From a launch-day slip to a launch process

The problem

Launches ran across many regional sites. One page went live on a regional site a day early because of a configuration problem, and was live for about 20 minutes before quality control caught and reverted it.

What I did

• The quality team investigated. My part was prevention: I created a standard workflow any region could copy, and proposed that the other segments adopt it. It became part of the publishing procedures.
• As Global Launch Manager, I reviewed the existing procedure on my own initiative, found the gaps and designed a documented process for the next two processor generations.
• It added checks at one week, three days and one day before launch, confirming that automations matched the server and each region's dates.
• Coordinated multi-region teams against fixed dates, and used sprints to improve between launches.

The result

Over two processor generations I published more than 30 pages and publications. The early publication never happened again, and the process was still in use after I left.