Goal Conversion Rates
Conversion rates are simply the percentage of the number of people who take action on an offer (i.e. signups, sales, etc) compared with the number of people who are shown the offer (i.e. the signup page, the product page, etc).
On the goals page, is simply the number of times each goal has been started (i.e. how many people were shown an offer), and how many times it has been completed (i.e. how many people converted).
The overall conversion rate is simply the number of people shown offers divided by the number of people who converted to those offers.
Why are my goal conversion rates slightly different from my test conversion rates?
Sometimes your goal conversion data may be slightly different from the test completion data.
ZenTester relies on a small javascript code being installed on your site, to display tests and send information back to the ZenTester servers about what tests are being displayed, and whether a user converts.
In order to work correctly, the ZenTester code needs 4 required actions to happen:
- It needs to be installed correctly;
- The user visiting your site must have javascript enabled;
- The page that the code appears on must load completely, and the code must connect to the ZenTester server;
- The page that is being tested must remain the same over time.
Here's how to find out which one of these 4 actions may be causing the difference in conversion data that you see, and what to do about it...
Making Sure the ZenTester Code is Installed Correctly
If the ZenTester code is not installed correctly - on the page that you are testing, as well as on the conversion goal page - you will typically see 0 impressions and/or 0 conversions listed.
For instructions on how to install the ZenTester tracking code,
or check it's installed correctly, visit:
http://www.zentester.com/training/videos/installing-the-zentester-t...
Users Visiting Your Site Must Have Javascript Enabled
If users do not have javascript enabled, they will be unable to see ZenTester split tests you are running on your website, and they will not be counted towards your conversions.
This will mean ZenTester lists both fewer impressions and fewer conversions. How many fewer? Well, it's estimated that between 0-3% of internet users do not have javascript enabled.
Unfortunately, there is no easy solution to this.
Page Loading and Code Connecting to ZenTester Server
Before a test (or a conversion) can be tracked, ZenTester's main server needs to receive a message from the ZenTester code installed on your site.
This message transmitted to ZenTester's server records whether all the tests displayed on the page is displayed correctly (i.e. is this a valid test impression?) and whether there has been a conversion goal met on that page (i.e. has there been a conversion?) - the two most important pieces of information for calculating conversion rate data for split tests.
If the page that the ZenTester code sits on does not load completely, the ZenTester code will not be able to check that everything on the page appears correctly (to validate your split test is displayed correctly) and then record what was displayed on the page.
And if ZenTester is not able to record what was displayed on the page you're split testing (e.g. your signup page), it will not record a conversion on your conversion goal page (e.g. the signup thanks page) because ZenTester is unable to record both the start and end of a test.
This can result in big differences in the conversion rate data recorded in ZenTester.
Typically, the two reasons why this may occur are:
- A user converts too quickly before a page has time to load;
- A page loads too slowly, and does not load before a user converts.
If you have stats on the time users spend on the page, look at these stats to see whether are converting within seconds of being on a page.
Consider reducing the file sizes of images, CSS and javascript loaded on your webpage; perhaps upgrade to a faster web hosting server; or clear your internet browser's cache and look at what elements on the page may be taking time to load on some users' computers.
The Page Being Tested Changes
As mentioned above, the ZenTester code validates that all test variables are displayed on a page correctly before it records a valid test beginning.
It's important that the page you're testing remains identical (if not, largely the same) from the time you begin a test, to the time you end a test. Adding or removing sections of the page will mean an invalid test - which is why ZenTester's code checks that the page loads correctly before reporting impression or conversion data.
If you've set up a Multivariate test inside ZenTester, and the page that the multivariate test appears on is later changed, the ZenTester code may think the page is not valid - and therefore may stop counting impression (and conversions) that relate to the altered page.
Typically, if this happens, your conversion rate data will be valid for a while - and then will suddenly stop matching your other stats when your page was changed.
The simple solution to this is to either revert your page back to how it previously looked - or set up a new Multivariate test inside ZenTester for the new version of the page.