NB: Generally naming conventions and documentation references to product name use Boxever rather than CDP/Personalize. This is partly due to the script reference using this domain but also the fact that it encompasses both CDP/Personlize. As described elsewhere in documentation we have used Boxever to personalise a component as an A/B test. We have also used Boxever to personalise the home page banner for people who view the "/Become-an-mvp" page. NNB: In the time we were not able to generate enough data for reporting but this would naturally be generated by Boxever being implented sitewide. We could also fire a "PERSONALIZATION" event so as to see the impact of this personalisation.
Decision Model Considering the time our decision model was based on the following example adjusted to our scenario: https://github.com/boxever/configurations/blob/master/Decision%20Model/Blueprints/Most%20Viewed%20Page/PageViews.xml The adjusted model can be seen in the following files: ~\docs\boxever\v1.2.xml ~\docs\boxever\PageViewDecisionModel.jpg The image shows a decision model which take two inputs:
- Number of days to include - the number of days to check
- Page Name - the page which triggers personalisation if visited Lines 2 and 3 (in PageViewDecisionModel.jpg at time of hackathon) define variables which take the above inputs Line 20 (in PageViewDecisionModel.jpg at time of hackathon) shows where the Guest's sessions page view events are compared to the page, iterating the pageCount variable if evaluating to true
Experiences An example of FreeMarker syntax can be found at the following link https://doc.sitecore.com/cdp/en/users/sitecore-personalize/compose-the-api-response-for-an-interactive-experiment.html Another reference on Freemarker https://freemarker.apache.org/docs/ref_builtins_hash.html The output from the Decision Model above can be seen in the image below. ~\docs\boxever\DecisionModelResponseConsumedByFullStackExperience.JPG The FreeMarker which consumes the DecisionModel above is in the image below ~\docs\boxever\FreeMarkerForFullStackExperienceAPIRespone.JPG
Experiments We used a Boxever Experiment to create an A/B test. The image below shows the allocation of traffic to our two responses MultiVariantTest.jpg We then made a simple API response as can be seen in the image below to indicate whether or not to show some content. The test being whether or not the content showing had a positive impact on our goal. We set a goal of visting the Application page.