W11 - How to Make Decisions

There are increasingly more moments for decision-making, especially for longer-term decisions.

You see many problems and must decide whether to invest and how much. Using AI integration with work as an example, it’s hard to determine the right level of investment. If you don’t invest you’ll regret it — last year we found an accounting scenario but invested little, and this year we’re already behind. If you do invest it’s hard to tell if it’s driven by FOMO. You tack on a few superficial AI features that look cool in demos, but in real use they fall short. After trying them out, no one wants to open them again.

Last week I came across three decision-making models in practice that are worth trying.

The first is used in this year’s business planning: demand level, demand space, implementation difficulty, and cost. It evolved from the RICE assessment method. From observation I’d add that the cost element should pay more attention to opportunity cost.

The second is a decision flow I saw earlier on X, which also complements single/dual gates. It starts by asking whether the decision is reversible, then how much time you have to decide, then how uncertain things are, and finally whether you would regret not acting in the long term.

The third is the loss-vs-gain four-quadrant model, which Jerry used live during last week’s planning prep. It’s a variant of the important/urgent quadrants; the core is to identify items where acting yields large gains, or not acting incurs large losses.