Companies rarely die from bad decisions. They die from decisions made four months after they mattered.
A bad decision at least produces information. You learn something, you correct, you move. A slow decision produces nothing except the slow leak of optionality — the competitor who ships first, the candidate who takes the other offer, the market window that closes while the analysis is still being refined.
Yet almost every incentive inside a growing company favors slowness. Nobody is ever criticized for wanting more data. No one is fired for scheduling another review. The person who says "let's think about this more" always sounds more responsible than the person who says "let's decide today."
The two kinds of decisions
The most useful distinction I know: some decisions are reversible and some are not.
For reversible decisions — a pricing experiment, a marketing channel, an org structure you can adjust in a quarter — speed dominates. The cost of being wrong is one cycle of learning. The cost of delay is the entire cycle you didn't run. You should make these fast, with visibly incomplete information, and treat the discomfort as the price of speed.
For irreversible decisions — a co-founder, an acquisition, a market you'll spend three years in, a fundraise on bad terms — rigor dominates. Take the time. Get the outside view. Be willing to walk.
Most organizational dysfunction comes from applying the wrong mode. Teams agonize over reversible decisions and rush irreversible ones, usually because the reversible ones are the ones everybody has an opinion about.
Ask of every pending decision: if this is wrong, what does it cost to undo? If the answer is "a quarter," stop deliberating and go.
Consensus is not alignment
Consensus means everyone agrees. Alignment means everyone will execute, including those who disagreed. These are different, and confusing them is expensive.
Consensus-seeking produces two failure modes. The obvious one is delay: you wait for the last holdout, and the holdout has learned that holding out works. The subtler one is dilution — the decision that finally emerges has been sanded down to whatever nobody objects to, which is usually the least interesting option available.
The alternative isn't autocracy. It's clarity about who decides. For every significant decision, three things should be explicit before the discussion starts: who owns the decision, who must be consulted, and by when. Everyone gets input. One person decides. Everyone executes.
Stated up front, this feels almost rude. Practiced consistently, it's the single largest speed unlock available to most leadership teams — because the people who were consulted stop fighting the outcome, having genuinely been heard.
Where the time actually goes
When a decision takes twelve weeks, the deliberation is rarely twelve weeks of thinking. It's usually:
- Two weeks of nobody being sure it's their call
- Three weeks waiting for a recurring meeting where it can be raised
- Two weeks gathering data that won't change the answer
- Three weeks of a stakeholder being unavailable
- Two weeks of re-litigating with someone who joined the conversation late
Almost all of that is structural, not intellectual. Which is good news: structural problems are fixable. A weekly decision forum with a standing agenda, a named owner per item, and a default of "decide or explicitly defer with a date" eliminates most of it within a month.
The information you're waiting for usually doesn't exist
There is a particular kind of paralysis where a team is waiting for data that would resolve the disagreement. Often, that data does not exist and cannot be obtained — because the question is about the future, and the future has not happened.
A useful intervention: ask the room what specific piece of information would change their mind, and how it would be obtained. If nobody can answer, the disagreement is not about facts. It's about judgment, risk tolerance, or something unsaid. Naming that is faster than another analysis.
What speed costs, honestly
Deciding faster means being wrong more often in absolute terms. That's the trade, and pretending otherwise is dishonest.
The reason it's still correct is that reversible mistakes are cheap and the learning compounds, while delay is a cost you pay with certainty and get nothing for. A team that makes twelve decisions a quarter and gets three wrong will outrun a team that makes four and gets all four right — because the first team has learned twelve things.
The discipline is in maintaining the distinction. Fast on the reversible, patient on the irreversible, and never confused about which is which.
If decisions in your company are taking longer than they should and you can't quite say why, that's usually a structural problem with a structural fix.