Customize GL Outlier Detection

Machine learning provides incredible insight when comparing submitted transactions to historical norms.

Configure General Ledger (GL) Outlier Detection to best suit your organization’s requirements.

Subscription

Company

Regional availability

All regions

User type Business user with admin privileges
Permissions

Administration

Application subscriptions: List, View, Subscribe, Configure
Prerequisites

GL Outlier Detection enabled and data model is ready

General Ledger journal entry approvals enabled

By default, GL Outlier Detection uses all enabled dimensions in your company as part of the evaluation process. It imposes no restriction on the amount of the submission.

Your company might determine that some dimensions are more important to include than others. Further, it might require that only submissions over a certain amount need to be scrutinized more carefully.

If you prefer to have a more customized setting for evaluation, use the Outlier Detection customization settings. This is needed only when the default values are not right for your company needs.

GL Outlier Detection customization is defined at the top level. Entities adhere to the settings created here.

Set the evaluation configuration

  1. At the top level, go to General Ledger > Setup > Configuration and scroll to Outlier Detection customization.
  2. Select a journal to customize.
  3. For each journal being evaluated, select the Currency to use.

    If you have different journals within different entities and currencies, set the currency for each one.

  4. Set the Materiality threshold.

    Transactions below this threshold are not material and therefore are not evaluated.

  5. In the Dimension priority column, define which dimensions to use in the evaluation process and the order of their importance. By default, all dimensions that are enabled in the company are used.
    • If no dimensions are listed, it means that all of them are used in the evaluation. Select Select to create a custom list of dimensions to be evaluated.
    • Under Selected items, drag the dimensions to match the order that you want the evaluation process to use. The higher the dimension is in the list, the more weight it carries in the evaluation process.
    • If dimensions have already been selected, select Add or edit to make any changes.

This is because some journals listed are entity restricted, and the Outlier Detection customization is done at the top level. Each currency is listed (entity-level journals can use different currencies).

How materiality works

For example, you have 2 transactions for approval in the Payroll journal. Your materiality is set at R50, and your dimensions are Class and Location.

The first submission is for R25.50. Because it’s below the materiality threshold, it is not evaluated.

The second submission is for R200.00. This submission is evaluated for class and then location. If there are no anomalies, no outlier is flagged. But if either the class or the location (or both) contains an anomaly, the submission is flagged for review.

No other dimensions are included in the evaluation.

Optionally, enable the Outlier Assistant to change the workflow and return flagged entries to the submitter.