Most companies don't lack decisions; they lack a record of why they made them.
Meeting minutes say that something was agreed, not what the team expected to happen or which alternative almost won. Six months later, people remember the outcome and quietly rebuild the reasoning to fit it, which is classic hindsight bias. In the same Atlassian report, just 20% of knowledge workers said they were confident their team had an effective process for quickly informing other teams of decisions that may affect their work. The report is self-reported research from a collaboration-software vendor, so read it as a signal rather than a benchmark. It fits an older pattern: in a 2018 McKinsey survey of 1,228 managers, respondents said they spent 37% of their working time making decisions, and 61% said most of that time was used ineffectively.
Log the decisions that are expensive to undo, not every decision.
A log that tries to capture everything dies within a month. A practical filter for a company of 50 to 500 people: log a decision if it commits more than an agreed budget threshold, binds you to a contract or a senior hire, affects more than one team, or would take months to reverse. Our guide to reversible vs. irreversible decisions shows how to score that last point. In most leadership teams this means a handful of entries per month, not dozens. Routine operational calls stay in your normal tools.
A useful entry fits on one page and answers eight questions.
Use the same structure every time so entries can be compared later:
- Decision: one sentence, written so an outsider understands it.
- Date, decision owner and approver.
- Context: the problem, the deadline and the constraints that applied.
- Options considered, including the ones you rejected and why.
- Key assumptions, each one testable (see how to track assumptions before you commit).
- Expected outcome, with a number, a time frame and your confidence level.
- Main risks and the stop rule or reversal trigger.
- Review date and the measure you will look at, plus links to the source documents.
More on Execution Intelligence — every week.
Join 2,400+ leaders getting the decision intelligence briefing.
Write the entry before the outcome is known, and never rewrite it afterwards.
The value of a log comes from capturing what you believed at the time. Have the decision owner draft the entry within 48 hours of the decision and the approver confirm it. Treat the log as append-only: later information goes in as a dated note underneath, not as an edit to the original reasoning. Recording a confidence level, even a rough one such as "70% likely to hit the target", is what later lets you see whether the team is systematically over- or under-confident.
Review on a schedule and judge the decision, not only the result.
Put a 30-minute review in the calendar each quarter and go through every entry whose review date has passed. For each one, ask: what actually happened, which assumptions held, which didn't, and was this a sound decision with bad luck or a weak decision that got lucky? Our article on reviewing a decision after launch covers that conversation in more detail. Over a year, the patterns that come out of these reviews (for example, consistently optimistic sales ramps) are worth more than any single entry.
Hypothetical example.
A 140-person logistics software company in Hamburg decides to replace its Benelux reseller with its own two-person sales team. The entry records the decision, the rejected options (keep the reseller, renegotiate margins, a hybrid model), and three assumptions: that most reseller-managed customers will renew directly, that the new team can be hired within four months, and that direct margins will cover the extra cost within 18 months. It sets a stop rule and a review date nine months out. At the review, hiring was on time but direct renewals were well below plan. Because the assumption was written down, the leadership team can see exactly which part of the reasoning failed and move to the hybrid option, instead of arguing about who wanted the change. The company and all figures are invented for illustration.
Keep the tooling simple and the access deliberate.
A shared document or spreadsheet with one row per decision and a link to the full entry is enough; the habit matters more than the software. Name one person responsible for keeping the log current. Some entries, such as those on individual employees, restructuring or acquisitions, will contain confidential or personal data, so restrict access accordingly and check your data protection obligations; this article is general guidance, not legal advice. If you run a decision through the NeuroAgent council, the Challenge and Synthesis stages and the AI-assisted boundary check against your constraints provide most of the raw material for an entry, and you can try it free without an account.
Research context
PMI's 2024 research reports an average project success rate of 73.8%, leaving substantial room for better execution systems.
Project Management Institute: Amanda · Social Connector of the Profession 2024 (2024)This independently published statistic provides context for the topic; it does not substantiate every claim in this essay.
One decision framework, every week.
The week's best articles on making and executing high-stakes decisions, backed by sourced research.