This section describes how to implement Adobe target V2 for AB testing through Adobe Launch. Adobe Target has V1 and V2 versions. V1 will be abandoned, so we will talk about V2 here, and there are two methods for deploying Adobe Target. One is It is a hard-coded form. One is through Adobe Launch. Here we use the Adobe Launch form.
Before the official start, you need to have the permissions of Adobe Launch and Adobe Target to achieve the following operations.You also need to have server permissions because you need to add some code to the source code.
Configure Adobe Target on Adobe Launch
The above information is generated by default, you do not need to modify it.
- Client Code is your company name.
- Organization Id is the company ID, which corresponds to the Client Code.
- The Server Domain was set when the company name was determined.
Need to pay attention to this paragraph in the head:deploying Launch asynchronously, we recommend hard-coding the following pre-hiding snippet immediately before the asynchronous embed codes in order to manage flicker. For more details, see:Adobe Target extension with an asynchronous deployment documentation section.
This is a piece of code that prevents the page from flickering. It needs to be deployed in front of the Adobe Launch embed code in the form of a hard-coded page.
Just click Copy prehiding snippet to Clipboard! To copy this code，we need to go to the server and add this code:
It should be noted that this code is js, but what you copied directly is not <script> </ script>, you need to add it yourself.
Extensions can then be saved.
Create a Rule named Adobe Target:
Create an EVENTS that fires when the page loads. The specific configuration is as follows:
Event Type: Library Loaded (Page Top)
Name: Core-Library Loaded(Page Top)
Create an ACTIONS, select the type of Load Target, this role is to initialize the mbox:
Then save it.
Behind the first one. Create a second ACTIONS, select the type of Add Params to Page Load Request, this role is to configure some defined data:
Then save it.
Behind the second one. Create a third ACTIONS, select the type of Fire Page Load Request, this role triggers sending data：
Then save it.
The final configuration effect is as follows:
Note that there must be at least Load Target and Page Load Request in ACTION, with Load Target first and Page Load Request last. You can add an Add Params to Page Load Request in the middle to send custom data.
Publish and Test Adobe Target
After configuration, we need to test the configuration, first we need to publish a library, and then use Adobe Experience Cloud Debugger to test:
You can see Adobe Target v2, which means the configuration is fine.
Setting up Adobe Target
It should be noted here that Adobe Target Premium only has Preferences and Properties. If it is Adobe Target Classic, you do not need to set these.
Set your account preferences to configure Target Standard or Target Premium to work correctly with your account. The following illustration shows the available settings on the Account Preferences page.
These four places are mainly set up:
- Custom Global Mbox is set to target-global-mbox, When there are no activities, it is empty by default, but when there is activity, it needs to be set to target-global-mbox. Remember that when the first ACTIONS was set earlier, this is to initialize target-global-mbox.
- Results and Reporting：Set options that determine what data is used for your results and reports.
- Mobile Viewport Configuration is to configure a mobile environment.（This is only available in Premium.）
Configure your preferences as desired, then click Save .
Property functions like a report suite, indicating where data is sent and which activity it belongs to.
Create an at_property, there are two deployment methods, one is hard-coded, directly the same as the previous anti-flicker, and one is to Add Params to Page Load Request in Rules:
Data needs to be set to send to the Adobe target normally. If it is not set, there will be an error of Unauthorized mbox host. Unable to load experimental resources.
Create an Activity
Return to Activities, click Create Activity on the right, select A / B Test
Then see the following interface，Looking at the middle position of the head, Adobe Target divides the process of setting the AB test into three stages：Experiences、Targeting、Goals & Settings:
This step mainly sets up different versions of the experiment.
The first is to modify the name of the activity in the upper left corner. Move the mouse to the vicinity to see the rename, and then modify:
The second part is to add Properties, located here:
Select the corresponding Properties:
Finally, select the editing mode, in Page Delivery:
There are two editing modes, one is Visual Experience Composer (VEC), one is Enhanced Composer, and the default is to use Enhanced Composer. Generally, VEC is recommended. This is interactive visual editing. It is more convenient, so we need to use Enhanced Composer shut down:
Click anywhere to edit it:
Finally, I changed Contact me to author and added a background color:
This step of Experience is complete, click Next:
This part is mainly the algorithm for setting the audience and distribution of the experiment:
- Audience is who participated in this experiment.
- The algorithm is TRAFFIC ALLOCATION METHOD
Goals & Settings:
This step is mainly the setting of the experimental goal, as well as the setting of other experimental time, description and other information.
Testing and publishing Activity
Save after setting.After saving, you will come to this page. This page displays relevant information about this experiment. We need to click Activity QA:
Copy the URL and open it in a browser to preview the effect of our experimental version.
The last step is to publish the experiment:
Our experiment is published, now we need to check if the data of the page is passed normally.
Open the page and find the mbox cookie in the browser tool. If the value of this cookie is long, like the following, it means that there is an experiment running, no experiment is running, only the sessionid, and the data is short:
Under the network, there is a delivery request. This is a request to find the adobe target. You can see the relevant data for the experimental settings.
Adobe Target V2 corresponds to delivery; Adobe Target V1 corresponds to json.
At this point, our basic experiments are set up.
Click the corresponding Activity directly, and then click Report next to Overview. If you set it to Analytics, then return to Adobe Analytics to see the report:
The experimental group has increased by 49.62%, and 76.58% is certain that this adjustment is improved.
The confidence of an experience or offer represents the probability that the lift of the associated experience/offer over the control experience/offer is “real” (not caused by random chance). Typically, 95% is the recommended level of confidence for the lift to be considered significant.
Obviously there is not enough certainty that the improvement is due to adjustment.