Using 1 Goal for Multiple Pages
I just got off the phone with an Optimizely engineer to discuss the results of a test we are running as i wanted to confirm how goals are attributed to a test.
It turns out that it's not what i expected and would skew all the tests i have ever ran.
1. Lets assume i have a custom event goal set up for a form on every page on site.
2. I set up a test targeted on simple match for 1 specific URL. The goal mentioned above is activated for the test
3. A user enters the targeted URL and is included in the test and a cookie is dropped. They are now entered into that variation. The user does not fill out the form on that page and therefore a goal is not activated.
4. Later on the same user comes back to the site and does not visit the page that the test is running on. They then submit a form that is using the same goal running in the test.
5. This goal is attributed back to the test. Even though the user did not convert on the page that the test is running on.
I was completely unaware that this was how Optimizely worked and would see this as invalid data for my tests.
The only way i can avoid this is to create unique goals with the tags added on only thge pages i wish to run the test on. This requires a huge amount of work throughout the year implementing tags, creating goals and testing that they are working for each and every test that is run.
Can anyone confirm that this is correct and if there is an easy way to get around this issue.
@DaveOC - Goals and Experiments are different things.
A user participates in an experiment.
A user performs actions on a website which may be goals.
The results tab shows the number of users who are in the experiment who recorded those goals.
If you have multiple forms on your website and have a goal called "form submit" that is recorded on every form, then a user could be in an experiment and could submit a form on an entirely different page and that goal woudl be reflected in the experiment's goals.
This is absolutely intentional! Consider that you are changing the "Add To Bag" button on your product page and one of your goals is "Orders". Since your experiment does not include the Thank You page, the goal *must* record independently from the experiment.
If you want to know how many people submit the form on the page you are modifying specifically for this experiment, then you must set up a click-goal that is specific to that page's form.
I completely understand that there are many instances where this method is useful.
For most of the experiments i run unfortunately is not.
For example lets say i run an experiment on 10 product pages to test a different form. If the user is placed in the variation 1 bucket then ends up submitting a form on a page not in the experiment, then the original variation will receive the conversion. Even though the page they converted on shows the original form.
Would be great if there was an option for conversion events to only be valid if they occur on a page that is set in the URL targeting of that experiment.