To avoid analysis paralysis, sort the decision before you analyse it. My rule is simple: if it can be reversed cheaply, decide this week. If it is a genuine one-way door, name the evidence that would change your answer, set the date you will have it and decide on that date. Most stalled decisions are not hard. They are uncomfortable to own.
I wrote the first version of this for the Forbes Coaches Council in 2022, and it still describes a block that CEOs and executive teams run into. The block is rarely the spreadsheet. Every stalled decision I've ever coached someone through had enough information behind it to act. "More analysis" has become the most socially acceptable way of not being the person who made the call. One more report, one more workshop, one more round of stakeholder input. Strip away the vocabulary of rigour and due diligence and what is often left is a leader who does not want to be the person who was wrong.
There is a structural reason as well. Value now sits across several business units and functions, so every decision seems to need sign-off from people who were not in the room where the problem was understood. That complexity is real, and it is the reason my advice rests on structure more than on personal courage. It still does not explain most stalls. Most stall because the person with the authority to decide is protecting themselves, not the organisation.
What the research says about how to avoid analysis paralysis
The cost of delay is better documented than most leaders assume. McKinsey's work on decisions reports that executives spend nearly 40 per cent of their time making decisions and believe most of it is poorly used. In a survey of more than 1,200 business leaders, fewer than half said decisions in their organisation were timely, and 61 per cent said at least half of the time they spend deciding is ineffective. The same research puts the cost of inefficient decision making for a typical Fortune 500 company at 530,000 days of managers' time a year, equivalent to about $250 million in annual wages (McKinsey, 2023). A smaller company will not reach those numbers, but the shape of the loss is the same: it is time that people spend waiting.
More choice does not rescue the situation. In experiments by Iyengar and Lepper, people offered a limited array of six options were more likely to buy, and reported greater satisfaction afterwards, than people offered 24 or 30; students given a limited set of essay topics also wrote better essays (Iyengar and Lepper, 2000). Schwartz and colleagues found that maximisers, people who need the best possible option, were less happy, less satisfied with their choices and more prone to regret than people who settle for good enough (Schwartz et al., 2002). Executives who are maximisers by temperament will feel every unresolved question as a risk, and the organisation pays for that feeling.
McKinsey's guidance for fast-moving conditions makes a point I quote often: when the environment is uncertain and defined by urgency and imperfect information, waiting to decide is a decision in itself (McKinsey, 2020). Delay has a cost, and somebody carries it. In most organisations it is the people one level down who were waiting for the answer.
Five questions that separate inquiry from stalling
When a leader keeps asking for more input, I stop the process and ask five questions. They separate real inquiry from delay, and they make the cost of a late decision as visible as the cost of a wrong one.
- Has the answer changed in the last two rounds of input?: If two rounds of data, workshops or stakeholder reviews have not moved the recommendation, you are no longer gathering information. You are stalling with better paperwork.
- Who is this extra review protecting?: Each additional sign-off layer protects someone from being the sole owner of the call. Name who. If the answer is you, that is the real reason for the delay.
- What does a wrong-but-fast decision cost against a right-but-late one?: Most leaders price only the cost of being wrong. Price the cost of being late too: lost momentum, disengaged teams, competitors who moved first. It is usually the larger number.
- Would the person closest to the problem make this call today?: If your subject-matter expert already has a clear recommendation and you are still deliberating, the gap is not information. It is your comfort with letting their judgement stand.
- Is the decision reversible?: Reversible decisions deserve speed, not ceremony. Keep the long process for the small number of calls that are genuinely one-way.
Sort the decision first, two-way door or one-way door
Jeff Bezos set out the cleanest sorting rule I know in Amazon's 2015 shareholder letter. Some decisions are consequential and nearly irreversible, one-way doors, and should be made methodically with deliberation and consultation. Most are changeable and reversible, two-way doors, and "can and should be made quickly by high judgment individuals or small groups" (Amazon, 2015 letter). He warned that as organisations grow they tend to apply the heavyweight process to everything, producing slowness, unthoughtful risk aversion and diminished invention. A year later he added that most decisions should probably be made with around 70 per cent of the information you wish you had, because waiting for 90 per cent usually means being slow (Amazon, 2016 letter). The same letter offers a phrase worth borrowing: disagree and commit. And it names the failure that ends many stalled debates: without early escalation, the default way a deep disagreement ends is exhaustion, and whoever has more stamina carries the decision.
| Decision type | Who decides | Process | Deadline rule |
|---|---|---|---|
| Reversible (two-way door) | The person or team closest to the work | Light: a short brief and one conversation | Decide this week; review the result in days or weeks |
| Cross-cutting, frequent and risky (for example pricing) | A cross-functional forum with a named chair | Clear objectives, measures and targets before debate | Fixed cycle; the chair closes it |
| Irreversible or big bet (one-way door) | Top leaders and the board | Deliberate; assign someone to argue each side | Name what evidence would change the answer, set the date, decide on it |
The categories draw on McKinsey's split of big-bet, cross-cutting and delegated decisions and on Bezos's two doors. The table is my synthesis for executive teams, and the discipline is the same in each row. Someone owns the call, and the deadline is a date rather than a feeling.
Give ownership to the people closest to the problem
In my experience a person with real authority over an outcome makes faster, sharper trade-offs than a person waiting for permission. McKinsey notes that delegated decisions are the most frequent and the least understood, with only about a quarter of surveyed organisations reporting high-quality, speedy delegated decisions. The advice it gives is practical: put responsibility firmly in the hands of those closest to the work, and clarify who has a vote and who has a voice (McKinsey, 2023).
I apply one test. When a team has debated a challenge properly, weighed the options and arrived at a recommendation, and the recommendation is sent back with a request to explore a few more options for no reason anyone can state, that is a leader who has not decided whether they trust the team. The kinder reading is that they are hiding indecision behind a request for analysis. Trust the recommendation or challenge it with a specific question. Then make the final call and own it publicly so everyone downstream knows where they stand. The work I do on executive team decision making and on honest upward information both start here.
A telecommunications case study on my work-with-me page shows what happens when ownership moves. A founder-dependent company at $15M, where every operational decision escalated to the CEO, installed decision rights and a leadership maturity framework. The published results are operational decisions fully delegated within 90 days and the founder freed for more than 20 hours a week of strategic work. These are my own figures from my site. If the same bottleneck sounds familiar, the Decision Bottleneck Challenge is a quick way to locate it.
Fail fast still needs judgement
"Fail fast" is thrown around as though speed alone were the virtue. A fast decision made without a way of noticing you were wrong is recklessness with better marketing. The leaders who do this well add one habit: before they act, they say out loud what result would tell them to reverse course, and by when they expect to see it. Pre-committing to the evidence that would change your mind is what separates disciplined speed from impulse. Without it, fail fast becomes permission to avoid thinking, which is a different route to the same paralysis.
Fail fast also gets misapplied to calls that were never reversible: a senior hire, a public commitment to a client, a restructure that changes reporting lines. Treat those with the tempo of a pricing test and you get burned, then conclude wrongly that moving fast is dangerous. Moving fast on the wrong category of decision is dangerous. The skill is sorting the pile correctly, and most leaders never do it explicitly, so every decision defaults to the slowest process available.
The inverse rule of analysis and urgency
In my observation the amount of analysis a decision receives is often inversely related to how urgently it needs making. Low-stakes, reversible calls collect endless workshops because nobody fears them. High-stakes, genuinely hard calls get rushed because sitting with them is unbearable. Naming that pattern in a room is usually enough to prompt laughter of recognition, and then a rethink. This is a pattern I have noticed rather than a measured finding, so test it against your own last ten decisions.
It matters for retention as well. People tolerate a leader who makes a wrong call far better than a leader who makes none. A wrong decision can be corrected. An absent one sits in the plan and rots it. The strongest people notice fastest that their recommendations go into a black hole, and they are the ones with other options.
A rule you can run this week
The rule I opened with is simple: reversible calls this week, one-way calls on a named date. The part people resist is the date. Once a one-way decision has a date, decide on it: not the date after, and not when the team feels ready, because readiness is a feeling and not a fact. Ask the room plainly what the delay is protecting. A six-month stall often ends in the next ten minutes.
Analysis paralysis is therefore a courage problem dressed as a thinking problem, and a design problem as well. The design half is yours to fix. Write down who owns which decisions, sort them into doors, and give each a date. For the structural side, my pieces on decision rights and on how today's leaders can make the best decisions carry the detail, and CapabilityAI brings the same decision log to the time between conversations. To work through it with your own leadership team, start here.