I went back over twelve completed projects and compared two numbers for each: the effort recorded against them, and the elapsed time from approval to go live. The gap was large and consistent, so I asked the project managers to account for it week by week.
The median project spent forty-one percent of its elapsed time waiting for a decision that nobody in my department was entitled to make.
They are not dramatic decisions. Which department owns a process that crosses three of them. Whether a data owner will sign off a field being shared. A legal opinion on one indemnity clause. Confirmation that a budget line released in principle in March is available in fact. Agreement on what should happen to the four hundred records that do not fit the new model. Each of these took between three and nine weeks, and almost none of the time was spent thinking. It was spent finding a forum, getting on its agenda, discovering the right person was absent, and going round again.
Our reporting hid this perfectly. A project sitting in amber with the note awaiting business input looks like a project being managed. It is a project that has stopped, and because the status stays the same for six weeks, nobody escalates, and the eventual conversation is about a missed date rather than about the eleven weeks of stillness that produced it.
Three changes, and none of them cost anything. Every project keeps a decision log alongside its milestones, with a question, a named decider and a date, reported at the same level as delivery progress. There is a standing forty-five minute slot each week attended by people with delegated authority from each function, which exists solely to answer questions that have got stuck. And decisions have a default: if no answer arrives within ten working days, the recommended option proceeds and is recorded as having been taken by the clock. That last rule has been invoked twice, and its real effect is that it almost never needs to be.
One uncomfortable finding stayed with me. A third of the slow decisions were slow because we had framed the question badly, offering the business a technical choice instead of a business one.
– Serguey Shinder
Top comments (0)