Testing Rating Management

Impact testing for Rating Management

You can use impact testing to see the impact that changing the rate book and rate routines have on policy premium for a group of policies. Impact testing generates test policy periods and rates them using the active rate books and selected comparison rate books. You can choose the test policy periods by product, jurisdiction, producer code, postal code, effective date, expiration date, and as of date.

In the base configuration, impact testing rates the baseline and test policy periods for all job types as submission jobs. You can configure impacting testing to rate a renewal as a renewal job rather than a submission job. For example, rate routines may include special steps for renewal jobs. Rate tables may contain factors based on job type.

In the base configuration, PolicyCenter prevents a renewal on a policy period which has a open renewal. If you make this configuration change, PolicyCenter can create a renewal on a policy period which has an open renewal. Making this configuration change impacts how PolicyCenter handles renewals in impact testing and in other areas of the product.

For information on configuring impact testing to rate renewals as renewal policy transactions, see Rate renewals as renewal jobs in impact testing.

Impact testing creates two graphs. The X-axis is divided into impact ranges. Each range represents a percentage change to the policy premium, such as No change or 0% up to 5%. In the Policies Affected graph, each bar represents the number of policies affected in each impact range. In the Financial Impact graph, each bar represents the monetary amount of change in each impact range.

The impact testing graphs only display changes to policy premium. The graphs exclude other costs, such as taxes, fees, and surcharges. However, you can configure impact testing to display other types of costs by modifying the RatingExportUtil Gosu class. In this class, change the generateIAStatistics method.

You can export the coverage and cost comparisons to Excel. For each policy, the Excel spreadsheet shows details for each cost on the policy when rated using the active rate book and the comparison rate book. The details include:

Baseline term amount
Baseline actual amount
Comparison term amount
Comparison actual amount

General guidelines for testability of Rating Management

This topic provides some guidelines to improve the testability of your Rating Management implementation.

Exclude `EffDated` objects from functions

Automated unit testing of code can be complicated if testing requires a policy. Creating a policy presents significant overhead to test creation and execution. Therefore, avoid passing effective dated objects to business logic where possible. When a function operates on atomic values, then unit tests can be small and simple.

Include data capture mechanisms

Test result or test execution data provides a useful way verify quality. PolicyCenter and the Guidewire platform provides data analysis tools and logging, but you can add additional logic to measure and verify the implementation.

Internal profiling

The Guidewire Profiler provides detailed information regarding which components consume time while processing a request. In addition, you can configure custom entry points. This data is particularly useful for profiling database interactions.

Performance measurement

If the base implementation does not support your performance data capture needs, you can configure data capture code at appropriate points. Results can be written to a log file, stored in the application database, or both. Since this approach is limited in scope, it can be used to capture and analyze production behavior without excess overhead.

Verifying inputs and outputs

PolicyCenter provides several ways to capture data inputs and outputs for test verification.

Impact testing enables you to capture large amounts of policy and rating information for analysis.
Rating worksheets display rating algorithm calculations on a policy. Generating worksheets may result in slower rating performance. Consider adding a way to turn worksheets on or off based on user input or specific policy conditions.

Through configuration, you can also develop other data capture mechanisms. Identify these needs early their implementation so that they can be built into the design.

Working with impact testing in Rating Management

You can use impact testing to see the impact that changing the rate book has on policy premium for a group of policies. Impact testing generates test policy periods and rates them using the active rate books and selected comparison rate books. You can choose the test policy periods by product, jurisdiction, producer code, postal code, effective date, expiration date, and in force on date.

Permissions for impact testing

To view the Administration > Rating > Impact Testing screen, you must have the Rate policies for rate impact testing permission.

The code for this permission is rateimpacttesting. In the base configuration, the Rating Analyst role has this permission.

If a user has this permission, impact testing analyzes and provides access to all policy periods that match the search criteria. With this permission, the user can access policy periods for which the user has insufficient permissions.

Impact testing warnings and recommendations

Impact testing is designed to work in a test environment on a copy of production data. Impact testing is accessible only when the server is in development or test mode. Because impact testing affects system performance and creates test policy periods that persist in the database, impact testing is not accessible in production mode.

When you run impact testing, impact testing creates test policy periods in the database. In the test environment, observe the following warnings and recommendations:

The work of users can interfere with impact testing results. During impact testing, Guidewire recommends that only the single user performing impact testing be logged into PolicyCenter.
Disable integrations to other systems. If you run impact testing with integrations enabled, the integration can send test data to these production systems. For example, disable the free-text search integration that uses the search engine Solr. The same applies to other integrations such as an integration with a billing system. To avoid unnecessary costs, disable integrations to systems that charge access fees.
While running impact testing batch processes, do not make changes in PolicyCenter that impact rating such as doing a policy change, canceling, or changing rate books. For example, the baseline creation batch process or the test period quote batch process is running. During this time, quoting test periods fails if a user changes or cancels a policy which has a baseline.
Do not include expired policies in the impact testing dataset. Specify an In Force On date to filter out expired policies.
Impact testing excludes archived policies.
Stop the Job Expire batch process to prevent PolicyCenter from unexpectedly expiring policies during or between impact testing runs. To manage batch processes on a test server, press ALT + SHIFT + T to display the Server Tools page, then select Batch Process Info from the Server Tools tab.

Generate test periods in impact testing

About this task

You can use impact testing to generate test policy periods rated using the current rate books and rate books that you select.

Note: Impact testing has several steps that may require a long time to complete. For these time-consuming steps, impact testing uses batch processes. While PolicyCenter is generating results, the user can navigate away from impact testing screens. The user can return later to view results and continue a previously started test. The following instructions include steps for continuing a previously started test.

Procedure

In PolicyCenter, navigate to Administration > Rating > Impact Testing.
PolicyCenter displays the Choose Policies screen.

Do one of the following:

To continue a previously started test, click Next with Previous Test Case to advance the wizard. Go to step 6.
Specify criteria to filter policies for testing. None of these fields is required.

Field	Description
Products	Choose one or more products.
Jurisdictions	Choose one or more jurisdictions.
Producer Codes	Choose one or more producer codes.
Postal Codes	Choose one of the following: Specify codes – One or more postal codes separated by commas. Range – Specify the minimum and maximum for a range of postal codes. The minimum and maximum are inclusive. PolicyCenter searches for a match that starts with the minimum and maximum that you enter.
Effective Date	The policy effective date must fall within this range.
Expiration Date	The policy expiration date must fall within this range.
In force on	Policies must be in force on this date to be included in the search results.

Click Search to display the list of policies that meet the search criteria.
Repeat specifying criteria in step 2 and search in step 3 until you are satisfied with the search results.
Click Next with Search Results. Alternatively, you can choose Next with Previous Test Case to continue and use the previous search results.
PolicyCenter displays the Create Baseline screen.
Depending upon the button that you clicked in step 5, do one of the following:
- Next with Search Results – Click Create Baseline to generate baseline policy periods on the policies in the search results. The policy periods are rated using the current rate books.
  For all policy transactions, impact testing rates the baseline policy periods as submission policy transactions. For information on configuring impact testing to rate renewals as renewal policy transactions, see Rate renewals as renewal jobs in impact testing.
  When you click Create Baseline, PolicyCenter runs the Impact Testing Test Case Preparation batch process. Since this batch process may take a long time to run, you may exit impact testing, and go to other PolicyCenter screens. After the Impact Testing Test Case Preparation batch process finishes creating baselines, the Create Baseline screen displays the baseline policy periods for each policy in the search results.
  After the batch process completes, the Next button appears.
- Next with Previous Test Case – If you are continuing a previously started test, the baselines are already created and appear on the Create Baseline screen.
The Create Baseline screen displays the baseline policy periods for each policy in the search results. The first column displays:
- A check mark if the baseline was successfully created
- An X if the baseline creation failed. Click the X to view more information about why baseline creation failed.
The Baseline Period column contains the submission number of the policy transaction that created the baseline policy period. Click this link to display the Impact Testing Policy Period Overview screen. This screen contains information about the baseline and test policy periods. The screen displays a baseline overview which includes premium details.
If you arrive at the Create Baseline screen by clicking Next with Previous Test Case, the screen has the following additional buttons:
- Click Reprocess Failures to reprocess all test periods that failed to generate baseline policy periods during the previous run.
- Click Recreate Baseline to regenerate all baseline policy periods.
Click Next. PolicyCenter displays the Select Rate Books screen. The rate books are grouped into Available Active Rate Books and Available Stage or Approved Rate Books panels. This screen only displays rate books from the policy lines contained in the selected products.
Select one or more rate books to move them to Selected Rate Books panel.

If you select more than one rate book, PolicyCenter rates the policy using the rate book with matching jurisdiction, underwriting company, policy line, and other factors. If more than one rate books applies, then PolicyCenter rates the test periods using the most recently changed rate book. When selecting a rate book for the test periods, PolicyCenter does not consider the rate book effective date.
Click Next to advance to the Testing Periods screen.

Similar to baseline policy periods, impact testing rates all test policy periods as submission policy transactions. For information on configuring impact testing to rate renewals as renewal policy transactions, see Rate renewals as renewal jobs in impact testing.
Do one of the following:
- To continue a previously started test in which the test periods are already quoted, click Requote Test Periods or go to step 12.
- Click Quote Test Periods to generate test policy periods rated using the selected rate books. When you click this button, PolicyCenter runs the Impact Testing Test Case Run batch process. You may exit impact testing and return to this screen later.
If you return to the Select Rate Books screen, the list of rate periods includes a Test Period column. If an applicable rate book was found, the $ indicates that the policy period was rated.

When the batch process completes, the Next button appears.
Click Next to advance to the Impact Results screen.

This screen displays Policies Affected and Financial Impact bar graphs. In both graphs, the X-axis is divided into Percent Change ranges. Each range represents a percentage change to the policy premium, such as 0 for no change or >5 for a 0% up to 5% change. In the Policies Affected graph for each bar, the Y-axis represents the number of policies in the impact range. In the Financial Impact graph for each bar, the Y-axis represents monetary amount of change in each impact range.

A table between the graphs displays the following information for each range:
- # – Number of policies affected in the impact range. This is the same information that the Policies Affected graph shows.
- % – Percentage of policies in the impact range.
- $ – Monetary amount of change in each impact range. This is the same information that the Financial Impact graph shows.
Click Create Excel Export File to export the test periods to Microsoft Excel format. When you click this button, PolicyCenter runs the Impact Testing Export batch process. You may exit impact testing and return to this screen later.
Click Download Excel Export File to open or save the results.
For each policy the Excel spreadsheet shows details for each cost on the policy when rated using the active rate book and the comparison rate book. The cost details include:

Baseline Rate Book
- Baseline Term Amount
- Baseline Actual Amount
Comparison Rate Book
- Comparison Term Amount
- Comparison Actual Amount
- Errors – Displays an error message if no rate book matched, or rating generated an invalid quote or encountered some other problem. The term amount and actual amount columns are blank.