Skip to content

Early-warning system

An early-warning system is an arrangement that surfaces a developing problem while the response options are still open. In organisational terms it means watching the things that move before the numbers do: decision latency, escalation volume, dependency slippage and where attention is actually going.

Early information restores options; after commitment the options have expired and the same information is an autopsy.

Why it matters when the plan changes

Financial reporting tells you about last quarter. By the time a bottom-line miss appears the causes have been running for two or three quarters, and the responses available then are expensive. Watching leading indicators buys back the interval in which cheap responses still work, which is the only interval that matters. Cross-unit slippage is often the earliest reliable signal precisely because it depends on colleagues managers cannot fully rely on, a weakness Sull, Homkes and Sull identified in Harvard Business Review.

The tension is that weak signals are ambiguous by definition. A system tuned to catch everything produces alarms nobody acts on, and one tuned to avoid false alarms catches things too late to matter. The useful design attaches a proportionate response to each level of signal rather than treating every alert as a crisis. This is the graduated response Ansoff proposed: a weak signal earns a question, not an escalation, because strategic information arrives too late if treated only at full strength.

In practice

A programme's decision latency doubles over two months and escalations to the sponsor triple. Neither appears in any report, because neither is measured. The milestone miss appears in the quarterly review four months later, by which time the available responses are descoping or slipping the date.

Evidence

What it cannot tell you

An early-warning system flags where attention should turn, but it cannot tell you what the developing problem actually is or how severe it will become. Weak signals are ambiguous by design; the same rise in decision latency can precede a minor stall or a major failure, and the system offers no way to tell which in advance.

Questions

Decision latency, escalation volume, how often decisions reopen, slippage on cross-unit commitments, and where senior attention is actually going. Sull, Homkes and Sull found in Harvard Business Review that only 9% of managers can rely on colleagues in other functions and units all the time, which makes cross-unit slippage an early and reliable signal.

Because financial results are the last thing to move. A bottom-line miss begins as a clarity miss and compounds for two or three quarters before it reaches the number, by which point the cheap responses have expired and only expensive ones remain.

By attaching a proportionate response to each level of signal rather than treating every alert as a crisis. A weak signal should trigger a question, not an escalation. Systems that escalate everything are ignored within a quarter and become worse than nothing.

Whoever owns the consequence, because only they can act on a signal in time to matter. Wikipedia's 2026 entry on early warning systems defines the arrangement precisely as one for detecting a developing situation while response is still possible, which is a duty, not a reporting task.

A question asked and an answer sought, not a programme launched. Most weak signals resolve into nothing. The value lies in looking early and cheaply, which is only sustainable if the act of looking does not commit anyone to a large response.