Onboarding with us is a simple and efficient process and in this article we will walk you through everything you need to consider when first getting up and running.
For each section we give a brief description of what it is and why its needed and we have attached the relevant help guide for reference.
First things first
We need to consider where we will be testing and what sorts of things do we want to track as part of monitoring and measuring the results of those tests. In other words are there certain metrics/goals that will help understand the performance of a test and what are they. Once we know this we can decide where to put the tag to enable both testing and metric/goal capture.
Tagging your site
If you follow the guide below correctly you will be hardcoding the tag into the head of the pages of your site and this is an ideal implementation. In some cases it may not be possible to get this hardcoded and it may be you use other methods such as tag managers to get started. This is fine and will work however it has some limitations such as how fast the test can load which can then introduce content flicker.
The help guide for tag implementation can be found here.
All the tag configuration options
There are lots of things you can configure within Webtrends Optimize and they all depend on your specific requirements. In some cases none of the below is needed but in some cases it may all be needed.
Consent management
Integrations
Conversion package
Page masking
Single page application handling
Consent management
Consent management refers to how we work alongside the consent management you have in place, sometimes for similar analytics tools. In the majority of cases we will need to be considered as a tool which tracks users behavior and therefore consent management becomes a consideration. We have different ways of integrating with these mechanisms and because of the nature of what Webtrends Optimize does it should be something you consider from day one. Here are some options:
Behind consent management
In the case where we have to sit behind consent management (where a user must click accept cookies for us to be shown) we will have a situation where the test cannot run (and the new page be shown) until cookies are accepted meaning users to the site will see 2 versions of the page. The first (original or control) when they land there and the second (the variation) when they accept.
β
Please find a doc relating to this here.No consent needed
In some cases, even though we track user behavior, we are considered essential and therefore we do not need to consider consent management.
Advanced method of consent management
Consent management refers to the tracking of user behavior and only tracking users who accept the policy. Because running a test consists of 2 key elements, the test itself and the tracking, we have the ability to break out the key elements into 2 workstreams. Essentially we can always run the test, even if cookies have not been accepted, but we only track a user once they accept the cookie policy. In this example a user never sees both versions of the page and we still comply with the necessity for a user to accept the policy.
β
Please find a doc relating to this here.
Integrations
Integrations can be turned on and off from the integrations screen accessible from the navigation. Some integrations require you to select a tag which the integration will work with, in most cases you can select 'All' here.
Integrations can be turned on and off at any time and all the help relating to integrations can be found here.
Conversion package/Global metrics
When we talk about a conversion package we are referring to goals or success metrics which are set up in Webtrends Optimize and are tracked as part of every project you run. This enables you to create global metrics and avoids you having to create and build these for every project. The conversions could be page based (visits to a page such as a checkout) or could be things such as events you want us to capture from data layers (maybe purchases or add to baskets). We have different ways of capturing this information and help can be found below. It is worth noting that if you plan to use the visual editor only, and only for small tests, there can be an argument that you do not necessarily need to create and build an entire conversion package as these can be added ad-hoc.
Full details on what is required can be found here. Please be aware, this is typically a task for clients with access to Javascript developers.
A version of the conversion package script can be found here.
Page masking
Page masking refers to how we can hide sections of a page to cover where changes (transformations) are being made and tests are being run, ultimately removing content flicker. For example if you want to test a hero banner on a page its likely you want to hide the original banner while Webtrends populates it with the new variation. Masking allows you to enable this and it is a key component in removing any content flicker.
β
Full page masking
This is where we are masking (hiding) the entire page until changes (transformation) have been made. You can specify how long this should happen and which pages if needed.
Partial page masking
We are able to hide sections of a page, or all pages. In lots of cases we will hide the majority of the page but leave the main navigation showing to show a user the page is loading. Partial masking can be on a single page, multiple pages or all pages.
Help for page masking can be found here.
Single page apps
Single page apps require a little bit more work as part of the configuration. A usual test relies on each page loading separately so we can assess whether a test should be loaded and run or whether there are conversions which should fire. For single page apps we do not have these page loads and therefore need to find a way to define when a page load has happened.
The key thing here is to understand if you will be using the visual editor, in this scenario we need to do some work to make this happen automatically. Alternatively we can provide hooks which will tell you, or your developers, when the page has changed and the decision can be made on your side. Each option has pros and cons but its all down to which bits of the tool you want to use and what is available on the site for us to use.
This will all be handled in the onboarding but for more information there are 2 articles which can explain this in a bit more detail.
Working with react can be found here.
Working with Gatsby can be found here.
Other considerations
CSP's
Content security policies are a modern security feature, which tell the browser which domains are allowed to operate on your website, and what permissions they're allowed.
We need to allow Wevbtrends services to run on the site and all the information on what's needed and how to do it can be found here.
Next steps
Setting up the first test
Ideally we want to set up an AA test to be able to validate all the configuration we have undertaken. An AA test assigns users into 2 variations, although no changes are made, and replicates a real life test. In this scenario everything will work as it would do for a true test but risk is removed by us not making changes to the page.
The AA test can validate everything from conversions that have been set up, integrations you have enabled and any masking/consent management. Its a perfect way to ensure everything is ready to go and it is a good place to start learning how to build tests and analyse the results.
Setting up an AA test is done easily through the advanced editor where you will only select a location to run the test on (we suggest sitewide) and there is no other configuration to consider.
You can find more information on setting up and running AA tests here.
What else should I check/consider?
Once everything above is complete and you feel happy its worth considering the following before committing to running any real tests you have lined up and ready:
Are all of my conversions tracking?
Are all of my integrations working?
Was the page masking working as expected?
Have I managed to use the force experiment tool?
Do I have what I need to build my test?
Now you have configured, checked and validated everything needed to be up and running with Webtrends Optimize.