Name: Creating Predictive Scores and Models in Predictive Studio in SAP Marketing Cloud
Uploaded: 2021-7-12
Duration: 405 s
Description: In this video, you'll learn how to create, train, and publish Predictive Models in Predictive Studio in SAP Marketing Cloud. Predictive models use algorithms and historical data to calculate scores that give you insights about your customers' future behavior.

Share feedback

<p><u><strong>Planned maintenance of Practice Systems</strong></u></p>

Planned maintenance of Practice Systems

Why create an SAP account?

Enhance your learning experience

SAP Learning

Choose your subscription

Manage your subscription

Live Sessions

Hands-on Practice

Certification

My Learning

Ready to start learning?

<div><div>In this video, you'll learn how to create, train, and publish Predictive Models in Predictive Studio in SAP Marketing Cloud. Predictive models use algorithms and historical data to calculate scores that give you insights about your customers' future behavior.</div></div>

Creating Predictive Scores and Models in Predictive Studio in SAP Marketing Cloud

Hi, my name is Lily. In this video, I'll discuss Predictive Studio in SAP Marketing Cloud. I'll show you how to create, train and publish predictive models. For this, the ship consumer buying propensity score will be used. With Predictive Studio. You can create predictive models that use algorithms and historical data to calculate scores that give you insights about your customers future behavior. You can create, maintain, train, and publish predictive models in the public cloud. A predictive model is based on a predictive scenario and one or more model fits a model fit is the training or the predictive model on a selected data set. A predictive scenario consists of one or more data sources and assign implementation methods that are tailored to your case. For consumer buying propensity, the business use case is to calculate how likely new customers are to buy one of your products. The data sources consist of data from customers who have recently bought the product. This is the auto ML, or APL automated Predictive Library on SAP Hana, which runs on the Hana database. Models that use a statistical implementation method such as SAP Predictive Analytics need to be trained. The training set you use is a target group, and the historical data of the members. Predictive modeling in SAP Marketing Cloud has a very simple user interface in order to identify and publish best model fits easily high speed performance is delivered, scores can be calculated in real time, or they can be calculated as a scheduled job and stored on the database. Optimization of business processes is done at various places like contact profile or contact engagement. Currently, the only supporting method available in SAP Marketing Cloud is SAP Predictive Analytics, also known as APL Automated Predictive Library. For you a business analyst, it means that you can easily run and manage predictive models directly in the Marketing Cloud application without the involvement of developers or data scientists. Now, I'll show you how easily new predictive models can be created. The goal is to create a model based on a consumer buying propensity for the product or Mr. Barista. The output score will indicate how likely it is that a consumer will buy the specified product. Let's open the Predictive Studio application and click on the create button placed on the upper right side. I'll name the predictive model and optionally add a description to identify it. Easier. In the future, I'll select the scenario from the dropdown. I'll select Consumer Buying Propensity from the Predictive Scenario list. This scenario is part of the standard shipment of SAP Marketing Cloud. It's based on interaction data and is included in the segmentation profile. Our consumers. Then I'll select the marketing area I wish to use from the dropdown, and I'll click Create to Save. Next I'll select the target object from the dropdown. This is an object such as a product or product category for which you want to make a prediction. I'll specified time frame for analysis. Then I'll choose a training set. The training set should contain enough relevant historical information to train your predictive model. The default target variable for consumer buying propensity is purchased since you're looking for customers who bought the product. The number of responses in the training set will be calculated using the target object and the target variable, and will appear once you save. In this case, the number of responses is the number of buyers of the selected product or target object in the past. You need to ensure that you have a certain number of responders or buyers in the training set. The number of members shows how many members your selected target group has overall. This will appear once you save. Click save. Now we can create the first model fit. I'll click create, add a name and optionally a description. The only available implementation method for predictive model training in SAP Marketing Cloud is SAP Predictive Analytics, also known as APL. I'll click create. Now I'll select all those predictors which I think are relevant for a good predictive model. Once a set of predictors is selected, the model can be trained. I'll click the trained button. Once the algorithm is finished, those contributing predictors which have an impact on the prediction appear with percentages expressing the contribution strength for a next iteration of model training. Single predictors can be removed or added from a repository. Once you have trained the model set, you'll see the training results. Predictive power indicates the quality of the model fit. The value between 0 and 1 corresponds to the proportion of information in the target variable. That can be determined by the predictors, which contribute to the current model. Fit prediction confidence indicates the robustness of the current model set when applied to a new data with the same predictors as in the training set, the estimation is expressed as a value between 0 and 1. The higher, the better initial number of predictors equals the number of entries in the list of contributing predictors. Number of selected predictors indicates the number of predictors you have selected for training of the current model. Fit number of Cat predictors indicates the number of predictors that contribute to the prediction. When training the model set. SAP Predictive Analytics checks whether the selected predictors contribute to the prediction and keeps only the contributing predictors. The list of contributing predictors displays all predictors that are considered in the current model training. And that's how easy it is to create predictive models. Now I'll go back to the predictive model detail screen to activate the model fit. Once the model set is activated, the whole predictive model changes status to being active and you can use it in segmentation. For more information on the Predictive Studio, go to the SAP Marketing Cloud SAP Help portal. Thanks for watching.