Scores let you know where you stand, whether it is a test score in academics or the final score of the Super Bowl, they are a quick way to measure past performance. Past performance is an important indicator for future performance and using scoring rules in business decisions can help predict the probability of certain outcomes.

Scoring rules are specific types of business rules that use scores to measure how likely a prediction is to come true. They are common tools in making business decisions and are often used to evaluate risk and prioritize opportunities. For example, scoring rules can measure the likelihood that:

• A borrower will pay back a loan
• A buyer of insurance is going to make a claim
• A credit card transaction is fraudulent
• Any single investment will provide a good return
• A patient has a certain condition

With these metrics, business leaders can better allocate resources and take action to optimize outcomes.

Implementing Scoring Rules

One of the great things about scores is that they summarize much larger sets of data. Simple scores may just incorporate a few data elements. More sophisticated scoring rules may aggregate scores from multiple more focused scoring rule sets and incorporate additional important metrics. For example, a risk scoring system that determines if an individual gets a mortgage may take into account a credit score as well as debt-to-asset ratio, and a score that measures the risk based on location and type of property being purchased.

With the ability of scoring rules to summarize data, decisions and actions can be made much quicker. For example, investment opportunities can be ranked and managers can allocate capital appropriately. Thresholds can also be easily set up. If the fraud score for a certain transaction falls outside a certain parameter, it can be denied and workflows that conduct greater security checks can be initiated.

Since scores summarize data into a single figure they can be very helpful in matching different entities. One example is matching the right product with a customer. An investment manager that is offering a variety of investment portfolios can match the risk and reward tolerance of the client to the risk-reward profile of a specific portfolio.

While scoring rules can assist managers in making decisions they can also be used to support automated processes. Data cleansing is a good example. By creating a set of scoring rules, systems can evaluate the likelihood that similar records are actually duplicates or not.

Scoring rules can be used in any number of applications. The ability for scores to summarize data and measure probability can also be instrumental in triggering actions and kicking of workflows to increase productivity.

Scoring Rule Transparency

While scoring strategies can be very powerful in summarizing data and optimizing decision making, they do not tell the whole story. A black box may be able to spit out a score but understanding the underlying models that calculate scores is important to making the best decisions.

Situations change and models drift. This can be on a macro level or a micro level. Extenuating circumstances may produce an unreliable score with low predictability in a specific situation. On a macro level, changes in markets or environments will degrade the ability of scoring rules to predict outcomes. Consequently scores and models need to be understood by the professionals that are using them to make decisions so they can identify where there are issues.

Scores are great for flagging and prioritizing, but they are not foolproof, final decisions are typically made by trained professionals.

Learn more about scoring rules by downloading our whitepaper How to Use Scoring to Simplify Complex Business Rules in Process Automation.