Every generation of management produces an instrument for turning intention into action, and every generation watches its instrument harden into ritual. Management by objectives was meant to replace supervision with purpose, and in a great many organisations it became an annual form. The balanced scorecard was meant to rescue strategy from the tyranny of the financial quarter, and it became a dashboard in four colours. Objectives and Key Results arrived as the corrective to both, and stands accused, in its turn, of being a quarterly exercise in writing down what a team was going to do anyway.
What is curious is not the decay, which is ordinary. It is the post-mortem, which never varies. When an instrument fails, we conclude that the organisation implemented it badly: the leaders were not committed, the culture was not ready, the people did not buy in. The instrument walks away from every accident without a scratch. This is a comfortable arrangement for the instrument and for those who sell it, and it may also be true. But a method that can only fail through misuse occupies a strange position. Its successes are evidence for it, and its failures are evidence about everyone else.
I offered a version of this observation in my introduction, and then spent seventy-five minutes listening for the moment my guest would blame the instrument. It never came.
My guest this week is Tim Newbold, founder of OKR Quickstart, which coaches organisations in Australia and the United States to adopt Objectives and Key Results and, more to the point, to keep using them after the coach has gone. He came through technology delivery, and dates his interest in outcomes to an internship at Australia Post, where a colleague of eighteen years’ standing told him that not one of the hundreds of specifications he had written had ever become a system a customer used. Newbold heard it as a prophecy. He also hosts Strategy Candy, a podcast on strategy, OKRs and product management, and he is unusually willing to say on air which of his own engagements went wrong.
What follows are the ideas from the conversation I have continued to turn over since.
