By now we have all heard about the “problematic” (*ahem*) launch of the federal government’s health insurance Web site. Today, a report by the consulting firm McKinsey & Company, sheds new light on the dysfunction that contributed to the failed launch of the Web site.
A bumpy launch of a massive Web site is not news. Virtually every Web site launch with this level of complexity and anticipated traffic was bound to have some glitches and problems in the first few weeks. That said, the report detailed some warning signs that, would they have been heeded, might have mitigated some of the failures.
Even simple Web sites, with no secure transactional functions, can benefit from best practices demonstrated by successful larger Web site launches.
Here are some of the key warnings from the report:
(1) Evolving requirements
This is likely the biggest challenge we face with clients in preparing for a Web site launch. Few of them want to take the time and effort up front to think through the project from end to end, outlining the specific requirements and user experience needs. This is not busy work, but essential to keeping design and development costs down.
Often clients are worried that their board or management “just wants to see something,” and they want to rush through this ESSENTIAL phase.
STOP. Fight the urge to circumvent the appropriate time for planning. That said, “agile development” methods allow for the release of software (let’s face it: launching a Web site is launching software) in phases, rather than one big splash. Don’t expect a fully functional site at launch, unless you plan to dedicate SIGNIFICANT development time without the site ever seeing the light of day.
(2) Too much dependency on third parties
We may be shooting ourselves in the foot here by suggesting you not use the expertise of outside consultants, designers or developers. But here it goes: the design and development of your Web site can not happen in a vacuum. You don’t get to go about your business and wake up one day and have a fully functional Web site that magically launches.
The role of third parties is to FACILITATE the internal discussions, CHALLENGE assumptions and sacred cows, IDENTIFY impacts on business processes, PLAN for maintenance (the launch is just the beginning, folks!), and MITIGATE risks.
(3) Not enough time for testing
This was perhaps the biggest mistake of all. This is the back end of the planning. By the time the site has been designed and developed, clients are anxious to get it launched so they can demonstrate to their superiors that they’ve accomplished something. Sometimes projects fall behind schedule and the temptation to make up for time by short-changing the testing phase can be overwhelming. This is a huge mistake.
If you are behind schedule and have some target launch date (and ask yourself, why did we set this date? Was it tied to some other significant occurrence, or was it an arbitrary random date picked out of the blue?) by which you MUST launch, then you need to revisit your requirements and scale back. Ask yourself, “What can we launch withOUT?”
Do not attempt to launch a site without proper review and testing.
(4) Launching too big
Too many clients think they will launch to all sorts of fanfare. They think they will launch a site that meets their entire functional wish list. They want to send out direct mail, take out ads and billboards.
Let the site settle down before you tell everyone about it. A limited launch, promoted internally and to “friends and family” first will reveal any lingering errors or bugs among the people you trust. Generally we recommend a soft launch, with the fanfare coming at least two weeks later. The bigger the site, the longer the soft launch.
Finally, be transparent with management and your executive board. Be honest about where you are and how you’re doing, and what issues may be coming and how you’re dealing with it. Managing expectations, minimizing surprises and keeping clear lines of communication will result in a much more successful launch of your site.