When things go badly

Despite many months of plan- ning, generous support from manage- ment, and a comprehensive test plan, a recent project to implement cloud- based software to support a rapidly growing business segment failed to achieve its aggressive goals. This failure imperiled the entire organization, as the software afected half the company's revenue stream and customers quickly became frustrated.

The supplier immediately blamed new software code and promised a quick fix. Five months later, the application still failed to perform. Management notified the board of directors and shareholders, and the share price has fallen.

The overwhelming majority of soft- ware projects don't meet stated goals on schedule and budget. But this company represented an extreme example, where missed goals harmed the enterprise's business and value. While the company developed a comprehensive project plan and assigned adequate resources, it did not enact an effective risk management plan. Consequently, when things went badly, the entire company was placed in jeopardy.

Strategic projects similar to this carry significant risks to a company's health. A well-designed and implemented risk management plan may mitigate the effects of project failure or even prevent failure.

A comprehensive risk assessment should be performed as a first step for IT projects. Risk is defined as the prob- ability of a certain outcome multiplied by the consequence of that outcome. In this case, management placed too much faith in the software vendor's ability to deliver a new application that addressed the client's complex requirements. This software was new and relatively untested, and the soft- ware vendor was a small company. When things went badly, the vendor did not have sufficient resources to resolve all problems quickly.

A well-thought-out risk assessment identifies these issues early in the process, preparing a backup plan that can address problems as they happen instead of play- ing catch-up.

Most software projects include a pilot to process and analyze a small number of transactions. Problems addressed early, before their consequences become large enough to affect the entire organization, have little or no downside. Once manage- ment is satisfied that the pilot is running to specifications, then the entire organi- zation transitions to the new application. Most strategic software projects include a back-out option so management can reverse implementation if things go badly and the project team requires additional time to resolve identified issues.

In addition, comprehensive, end-to- end testing should be included in any strategic software project. Problems iden- tified during testing can be resolved with no risk to the client -- problems that could be catastrophic should they occur in the production environment.

While placing blame on the software supplier may be warranted, this rarely produces a satisfactory outcome. Soft- ware suppliers, like any other supply chain 22 Industrial Engineer partner, should be selected with great care using well-conceived criteria, includ- ing past performance, financial health, ability to resource complex projects, and history in resolving problems. Should the company fail to perform, blame rests with management's decision to include the supplier as a trusted partner.

Project failures represent significant costs. Every project plan should include risk management, and this may well be the most important element of the plan. Behind every risk management plan is a comprehensive risk assessment. Project managers would be well-served to assign sufficient time and resources to develop- ing the risk assessment, ensuring that identified risks are avoided, and imple- menting risk mitigation steps if a problem should occur. d

The supplier blamed new software code.

Paul Engle is a management consultant with an MBA in finance. He has more than 25 years of experience in management, operations, product development, sales and marketing, strategic plan- ning and business process improvement. You may contact him at [email protected].