What a decision log is actually protecting
Every organisation makes hundreds of consequential decisions a year and remembers almost none of them accurately. Within eighteen months the people change, the context evaporates, and what survives is the outcome without the reasoning. New executives then reverse choices that were right, repeat ones that were wrong, and reopen debates that were settled for reasons no one can produce. The log exists to stop the organisation paying for the same decision twice.
The log, field by field
| Field |
What to record |
What goes wrong without it |
| Date and reference |
When the decision was made and a short identifier so entries can be cited |
Decisions become "that thing we agreed last spring", and two people mean two different meetings |
| The decision as made |
One sentence a minute-taker could defend: what was committed, for how much, effective when |
People remember the discussion instead of the commitment, and each remembers the version that suits them |
| Options considered and rejected |
Each real alternative in one line, with the actual reason it lost |
Rejected options return with every reorganisation, presented as new thinking, and cost the same analysis again |
| Decider and consulted |
The named person or body that decided, and who was consulted before they did |
Success has many parents and failure has none: "we all agreed" degrades into "nobody decided" the moment it matters |
| Assumptions at the time |
The two to four beliefs the decision rested on, stated as facts that could later prove false |
Hindsight rewrites the decision as obviously right or obviously reckless, and nothing is learned either way |
| Expected outcome |
What success was supposed to look like, with a number and a date where possible |
Later reviews compare results against remembered hopes, which shift to match whatever happened |
| Review date or trigger |
When, or on what event, the decision comes back for inspection |
Decisions are only ever revisited when they fail in public, which is the most expensive possible review mechanism |
Keeping the log alive
One person holds the log: a chief of staff, the PMO lead, or whoever runs the leadership meeting. Entries are written within a day of the decision, in five minutes, ideally as the meeting's final agenda item. Set the threshold explicitly: anything that commits budget, people or direction beyond the current quarter gets an entry, and anything below it does not, so the log stays short enough to be read. Its value compounds slowly. For the first month it feels like administration; by the second year it is the only reliable account of why the organisation is shaped the way it is.
Why decision logs die
- Only formal committee decisions get logged, while the consequential ones made in corridors and one-to-ones never enter the record.
- Entries are reconstructed a week later from memory, which converts the log from evidence into narrative without anyone noticing.
- The assumptions field is left blank because the assumptions felt obvious at the time, and obviousness is exactly what does not survive.
- The log becomes ammunition: once entries are used to assign blame, people negotiate the wording before it is written, and the record is finished as a learning instrument.
Reading the log as a diagnostic
A year of entries is a dataset about how your organisation decides, and it is more revealing than any survey. The same assumption failing three times, options that were never genuinely alternatives, review dates that pass unobserved: these patterns are hard for insiders to see because insiders are in the data. Senior operators from outside can read twelve months of entries and tell you not what to decide next, but what your deciding machinery systematically misses. That is a short, unusual and high-yield use of a confidential brief to the Global Board.
