Gerald Ajam

“A good plan specifies the outcome and the path to it”

Specify where cause and effect are known. Where they emerge, shape the conditions and probe.

Status Position

Published September 2026

All counterpositionsImplementation

Usually said

“A good plan specifies the outcome and the path to it”

The shift

From: Specify the path

To: Specify where you can, probe where you cannot

More precisely

Where cause and effect are known in advance, a good plan specifies the outcome and the path. Where they emerge from interaction, it shapes the conditions, runs safe-to-fail probes, and says beforehand what would make it amplify or stop each one.

In short

A plan with a clear outcome and path is easy to resource and to hold to account, and for many tasks that is right. But where results depend on how people respond, nobody can know the path in advance, and a specified one gets followed after it stops working. The precise version specifies where it can and probes where it cannot.

The argument

From the claim to a better one

  1. The strongest formulation

    A plan without a clear outcome cannot be held to account, and a plan without a path cannot be resourced. Staff deserve to know what they are working towards, and how.

  2. Why it is attractive

    It is legible to the principal, the cluster superintendent and anyone else who reviews the school. Milestones can be checked, and a missed one is a clear signal.

  3. Where it holds

    It holds where cause and effect can be worked out before acting: building a timetable, moving to a new reporting system, training staff on a procedure. Expertise can find the answer, and a specification gets it delivered.

    Where it breaks

    It breaks where the outcome depends on how people respond to each other and to the plan itself: how students take to new feedback, how staff use a new rating scale. There the path cannot be known in advance, and a specified one gets followed long after it has stopped working, because the milestones still look met.

  4. What evidence matters

    Snowden and Boone (2007) on matching how a leader decides to whether cause and effect can be known in advance. Glouberman and Zimmerman (2002) on the difference between complicated and complex problems. Mintzberg (1994) on strategies that emerge in action rather than being planned. What matters is whether the outcome depends on interactions nobody can predict, and whether the plan has a way to learn that its path is wrong.

  5. So, more precisely

    Where cause and effect are known in advance, a good plan specifies the outcome and the path. Where they emerge from interaction, it shapes the conditions, runs safe-to-fail probes, and says beforehand what would make it amplify or stop each one.

Reading it

The precise version, word by word

Where cause and effect are known in advance1, a good plan specifies the outcome and the path. Where they emerge from interaction, it shapes the conditions2, runs safe-to-fail probes3, and says beforehand what would make it amplify or stop4 each one.

  1. Known in advance

    Expertise or analysis can work out the answer before acting, as with a timetable or a procurement.

  2. Shapes the conditions

    Sets the boundaries, resources and routines that make good outcomes likelier, without fixing what the outcome must be.

  3. Safe-to-fail probes

    Small tries, several at once, each cheap enough to drop if it goes badly.

  4. Amplify or stop

    Criteria, set before a probe starts, for when to grow it and when to end it, so the decision does not rest on who championed it.

What it changes in practice

  • Sort the plan’s parts into those whose result can be worked out in advance and those whose result depends on how people respond.
  • Specify the first kind. For the second, run two or three small probes at once instead of one large launch.
  • Write the criteria for growing or stopping each probe before it starts.

A case, in general terms

A school plans a new approach to feedback. Buying the software, setting the calendar and training staff on the tool can all be worked out in advance, and the plan specifies them. Whether written comments, whole-class feedback or short quizzes change what students do next cannot. The school tries all three in different departments for six weeks, having agreed beforehand what result would make it extend each one.

Singapore