Creating and Starting Your Experiment
Create a hypothesis and set a goal
Forming a hypothesis and setting a goal gives your experiment direction and provides a frame for evaluating the results. After forming your hypothesis, you need to set your goal in the tool. Do that by logging into your analytics account and clicking “ADMIN”. From there, click on the account and property that you want to modify, and in the third column click on “GOALS”. You’ll be taken to a page where you can click on “CREATE A GOAL”. If you aren’t picky, you can just use a predefined goal like “made an online payment”, “signed up for a service”, or “made an appointment”. You can even set up a custom one. Finally, you add in the URL of the page that the goal applies to into the destination box and you are set to go. Once you’ve clicked “CREATE GOAL” you should see it added to your list.
Prepare your variations
It’s worth noting that you do actually need to prepare your variations. With Umbraco, you create new page for each variation and tweak the content in the direction you think will help you meet your goal. The changes you make should be reflective of your hypothesis and goal (E.G, if your goal is to increase newsletter signups, you should change your call to action).
Creating your experiment
Now that you’ve made an assumption about what will improve your site, created a direction, and made the tweaks, it’s time to bring the experiment to life. Here’s how:
- Navigate to the experiment creation screen – it all starts with navigating to the experiment creation screen. Once into your account, just click on “REPORTING” and in the left hand column choose “EXPERIMENTS” from the “BEHAVIOUR” category, and the interface to create your experiment will be right in front of you.
- Enter the URL of the page- on the experiment creation screen, the first thing you want to do is to enter the URL of the page that you want to test (the original). It could be any kind of page–a blog post, a product landing page, your site’s home page, etc.
- Name the experiment- you might have several different experiments going at one time, so it’s a good idea to give yours a descriptive name. How about “Landing page CTA change”.
- Choose your objective- since you’ve already set up your goal in Google Analytics, now is the time to choose it from the dropdown menu. Or, if you want to get down to business ASAP, you can select a predefined one.
- Select your traffic %- it’s a good idea to include 100% of your traffic in the experiment; however, one caveat is that if you’re running a big and risky experiment, then you might want to scale it down a bit. That way, if it’s broken or buggy, you won’t lose out on a great deal of potential revenue. Directing 50% your hard-earned traffic to a broken checkout page would be a shame.
- Configure the variations- enter the URLs of the variation pages you wish to evaluate into the corresponding boxes. If you want more than one variation just click on “ADD VARIATION” and add it in.
- Deploy the experiment code- once you’re done with those steps, you are brought to a screen where you are given the choice to get your code and insert it manually or have it sent to a web master. If you’re good enough at coding, then you can just insert the code into the right place inside your Umbraco templates. Make sure to follow GA’s instructions! Most opt for a developer to help them with this. This part is tricky but luckily there are plugins that will do this on your behalf, more on that later.