Launching a new product or service typically involves many moving parts—including information technology (IT)—that must all come together simultaneously. Often, the long lead times that complex IT components involve become a drag on time-to-market. Yet, as the examples to follow illustrate, efforts to speed the process by buying or building workaround systems frequently backfire.
Launched on October 1, 2013 with great fanfare, Healthcare.gov promptly crashed. Applicants struggled to access the website and, when they finally did, computations were wrong, accounts were not recognized, insurance links didn’t function and live chat didn’t work. The site improved over time, but only after copious delays, frustration, manual workarounds, bad data, confusion, security problems and great expense.
Why the debacle? First, the scope was far too large. Overnight, the Affordable Care Act sought to displace one-sixth of the U.S. economy with no fallback plan. Second, the plan involved a “Big Bang,” with everything launching at once although many modules were untested and others had yet to be written.
Third, specifications for the project were continuously changing—right up to the launch date. Fourth, leadership forged ahead with the initiative despite copious warnings and danger signs.
HERSHEY’S GREAT WHITE HOPE
In the late ’90s, Walmart and other retailers were demanding more inventory and logistics information from suppliers. To adapt, Hershey decided to replace its legacy systems with “Enterprise 21,” a sophisticated combination of SAP, Siebel and Manugistics software. However, the company had little experience implementing a huge system that required many simultaneous business process changes.
After three years in development, some modules launched but several were delayed. In July 1999, with critical Halloween orders pouring in and Y2K looming large, management succumbed to pressure, scrapping the company’s legacy systems and launching the new system, “Big Bang.”
Problems in order fulfillment, customer service and shipping quickly arose, devastating the candy company’s most crucial selling season. Unable to get Hershey products, retailers switched to competitors. Hershey’s sales for 3Q 1999 dropped $150 million; profits declined 19 percent and the stock dropped by 8 percent as a result.
APPLE GETS BITTEN
When the recent live, video-stream introduction of Apple’s iPhone 6 glitched, all viewers saw for 25 minutes was a fixed test pattern and all they heard was Tim Cook’s presentation—in Mandarin.
Two days later, the Apple Store website crashed, frustrating customers who had queued overnight to place early orders. While the iPhone 6 is now a success, its launch was marred by these support issues, which were not unusual for the company’s launches.
Clearly, new product or service launches can be badly tarnished by problems with the supporting IT systems. To ensure that the IT workstream is on target, CEOs should:
Appoint a Program Czar
• Designate someone with total accountability and responsibility for the new product launch and the ability to spend the majority of his/her time on the program.
Avoid or Limit New Software
• Implement “mega projects” that span the entire company and require business process changes in small pieces.
• If new software is required, test and install it well in advance of the product launch date.
• Keep the old legacy systems operational until the new software is proven.
Understand All Workstreams
• Review the overall program schedule, questioning anything that isn’t clear. Stop everything if there is no documented plan or if it is superficial.
• Understand and monitor milestones.
• Ensure there are fallback strategies for things that go wrong.
Ensure the IT Components Are Complete and Sufficiently Robust
• Commission an IT workstream assessment by another department or an outside firm.
• Review the program weekly, then daily as the launch date nears.
• Be receptive to bad news.
CEOs can reduce risk and avoid sloppy and embarrassing launches by following these recommendations. An ounce of prevention is worth more than a pound of cure.