Jayesh Jain provides an overview of fallout management....

Fallout management is the ability to resolve fallout and allow a transaction to continue processing.

The most common fallout in a typical business environment is order fallout, or order failure which occurs when an order fails during processing.

This fallout is very important because it directly impacts revenue and customer satisfaction. The most common fallout scenarios are:

• Failure in a downstream system; for example, a failure in an order processing system or supply chain management.

Scenarios include:

• The data was received by the downstream system, but was missing or incorrect and could not be processed. When a downstream system detects missing or incorrect data , it returns an error which fails the transaction.

• An internal error due to data corruption.

• An internal application has occurred in the downstream system.

• Failure in a network or system resource, typically due to network connection, database or middleware failure, or a downstream system being down or running slow.

• Failure in a run-time initiating system execution, which usually happens when an unresolved dependency prevents a transaction from being processed. For example, an order cannot be processed because the address validator application or credit check application is down.

• Failure when a transaction is created in an initiating system; for example, an order cannot be dispatched because the delivery address is missing. If the initiating system fails business/data validation it may accept the transaction but immediately places it in a failed state.

Managing fallout

Detection - You must be able to detect when a transaction has failed. For example:

• A process can transition a transaction to a failed state. When it does, you can specify to send a message to a fallout management team.

• The fallout management team can monitor the progress of transactions, paying special attention to the failed ones. When a transaction is identified as a fallout, the transaction state should typically be changed to the failed state.

Notification - Depending on the error, you might need to notify other systems that a problem occurred and the reason for the failure. For example, the CRM system must be updated if an order has failed so that the CSR can inform the customer.

You can configure your system to send notifications to fallout management teams or to notify a trouble-ticketing application.

Recovery/Correction - Corrective measures must be taken automatically or manually to fix the problem and then you can resume or restart the transaction. You need to search for failed transactions and examine their behaviour to determine the cause of the error to prevent transactions failing in future.

Typical transaction recovery can be carried out in various ways. For example:

• A revised transaction can be re-submitted from the initiating system after correcting data.

• The fallout management team can edit the transaction data and resume processing.

• The failed transaction can be terminated and a new transaction submitted.

Ideally you should invest in an automated fallout management, which corrects errors as it occurs, though implementing these systems can be very expensive.

Alternately you can model manual fallout management, which supports manual intervention to correct errors. This requires businesses to be proactive.

If you are investing in fallout management applications, it should provide auto correction, correction assistance, manual correction queue handling and fallout dashboard, reporting, notification, rules engine and orchestration.

Investing in fallout management makes business sense because an incomplete order costs money and completing the order makes money.

Jayesh Jain works in information technology and services and was recently appointed as vice president, membership, for IIBA's (International Institute of Business Analysts) New Zealand chapter.