Calm Frequencies: A Broadcast Discipline for Incident Communications
Photo: Christopher Michel, CC BY 2.0, via Wikimedia Commons
Incident communication often fails not because the facts are unknown but because the rhythm is wrong. Updates arrive in bursts, then stop. Severity labels mean different things to different teams. Stakeholders ask for certainty while engineers are still discovering the shape of the problem. The Shipping Forecast offers a better model: a familiar cadence, plain language, area-specific warnings, and forecast windows that make uncertainty explicit without causing alarm.
The forecast is not dramatic. It is disciplined. It tells listeners what matters, where it applies, how strong it is likely to be, and when the next update will arrive. Software teams can borrow that discipline for status pages, incident channels, and executive updates.
The Discipline of Saying the Same Thing at the Same Time
A shipping forecast is broadcast on a schedule. The schedule is part of the service. Users know when to listen and what they will hear. Incident communications should work the same way. During a significant incident, set a cadence early: for example, an update every thirty minutes for the first two hours, then hourly until resolution. Publish the cadence in the status page and the incident channel.
The cadence should not depend on new information. If there is nothing new, say so. Silence is not calm; it is ambiguity. A short update that confirms the position is better than a long update that speculates.
Plain Severity, Specific Areas
The Shipping Forecast uses sea areas. It does not say that the whole country will be windy; it says where the gale is expected. Incident updates should name the affected service, region, customer segment, and journey. Avoid global statements when the impact is local.
Severity language should be plain and stable. Define what 'major' means in terms of customer impact, data integrity, and safety. Publish the definition. Then use it consistently. Stakeholders should not need a decoder ring to understand whether a payment outage is more serious than a search delay.
Forecast Windows, Not Heroic Silence
Engineers often delay communication until they have a fix. That instinct is understandable and usually wrong. The forecast model separates the warning from the resolution. It says what is happening, what is likely, and when the next forecast will be issued. Incident updates can do the same: state the current impact, the working hypothesis, the next action, and the next update time.
This is not an admission of failure. It is operational maturity. Stakeholders can plan around a known window. Support teams can prepare scripts. Executives can decide whether to escalate. The alternative, a heroic all-hands update at the end, leaves everyone else guessing in the meantime.
Before: Set the Frame
Before an incident, publish the communication contract. Define severity levels, update cadence, ownership, and channels. Make the status page the canonical source. Include a short template that any incident commander can use.
A before-incident template might include: service and journey affected; severity and definition; start time; current customer impact; known workaround; incident commander; next update time. Keep it short enough to complete under pressure.
During: Cadence Over Drama
During the incident, use the template repeatedly. Lead with impact, not with the chronology of engineering effort. Name the area. Give the next update time. Avoid blaming individuals or vendors. Avoid promising a fix time unless it is genuinely known.
A during-incident update might read: 'Card payments are failing for approximately 8 per cent of UK transactions. The cause is under investigation. Customers may see a declined message. We will update at 14:30.' That is calm, specific, and useful.
After: The Settled Summary
The Shipping Forecast does not end with a dramatic conclusion. It moves to the next area and the next period. Incident communications should close with a settled summary: what happened, who was affected, what changed, what remains uncertain, and when the review will be published.
The after-incident update should not be a victory lap. It should be a record. Link to the post-incident review, state the remediation owners, and give a date for the next update. Then keep the cadence until closure.
Calm communication is not a soft skill. It is a control. It reduces duplicate questions, steadies stakeholders, and gives engineers room to work. The Shipping Forecast has survived for decades because it is predictable, specific, and honest about uncertainty. Incident communications can aspire to the same standard.