Oracle® Social CRM Applications Administration Guide > Administering Oracle Sales Prospector >

Generating the Prediction Model


After you have imported your data into Oracle Sales Prospector, you can generate a prediction model for your sales representatives to generate their sales and product predictions. Data model building is the science behind making recommendations. From your data, Oracle Sales Prospector evaluates the kind of products your customers have purchased, when they purchased them, and finds a pattern within the purchases to make a prediction or recommendation.

Whenever you make changes to your data, with the exception of target customer by user data (user-to-customer account mappings), you are required to regenerate the prediction model. When you add or change user-to-customer account data, you affect only the visibility of the prospects, but not the prospects themselves. The following are examples of when you must regenerate the prediction model:

  • You import new customer records.

    No predictions exist for these records until you rebuild the prediction model.

  • You import new product records.

    No predictions exist for these records until you rebuild the prediction model.

For each attempt to build the prediction model, Oracle Sales Prospector sends you an email indicating success or failure, with a log file attached.

NOTE:  Complete this procedure after you have successfully imported your customer data into Oracle Sales Prospector.

To generate the prediction model

  1. Log in to Oracle Sales Prospector as a registered data manager.
  2. Click Manage Data at the top of Oracle Sales Prospector.
  3. In the Manage Data window, click Build Model.

    Oracle Sales Prospector uses Oracle Data Mining to build sales predictions by identifying gaps in the selling patterns in a customer's order history when compared with the customer's peers. Peers are identified using the attributes of customers.

    Oracle Sales Prospector typically takes 15 minutes to generate the prediction model, however, your results might vary. To maximize performance when generating a prediction model, close any nonessential applications that are running on the Oracle Sales Prospector server. After Oracle Sales Prospector generates the sales prediction model, it sends you an email to notify you that it has completed the prediction model generation. If Oracle Sales Prospector encounters an error while attempting to generate the prediction model, it sends you an email with the details of the error.

  4. To view prospects after you have successfully built the prediction model, refresh your view by clicking Home.

    (Optional) If Oracle Sales Prospector encounters an error condition, see Troubleshooting the Prediction Model Generation for information on how to resolve the problem.

Oracle® Social CRM Applications Administration Guide Copyright © 2009, Oracle. All rights reserved.