Management latency

The elapsed time between something changing in a business and the person able to act on it knowing about it. Most companies measure system uptime to the second and leave management latency unmeasured.

The part that costs money

Engineering measures latency to the millisecond and pages a human when it degrades. The same company will let eight days pass between a deal slipping and anyone deciding what to do about it, and no one will call that an outage, because nothing in the management layer is instrumented.

The measurement is four timestamps and most companies can produce none of them: when the thing changed, when a person able to act knew, when a decision was made, and when the action shipped. The gaps between them are where the cost lives, and each gap has a different fix — the first is a noticing problem, the second is a meeting-cadence problem, the third is an ownership problem.

Cadence is the usual culprit and the usual defence. A business that changes continuously is reviewed weekly, which sets a floor on latency that no amount of dashboard freshness can lower. The warehouse being current to the hour does not help if the only moment anyone looks is Friday.

What it looks like

The deal slips on Tuesday. It surfaces in the forecast call on Friday. The response is agreed the following Monday and ships Wednesday. Management latency: eight days. The same company alerts on 400ms of API latency at three in the morning.

How you would actually measure this

• The four timestamps, per incident: changed, known, decided, acted. Start with ten recent examples; the median is enough.
• Review cadence as a hard floor — a weekly meeting cannot produce latency below about seven days for anything it owns.
• Share of material changes first noticed by a person rather than raised by a system.
• The gap between data freshness and human freshness, which is usually where the surprise is.

What it is not

• It is the same as data freshness. — The warehouse can be current to the minute while the person able to act finds out on Friday. Freshness is a property of the pipeline; management latency is a property of the organisation.
• It is the same as reporting frequency. — A weekly report can carry a fact that was already three days old when it was compiled. Frequency sets a floor on latency, it does not measure it.
• Faster meetings fix it. — Meeting cadence only addresses the gap between known and decided. If nothing noticed the change in the first place, moving the meeting earlier changes nothing.

Related terms

• Organizational attention (https://www.themeetpatel.com/glossary/organizational-attention) — The finite capacity of a company to notice, prioritise and act on what is actually happening inside it. Storage buys none of it, and most tools spend it.
• Decision debt (https://www.themeetpatel.com/glossary/decision-debt) — The accumulated cost of decisions a company deferred because the information required to make them was expensive to assemble. It compounds quietly, and is usually repaid during a crisis.
• Autonomous decision intelligence (https://www.themeetpatel.com/glossary/autonomous-decision-intelligence) — Autonomous decision intelligence is a system that connects a company’s systems, reconciles where they disagree, monitors what matters, investigates what changed, and puts an evidence-backed decision in front of a human — then keeps watching whether the call worked.