Decision Latency: The Silent Tax on Every Good Idea
Decision latency is the silent tax most organisations pay without seeing it on any invoice. It is the gap between the moment a decision could be made and the moment it actually is, and in the companies we study it consumes weeks. A pricing change sits in review for a month. A hiring decision waits for one more meeting. A customer escalation bounces between two managers, each waiting for the other to own it. Nothing is broken, everyone is busy, and the organisation is slower than its smartest competitor by exactly the amount of time its decisions spend waiting.
This piece unpacks where decision latency comes from, why it grows as companies succeed, and what leaders can change structurally, in a quarter, to shorten it. The argument draws on decision science and on patterns observed across South African and international firms, and it ends with a measurement any team can start using this week.
What Decision Latency Costs
Latency compounds quietly because each individual wait looks reasonable. A week of review here, a follow-up meeting there. Multiply across every decision in a quarter and the effect is dramatic: product launches timed to a market that has moved, candidates lost to faster competitors, customers churned while an answer was pending. Research on organisational decision-making has repeatedly found that the speed of a decision predicts outcomes better than its predicted accuracy, fast deciders who adjust outperform slow deciders who optimise, because feedback arrives sooner and more often.
The cost is not only external. Slow decisions consume the organisation’s scarcest resource twice: once in the waiting, and again in the re-work of re-briefing, re-contextualising, and re-litigating after the delay. Teams learn that decisions evaporate, so they hedge, and hedged teams ship hedged work.
Where Latency Actually Comes From
Four structural sources explain most of it, and none of them is laziness.
Undefined decision rights. Nobody is sure who decides, so everyone waits for consensus that never arrives, or the decision escalates to the top because that is the only path everyone trusts. In practice the highest-paid person’s opinion wins by default, and the queue for their attention is the queue your decision joins.
Consensus culture. Organisations that grew through collaboration often institutionalise the belief that everyone must agree before anything moves. Consensus is appropriate for values and irreversible choices. For reversible ones it is a tax, and most operational decisions are reversible. Jeff Bezos’s distinction between one-way and two-way doors captures it: treat a reversible decision like an irreversible one and you have spent irreversible time on it.
Information hoarding as risk management. Waiting for complete information feels prudent, but the marginal value of another data point falls while the cost of delay stays constant. Most decisions in a business are made with 70% of the information, and the last 30% changes the answer less often than anyone expects.
Meeting-mediated decisions. When the only legitimate place a decision can happen is a meeting, decisions happen at the cadence of calendars, not at the cadence of need. Two-week meeting gaps become two-week decision latency, automatically.
The Structural Fixes
Latency responds to structure, not exhortation. Four changes shorten it measurably within a quarter.
1. Name a decider for every decision type
Not every decision, every decision type. Pricing, hiring, architecture, brand, incident response: each gets a named owner with a documented boundary of what they may decide alone, what needs consultation, and what escalates. The Amazon framework is the cleanest version: disagree and commit, where the decider consults, decides, and dissenters commit without re-opening. Consultation is not consensus, it is input with a deadline. Write the decision rights down; unwritten decision rights are a rumour, and rumoured authority produces the slowest decisions of all.
2. Time-box reversible decisions
Classify decisions by reversibility, then set a clock. Two-way-door decisions get days, not weeks, and the decider is empowered to move at 70% information. The discipline that makes this safe is a written record: what was decided, what was known, what would change the answer. Reversible decisions made fast and recorded honestly are how organisations learn, because they generate feedback loops that committees never see.
3. Replace decision meetings with decision documents
Move the context before the meeting: a one-page decision memo circulated in advance, with the options, the recommendation, and the decider named. The meeting, if it happens at all, reviews rather than hears. Amazon’s six-page narrative memos and silent reading openings exist for exactly this reason. The meeting stops being the place where people first encounter the problem, which is where half of every traditional meeting’s time goes.
4. Measure latency like you measure delivery
You cannot fix what you do not measure, and almost nobody measures decision latency. Start simple: log decisions in a shared register with two dates, when the need was identified and when the call was made. Review the distribution monthly. The teams that do this discover their latency is concentrated in a handful of recurring decision types, and fixing three types often removes most of the tax. What gets measured gets managed applies to decision flow as much as to code flow. Our earlier piece on why smart teams make slow decisions examines the psychology behind the queues this log will expose, and the meeting that should have been a decision is the companion failure mode. McKinsey’s research on organisational decision-making quantifies the payoff of faster, more distributed decisions, and Herbert Simon’s work on bounded rationality remains the intellectual foundation for the 70%-information standard above.
Where Speed Must Not Win
Latency reduction has a boundary, and mature leaders know where it sits. Decisions that are irreversible, that carry legal or safety consequences, or that determine values deserve the slow path. Hiring a leader, entering a market, changing a security posture, these warrant full information and genuine consultation. The discipline is not “decide everything fast”. It is deciding deliberately which decisions deserve which speed, and refusing to let slow-path habits leak into fast-path choices. Companies confuse the two in both directions: they deliberate over logo colours for a month and hire executives after two interviews.
A 30-Day Starter Plan
Four weeks, one hour of leadership time each. Week one: list the ten decision types that recur most in the organisation, and name a decider for each with their consultation boundary. Week two: pick the two highest-latency types, and time-box them, reversible ones get a 5-working-day clock from proposal to call. Week three: switch those decision meetings to the document-plus-review format, context circulated 48 hours ahead. Week four: start the decision log and hold the first latency review. Organisations that run this cycle once rarely go back, because the feeling of decisions moving at the speed of need is habit-forming.
Frequently Asked Questions
What is decision latency?
Decision latency is the elapsed time between when a decision could be made and when it is made. It excludes the time spent actually analysing; it measures the waiting, the unclear ownership, the calendar gaps, and the consensus loops. In most organisations it is the largest untapped source of speed, and unlike headcount or budget, shortening it costs nothing.
How do you decide fast without being reckless?
Classify by reversibility. Reversible decisions get time-boxes and a 70%-information standard, with the reasoning recorded so the choice can be revisited. Irreversible, legal, and safety decisions keep a deliberate process. Recklessness is not speed but speed applied to the wrong decision class, and most organisational decisions are more reversible than their owners believe.
Who should be the decider when ownership is unclear?
The person closest to the consequences, with authority proportional to the blast radius, is the default. Where stakes cross teams, name a single accountable decider with a consultation list and a deadline. Shared ownership is the structural cause of latency, and an explicitly named decider with a documented boundary resolves it faster than any process redesign.
Does faster decision-making reduce quality?
Not when the speed comes from removing waiting rather than removing thinking. Studies of organisational decisions find that faster deciders tend to make better long-run choices because they act on feedback sooner and correct more often. Quality lives in the feedback loop, and latency starves the loop. The exception is irreversible decisions, which keep their deliberation.
How do we get consensus without slowing down?
Replace consensus with consultation: everyone with material input is heard before the call, then the named decider decides and the team commits. Disagree-and-commit is the working compromise, dissent is voiced and recorded, then executed faithfully. True consensus belongs on values and irreversible strategy; for everything else it is latency wearing a virtue’s clothing.
The Takeaway
Decision latency is the tax your organisation pays on every good idea, and it is payable in structure rather than salary: name deciders by decision type, time-box the reversible ones, move context into documents so meetings decide rather than brief, and measure the gap between could-decide and did-decide until the distribution shrinks. The organisations that feel fastest are not the ones with the best opinions. They are the ones whose decisions spend the least time in queues, and whose people trust that a call, once made, will hold. Start with the decision log this week. The first month’s latency review will tell you exactly where your quarter has been going, and the second one will show it shrinking.
