A fault takes out service across part of a region at eight in the morning. Within twenty minutes the support operation receives more contacts than it normally handles in a day.
The status page says investigating. It still says investigating at two in the afternoon. Every one of those customers contacts again, because the page has given them no reason not to.
Why outages break support operations specifically
Demand is not elevated, it is multiplied. No staffing model survives a hundredfold increase over an hour, and the contacts are almost entirely identical.
Why customers contact rather than reading
Because the published information does not answer their question. They do not want to know whether there is an incident, they want to know whether theirs is part of it and when it ends.
Why investigating is the worst possible status
It conveys no information, sets no expectation and gives somebody every reason to check again in twenty minutes and then contact you.
Why an optimistic estimate is worse than none
An estimate that passes without restoration produces a second wave of contacts angrier than the first. Under promising is the only workable approach.
Fix one: confirm the fault by location immediately
A postcode or account lookup that says yes, your area is affected. This single capability removes the majority of contacts.
Fix two: give a restoration estimate and honour it
Even a wide one. Customers plan around an estimate, and the absence of one is what produces repeat contact.
Fix three: update on a schedule and say when the next update is
By two o'clock, and then at two o'clock regardless of whether anything has changed. Predictable updates stop people checking constantly.
Fix four: distinguish a local fault from an area outage
Customers with an individual fault during an area outage are told to wait and never get fixed. Separating the two streams matters enormously.
Fix five: offer registration for updates
A notification when service is restored. Customers who have one stop contacting, which is the entire objective.
Fix six: give the diagnostic steps for individual faults
Router restarts, cable checks and the obvious things, so that the genuine individual faults are filtered from the ones that are not.
Fix seven: handle the vulnerable customer register properly
Customers dependent on connectivity for medical or safety reasons need prioritising and a different contact route, which requires the care described in how to handle sensitive questions in your chatbot.
Fix eight: answer the compensation question upfront
Whether any applies, how it is calculated and whether it is automatic. It is the second wave of contacts after restoration and it is entirely predictable.
Fix nine: keep the billing questions separate
Outage contacts and billing contacts arrive together and need different handling, which is the problem in service outage and billing questions that swamp support.
Fix ten: keep the switching enquiries flowing
An outage drives customers to consider leaving and new customers to enquire, which is the traffic described in switching and new connection questions for providers.
Fix eleven: borrow from equipment fault triage
The same problem of separating a real fault from a user fixable one appears in integrated appliance questions that complicate repairs.
Fix twelve: absorb the volume automatically
An assistant that confirms the outage by location, gives the current estimate, registers a customer for restoration updates and walks individual faults through diagnostics handles the identical contacts at a scale no team can. See features, start at try, and the sector view sits under b2b.