3 mental models that cover 80% of your decisions

Hey Everyone,

You've probably read about dozens of mental models. Reversible doors, pre-mortems, first principles, second-order thinking. How many of them did you actually use in your last big decision?

For most leaders, the answer is zero. Under pressure, all those frameworks disappear and you fall back on instinct.

We put together a three-model toolkit that covers how fast to decide, when to walk away, and what to say no to.

Why three is enough

You've probably read about dozens of mental models. The problem is that under stress, time pressure, and political weight, you won't reach for any of them.

Cognitive scientists have a simple explanation. The kind of careful, deliberate thinking that frameworks require overloads your available working memory. Stress, fatigue, and time pressure reduce working memory, so your brain falls back on instinct instead.

The fix is to practice a very small number of models until they become automatic. Two or three, rehearsed often, will outperform a library of fifty that you vaguely remember from a book.

Model 1: Reversible or irreversible? (for speed)

Jeff Bezos sorts every decision into two types. Type 1 decisions are irreversible, like signing a ten-year lease or acquiring a company. Type 2 decisions are reversible, like testing a new pricing page or trialling a hire with a 90-day review period.

Type 1 decisions deserve slow, careful deliberation. Type 2 decisions should be made fast, at the lowest level possible, with roughly 70% of the information you'd ideally want. Waiting for 90% makes you slow, and being slow is almost always more expensive than being occasionally wrong on a reversible call.

The mistake Bezos sees repeatedly inside large organizations is treating Type 2 decisions like Type 1 decisions. Teams over-analyze reversible choices, stack meetings on top of meetings, and lose weeks on calls they could test and reverse in days.

Before your next decision, ask one question. If this doesn't work, can we reverse it within 30 days without major damage? If yes, decide now with what you have. If no, slow down and bring in more perspectives.

Model 2: Set the kill criteria before you start (for exits)

The hardest executive skill is walking away from something you've already invested in. Sunk cost bias makes this worse the more time, money, and identity you've poured into a project.

Annie Duke, the decision strategist and former poker professional, recommends writing down your "states and dates" before you commit. A state is a specific condition that would tell you to stop. A date is when you'll check.

Shopify's 2023 decision to reverse its logistics strategy is a recent example. CEO Tobi Lütke had acquired Deliverr for $2.1 billion in 2022 and invested heavily in building a fulfillment network. A year later, the company sold the logistics operation to Flexport, cut 20% of staff, and booked $1.7 billion in charges. Lütke described the decision publicly as getting back to what Shopify does best.

That reversal was painful and expensive. It would have been less painful with pre-committed exit criteria written earlier. Something like "if logistics margins haven't improved by X% by Q2 2023, we revisit."

For your next initiative, write two things down before you approve it and share them with your leadership team so the criteria exist outside your own head, where bias can quietly adjust them later.

What would have to be true in 90 days for this to be on track?
What would tell us to stop?

Model 3: Say no to good things (for focus)

Steve Jobs cut Apple's product line from roughly 350 products to 10 within weeks of returning in 1997. What followed was the largest stretch of value creation in business history. Jobs later said that deciding what not to do is as important as deciding what to do, and that he was as proud of the things Apple hadn't done as the things it had.

McKinsey's research on corporate performance backs this up at scale. Companies that reallocated more than 50% of their capital spending across business units over a ten-year period created roughly 50% more value than those that spread resources evenly. Only 8% of firms escaped the middle tier where the majority earn near-zero economic profit.

The hard part is that the things you need to say no to are usually good. They're reasonable initiatives with reasonable business cases, proposed by people you respect.

Saying no to bad ideas is easy. But saying no to good-but-wrong ideas is what separates the companies that break out from the ones that stay stuck in the middle.

Before your next planning cycle, write a "won't do" list alongside your priorities. For every initiative you approve, name one you're stopping or starving of resources. If nothing comes off the list when something goes on, your focus is an illusion.

Try this today

Pick one decision you're currently weighing. Before you discuss it again, write down your answers to two questions.

Is this reversible or irreversible? If reversible, set a deadline to decide by end of week. If irreversible, write down what would need to be true for you to walk away in 90 days.
Then look at your current project list and circle one initiative that would not make the cut if you could only keep half of what you're doing.

Bring that to your next leadership meeting.

Go deeper

👉 Farnam Street: Reversible and Irreversible Decisions – a clear breakdown of Bezos's Type 1 and Type 2 framework with practical examples

👉 Harvard Business Review: Before You Make That Big Decision – Kahneman, Lovallo, and Sibony's 12-question checklist for catching bias before a major call

👉 McKinsey: The case for behavioral strategy – the original research showing process quality matters six times more than analytical rigor

👉 McKinsey: Strategy Beyond the Hockey Stick – how capital reallocation separates the top 8% of companies from the stagnant middle

Coming up tomorrow

Tomorrow, we'll share how to tell if your exec scorecard is measuring the right things, and what to cut.

P.S. How many mental models can you name off the top of your head versus how many you actually used in your last big decision?