optimize. In other words, a visitor who is included in your experiment has an equal chance of seeing any of your variants. If you want to direct more, less, or even all of your traffic to a specific variant, you can adjust your variant weights on the experiment details page.
Click the weight and then you can select custom percentages option to adjust the weights. Here I keep the even splitting setup. By the way, when you want to do A/A test, you can set 100 percent for one specific variant.
6. Set up the Audience targeting. You can either keep the default setup by which All visitors to the page(s) targeted above or customize the audience group counted in your experiment.
Basically, customizing your targeting audience can help your experiment focus on a specific group of visitors such as visitors led from Google ads, visitors from mobile device, and visitors from a geopolitical city of interest. To customize your audience,
A. Find Audience targeting at the Targeting and variants section and click Customize.
B. Select one type. For example, if I want to focus on visitors from mobile device. Then I need to select Device category option.
C. Then, you can choose the condition for filtering your experiment subjects.
D. More details at https://support.google.com/optimize/answer/6283435?hl=en and https://support.google.com/optimize/answer/6283420.
7. Set your objectives for the test.
Then, you’ll need to set your objectives for the test. Scroll down to the Measurement and objectives section and select “Add Experiment Objective.” Optimize users can use up to three pre-selected objectives per experiment and see data for those objectives in Optimize reports.
A. Select a Google Analytics view.
Note: This step will help Optimize introduce goals (objectives) defined at the selected Google Analytics view if you have done this.
B. For the PRIMARY OBJECTIVE,
a) click Add objective.
b) You can either choose one from a list (also including goals defined/created in Google Analytics view. See https://support.google.com/analytics/answer/1032415?hl=en#goal_type ) or create a custom one.
c) Here I just choose pageviews from the list as a primary objective of my experiment which is a typical objective. Recall that I have mentioned that the goals defined in your selected Google Analytics view at step a)) will also be shown in this list. Since I did not create any goal in my Google Analytics view, there is no goal listed here.
d) You can also add at most two additional objectives (you can have at most ten if you are Optimize 360 user) at ADDITIONAL OBJECTIVES.
C. For more details about how to create an objectives, you may refer to https://support.google.com/optimize/answer/7018998 and https://support.google.com/analytics/answer/1033068.
8. You may check whether we have correctly installed Optimize at section Settings by clicking Check installation:
9. By now, we are almost done with the main setup of the experiment. Start your experiment.
A. You can set the experiment duration by clicking the clock button at the top of the experiment editor.
B. An editor is popped up. Schedule your experiment duration and click Done.
C. Click start to launch this experiment.
10. For more details about how to create an A/B test, you may refer to https://support.google.com/optimize/answer/6211930.
4.5 Analyze Experiment Results
Once your test has gone through the time duration, you can navigate to the Reporting tab in Optimize to see your results. Your results will be broken down into “cards” with data about how your variants performed.
For more details about the elements in the reporting page, you may refer to https://support.google.com/optimize/answer/6218117.
helpful
ReplyDelete