Decision-Making
Decision Making
What is decision-making?
- Choosing the best of several options
- Perfect information doesn’t exist (bounded rationality)
- Guard against bias (confirmation bias, groupthink)
The first thing that changes in an organization once it treats decision-making as a competency is the evaluation standard. Before, a good outcome meant a good decision and a bad outcome meant a bad one. But the quality of a decision should be judged by the information and logic available when it was made, not by the result. If you measure a reckless investment that survived by luck and a well-grounded investment that failed because exchange rates swung by the same yardstick, the organization learns not how to decide well but how to avoid accountability. Once the habit of recording the rationale for each decision on a single page takes hold, post-mortems shift from blaming people to reviewing the judgment structure.
In practice, simply splitting agenda items into two tracks speeds things up. Reversible decisions are made by the responsible executive alone once about 70% of the information is in, while hard-to-reverse decisions add a step where the case against is heard first. Here, keep consensus and consultation distinct. Gather opinions widely, but there must be exactly one person ultimately accountable, and that name must be written on the agenda sheet. The boundary with problem-solving is clear as well. Problem-solving is about identifying the root cause all the way through; decision-making is choosing within a deadline, before the cause is fully established.
Bias doesn’t disappear through willpower, so you block it with mechanisms. Typical examples: a pre-mortem, where right before a decision people are asked to 'assume this choice failed a year from now and write down why'; a red team, with one person assigned the opposing role on each agenda item; and a rule to circulate materials 24 hours before the meeting and have the most junior people speak first. Once a senior person speaks first, the room’s conclusion is effectively set. Still, these mechanisms only work after it has been shown a few times that dissenters face no penalty.
The most common mistake in practice is believing that gathering more data leads to a better decision. As data grows, confidence rises, but accuracy doesn’t rise nearly as much, and while the review drags on, the options themselves disappear. Another is assuming that because no one objected, everyone agrees. Silence is often not agreement but reservation, and at the execution stage you hear 'I never felt good about this from the start' and schedules slip. In organizations that blame only the decision-maker when results are bad, you end up with meetings where no one raises a hand first.
Inwoo LeeExpert in leadership, communication, and performance management
John KimInstructor John Kim, CEO of Nanum Consulting and author of “B2B Marketing Design”Decision science was established through Herbert Simon’s 'bounded rationality' and Kahneman and Tversky’s research on judgment and bias, among other work.
Consider a hypothetical case. Company A, an auto-parts maker, is a mid-sized firm with 320 employees and KRW 118 billion (≈ US$88,283,700) in annual revenue. A KRW 13 billion (≈ US$9,726,200) line-expansion proposal had been put on the executive meeting agenda and pulled off again for eight months, and even small items like switching suppliers stalled while waiting for the CEO’s sign-off. On the shop floor, people started saying 'It’ll just get postponed again anyway.'
- A strategic planning team lead reviewed the 47 executive meeting agenda items from the past year. 19 ended with no conclusion, only 'let's revisit this next time,' and 7 of those came back to the table three or more times.
- The CEO split agenda items into Class A and Class B. Hard-to-reverse items such as capacity expansion or setting up a new entity were Class A; swapping a supplier or choosing a training program were Class B, and for Class B items the responsible executive was delegated to decide alone once about 70% of the information was in.
- The head of quality finalized the replacement of inspection equipment, classified as Class B, in just two weeks. The production team lead, who had opposed it to the end, agreed once given a condition: 'We’ll run it for one quarter, and if the defect rate isn’t under control, we’ll go back to the original equipment.'
- For the Class A KRW 13 billion (≈ US$9,726,200) expansion, a 30-minute red-team segment was added before the meeting. A finance team manager pointed out that 'a KRW 1,450 (≈ US$1) per dollar exchange-rate scenario is missing entirely from the cost assumptions,' and the CEO cut the investment from KRW 13 billion (≈ US$9,726,200) to KRW 9.2 billion (≈ US$6,883,100) and carried it out in two phases.
- Six months later, the average time to decide per agenda item fell from 34 days to 11, and resubmitted items dropped from 19 to 4. Two decisions were reversed after delegation, but each loss was KRW 30 million (≈ US$22,400) or less.