Making Technology Work: How to Corral Your Runaway IT Projects

The Standish Group’s latest report reviewed 3,555 technology projects with labor costs of $10 million or more and found that a whopping 41 percent were complete failures (abandoned or started anew). Some 52 percent were assessed as “challenged”—way over budget, behind schedule or failing to meet user expectations. The bottom line? Only 7 percent could be considered successful.

McDonald’s Fails to “Innovate”
In 2002, McDonald’s pulled the plug on its five-year $1 billion “Innovate” project and wrote off $170 million. Conceived two years earlier, the project intended to allow corporate management to see inventory and sales figures at any of 30,000 stores at any time. It would give them an instantaneous birds-eye view of the entire system—from crew-member scheduling to cooking oil quality and temperature.

“Top management must be “sold” and on board to fully support projects.”

Management claimed the decision stemmed the need to reduce capital spending for the decision. However, the issue was likely sheer size, a plan to replace virtually every major back-office system. Plus, it was technically unrealistic to attempt to connect 30,000 stores in real-time—many in locations where little or no broadband connectivity existed.

Avon’s Broken “Promise”
In late 2013, after four years of development, Avon Products cancelled the “Promise” project, writing off $125 million in the process. Using a new SAP back-end system and a tablet-enabled e-commerce front-end ”Promise” intended to allow sales agents to showcase products on tablets, immediately check inventory and place orders online.

When piloted in Canada, basic functions, such as logging-in, saving orders and checking inventory did not work. “Promise” performed so poorly that many commission sales agents left the company, prompting management to cancel the entire project.

FoxMeyer Fails at ERP
FoxMeyer Drug went from profitability to bankruptcy in one year and was eventually sold to archrival McKesson for $80 million. Many blame its demise on problems with an ambitious $100 million ERP system the company installed. Implementation required extensive SAP modifications and caused real-time integration issues with software from different vendors.

When the launch failed, orders were incorrect, some shipping twice and some never shipping at all. There was no fallback plan; management had “bet the company” and lost.

Lessons from Failure
CEOs can avoid debacles like these by being wary of:

PROJECTS DRIVEN BY IT The project should be steered by a business executive singularly responsible for its success and ROI. Both McDonald’s and FoxMeyer’s business executives seemed detached from the project details. THE LESSON Do not start projects if users don’t have the time or interest to participate.

MEGA-PROJECTS McDonald’s $1 billion “Innovate” was so huge, complex and technically unrealistic that failure was nearly inevitable. THE LESSON Projects transforming several business units simultaneously should be broken into smaller chunks so that problems are contained.

BIG-BANG PROMISES If nothing will be implemented until “the end,” you are cruising for a disappointment. THE LESSON Break the project into smaller chunks—with tangible mid-project deliverables every 30-60 days—that management can see and touch.

LACK OF EXECUTIVE SUPPORT If you or your executive committee is lukewarm about the project, its success will be challenged. THE LESSON Top management must be “sold” and on board to fully support projects.

EXCESSIVE OVERTIME The quickest and easiest fix for troubled projects is overtime, which often goes unnoticed. THE LESSON A spike in project overtime is a red flag.

LACK OF METRICS Cost and person-hour budget overruns suggest project slippage. THE LESSON Use metrics granular enough to provide monthly budget variances.

MISSED MILESTONES Missed milestones are another red flag. THE LESSON Milestones must be clearly defined and weeks apart, rather than months or quarters.

SCOPE CHANGES Delaying or canceling features and functions signals issues. THE LESSON If you hear “we removed that function to make this milestone,” dig further.