Create Decision Matrices for Rating
Learning Objectives
After completing this unit, you’ll be able to:
- Create a decision matrix.
- Explain how to import a CSV file into a decision matrix version.
- Identify common errors.
- Prepare a decision matrix version for use.
Create a Decision Matrix
With all the pricing data carefully reformatted, Solution Architect Justus is ready to build the decision matrices that powers rating procedures at Cumulus. Thanks to the initial dating formatting, creating and configuring a new decision matrix is quick and easy.
To start, navigate to Lookup Tables and choose New. For lookup table type, choose Decision Matrix. Next, give it a name and select the record type. There are two record types available: Standard and Grouped. For more detailed information about each type, check out Decision Matrices.
Justus starts by creating a new standard decision matrix named ageFactors. This matrix takes a customer’s age as input and outputs a rating factor that directly impacts the final premium. Due to the slightly higher risk associated with older individuals, insurance premiums often incorporate age. In this matrix, the age factor increases as the customer’s age increases.
Once he creates the matrix, he automatically sees a version named ageFactors V1. Each decision matrix can have multiple decision matrix versions, but the first version is created automatically with the name [matrix name] + " V1".
You can add more versions if needed. Matrix versioning is a powerful way to automatically set up future pricing changes. For example, consider a Cumulus insurance plan where certain rating tables need adjustment each year to reflect updated risk data. Instead of creating entirely separate matrices and expression sets, Justus simply creates two versions of the affected matrices. The expression set can then dynamically reference the correct version based on the date.
Import CSV Files
A brand-new matrix has no table data yet. You can add the rows and columns manually or automatically by importing the table data. Importing is the more common and efficient option, especially when dealing with tables with a lot of data.
The rating table is saved as a CSV file. You upload the CSV file to the lookup table to create a decision matrix.
Here’s the rating table for Cumulus’s simple age factor.
age |
ageFactor |
|---|---|
18 |
0.9 |
30 |
1.7 |
50 |
2.8 |
During import, define each column header with the following information.
-
Header Type: Specify whether the column is an input or output.
-
Data Type: Select from options such as number range, text, percent, currency, or Boolean.
For this age factor table, the column headers should be:
-
age: input, Number Range
-
ageFactor: output, Number
After the CSV file upload is complete, the table is visible, and the file is listed on the Related tab of the decision matrix version.
Troubleshoot Your Matrix
What happens if you have text in a number column? What if you specify all input columns and no output columns? You get errors.
To keep your matrices error-free, remember these essentials.
- The table must have at least one input and one output.
- Each row must have a unique combination of input values.
- Data in columns must correspond to the specified data type. For example, text doesn’t belong in a number column.
Some errors relate to the data in the table. The system analyzes your uploaded file for errors. If it detects problems, it creates an error file with messages, such as warnings about duplicate rows or mismatched data types, placed within the context of your CSV data.
Column headings represent rating variable names and often match product specs. Watch out for typos in those heading names. Those mistakes can show up much later in testing and are typically more challenging to identify.
Prepare a Decision Matrix Version
You found an issue during troubleshooting. Now, here are some ways to fix that issue.
-
Edit columns: Delete columns or adjust column headers, header types, or data types. Remember to save any changes you make.
-
Upload CSV file: Upload the original file with a repaired CSV file.
-
Remove files: Delete files related to the decision matrix. For example, if you uploaded the wrong CSV file, you can delete it.
To make future changes to the matrix, create a new version. This new version preserves your existing configuration, so you can modify the columns and upload updated data as needed.
Final Steps
With his data in order, Justus has successfully built his decision matrix. Excellent! Can he immediately use it in an expression set and get output? Not quite yet. The final step, which is easy to forget, is to enable the matrix. Edit the version, set the desired date range and rank, and select Enabled. When enabled, the matrix is in use, and the data can’t be changed.
Well done. You now know how to import CSV files, define column headers, and troubleshoot errors. The matrices are enabled and ready to power premium calculations.
In the next unit, explore how to create expression sets to determine the final pricing for insurance products.