Risk analysis and minimizing risk

Network Optimization provides features that optimize the network to minimize the risk level and methods to evaluate risk in your supply chain.

• Minimize risk and respect maximum risk levels - Risk is included in the solver objective when you run Sequential Optimization with risk as an objective. You can apply constraints at the product and network levels that define the maximum risk as the solver minimizes risk based on aggregated risk scores. These constraints can also be applied when not optimizing for risk (they are not part of the objective in that case). Output tables report aggregated risk scores and summaries of the percentage of high, moderate, and low risk flows in the solution.

Network Optimization provides additional methods to evaluate risk in your supply chain. These are not dependent on optimizing for risk:

• Risk Exposure Summary - This output table provides a summary of risk-oriented metrics in your current supply chain network configuration.

• Composite Risk Scores - This feature enables you use to risk scores applied to sites to generate a composite risk score for your network. This composite score combines a number of site-based risk scores and a network criticality weight applied at the node level. You can use the information to analyze the relative risk posed by the various suppliers in your network. These scores do not affect the optimization results unless you are running Sequential Optimization with risk as an objective.

Optimizing to minimize risk

Risk is included in the optimization objective when you run Sequential Optimization with one or more of the risk-based objectives/constraints defined. Risk-based objectives are available in the Sequential Objectives table. Define records with one or more of these objectives, then run Sequential Optimization:

• Total Network Risk - This objective is based on the weighted average aggregated risk across all the facilities in the network.

• Maximum Network Risk - This objective relaxes the risk constraint defined by the Maximum Network Risk value specified in Network Optimization Options, and puts the deviation term into the objective.

• Maximum Product Risk - This objective relaxes the risk constraint defined by the Maximum Risk value specified in the Products table, and puts the deviation term into the objective.

The average risk score is a weighted average of risk scores by site throughput or revenue. See Determine the risk formulation basis.

Maximum risk constraints

You can apply a risk constraint at the product level by entering a Maximum Risk value between 0 and 100 in the Products table. The product risk is the weighted average site risk score of the sites associated with the product. The site risk scores are weighted by the product flows or revenues through the sites to calculate the weighted average site risk score. Site risk scores are based on the weighted average of user entered scores in the Risk Profile table when a Risk Profile is assigned to a site. It is not averaged across products. See Composite risk scores. The Maximum Risk constraint is against this product risk weighted average.

You can also define a network level risk constraint in Network Optimization Options. On the Modeling Options tab, turn on the Enable Maximum Network Risk Constraint switch and enter a value for Maximum Network Risk.

The values for both the product level and network level maximum risk are in the range 0 to 100, where 0 indicates no risk tolerance. Risk is bucketed into low (less than 33.33), moderate (33.34 to 66.67), and high (greater than 66.67) when reporting in the Network Risk Summary and Product Risk Summary output tables. If you define maximum risk constraints at both the product and network level, the more constraining value is applied.

If you are not using Sequential Optimization to solve for risk, the maximum risk constraints are not part of the objective, but they are still applied as constraints.

Determine the risk formulation basis

Use the Risk_Formulation_Basis parameter in the Config_NO table to configure how the risk score is applied. Set it to "Revenue" if you want to multiply revenue with risk score as the risk term in the objective, otherwise set it to "Flow" to multiply the throughput with the risk score.

Risk information in output tables

Several output tables report information based on the use of risk as a constraint and as an objective:

• Network Summary - provides an aggregated risk score.

• Risk Profile Summary - provides an aggregated risk score in addition to the summary information for site risk profiles.

• Site Risk Summary - provides an aggregated risk score at the site level.

• Site Product Risk Summary - provides an aggregated risk score at the product level.

• Network Risk Summary - provides an aggregated risk score and summaries of the percentage of high, moderate and low risk flows in the solution at the network level based on optimizing for risk.

• Product Risk Summary - provides an aggregated risk score and summaries of the percentage of high, moderate and low risk flows in the solution at the product level based on optimizing for risk.

• Risk Exposure Summary - includes metrics that help you to identify potential vulnerabilities in the supply chain network. For example, the table reports the number and percent of customers with demand served by a single source and similar metrics for products with a single manufacturer.

Composite risk scores

For the supplier nodes in your supply chain model, you can augment the supply chain information with risk scores from Coupa Risk and Performance Management (RPM). This feature enables you to perform a more complete risk assessment of suppliers by considering a number of metrics including financial score, judicial score and news sentiment. The RPM scores can be populated in your Decision Data Model (DDM). From there, this data can be pulled into your model for use in Network Optimization. You can also enter risk profile information directly in the model.

You can apply a relative weight to each score, and an overall node criticality weight which is applied to the actual solution flow when calculating the composite risk scores. For example, supplier financial risk may be more significant to you than other risk factors. In this case, you apply a higher weight to the financial score than to other scores to represent this condition.

The standard risk scores are the following:

• Financial Score - A higher value indicates the entity is more financially sound, and a lower value indicates a higher risk of the entity going bankrupt.

• Judicial Score - A higher value indicates an entity less likely to be involved in legal proceedings, and a lower value indicates the entity is involved in legal proceedings.

• News Sentiment Score - The value represents sentiment analysis of the entity based on a list of news sources. A higher value indicates that news sentiment of the entity is favorable.

• Screening List - The value indicates whether or not the entity is on a screening list in one or more geographies. A value of 0 indicates that the entity is not on any screening lists and is therefore at a lower risk. A value of 1 indicates the entity is on one or more screening lists.

• Residual Score - A higher value indicates the entity has reduced the risk of a mishap, and a lower value indicates a higher risk of the entity experiencing a mishap.

• Inherent Score - A higher value indicates the entity has a low risk of a mishap, and a lower value indicates a higher risk of the entity experiencing a mishap.

• Business Continuity Score - A higher value indicates the entity is able to operate without disruption to business, and a lower value indicates that the entity is likely to suffer business disruptions.

• Performance Score - A higher value indicates the entity is performing within expectations, and a lower value indicates that the entity is not performing to expected levels.

• Labor Score - A higher value indicates the entity is maintaining the needed labor force, and a lower value indicates that the entity may not be able to maintain the required labor force.

• Geography Score - A higher value indicates the entity is situated geographically in a location that supports its business, and a lower value indicates the entity is in a geography that may be problematic for its business.

Additionally, you can define additional risk factors using the four Custom Score columns. For these columns, a higher value indicates a higher risk.

You associate profiles with suppliers in the Sites table. Using scenarios, you can exclude or constrain the activity for high risk suppliers to assess the impact on your network. You can also analyze the effects of various weighting approaches.

When you optimize the model, a set of output tables are populated that provide the risk scores per profile along with the percentage of total flow they affect. In addition to the scores defined in the Risk Profile table, the output tables include a Composite Risk Score that takes into account the weights applied to each of the individual risk scores:

• Risk Profile Summary- This output table provides risk scores at the risk profile level.

• Site Risk Summary - This output table provides risk scores at the risk profile-site level.

• Site Product Risk Summary - This output table provides risk scores at the risk profile-site-product level.

The Composite Risk Score, available in each of these output tables, is calculated as:

((100 - Financial Score)* Financial Weight + (100 - Judicial Score) * Judicial Weight + (100 - rescaled News Sentiment Score) * News Sentiment Weight + rescaled Screening List * Screening List Weight + Residual Score * Residual Weight + Inherent Score * Inherent Weight + (100 - Business Continuity Score) * Business Continuity Weight + (100 - Performance Score) * Performance Weight + (100 - Labor Score) * Labor Weight + (100 - Geography Score) * Geography Weight + Custom Score 1 * Custom Weight 1 + Custom Score 2 * Custom Weight 2 + Custom Score 3 * Custom Weight 3 + Custom Score 4 * Custom Weight 4 + % Total Flow * Node Criticality Weight) / (Financial Weight + Judicial Weight + News Sentiment Weight + Screening List Weight + Residual Weight + Inherent Weight + Business Continuity Weight + Performance Weight + Labor Weight + Geography Weight + Custom Weight 1 + Custom Weight 2 + Custom Weight 3 + Custom Weight 4 + Node Criticality Weight)

When the model is solved, both the News Sentiment Score and Screening List input values are rescaled in the calculation to align with the 0 to 100 range for the other scores.

The Network Summary includes an overall Network Risk Health column that you can use to evaluate the high level effect of supplier risk mitigation strategies on your supply chain.