Reporting a Power Outage When the Utility App and Home Wi-Fi Are Both Down

The outage app is the tool most people reach for when the lights go out, and it is the one least likely to answer. That is not bad luck. An app or an outage map is the reporting path with the most moving parts, and every one of those parts is under stress at exactly the moment it is needed.

The fix is an ordering, not a repair. Report by phone or by text first, use the app only to check status afterwards, and treat the outage map as the last thing to consult, not the first. Nothing about that ordering depends on which utility serves the address, which is why it is worth setting up before a storm rather than during one.

The reporting order that works while the app does not

1. Stop reloading the map. A map that spins is telling you the servers are saturated. Repeated reloads add to that load and change nothing about the report.
2. Call the utility's outage line from the number on the bill. Utility outage lines are typically automated, matching a caller ID or an account number to a service address, so a report can usually complete without a human answering.
3. If the utility runs a text-message reporting service, send the outage keyword. Enrolment usually has to happen in advance, which is the reason this belongs on a pre-storm checklist rather than a storm-night one.
4. Note the confirmation, then stop. A reference number or a return text is the whole point of reporting. Filing again in five minutes does not raise priority.
5. Check status on cell data, not home Wi-Fi. The home router lost mains power at the same instant the lights did. Anything that says "connected, no internet" on a phone is describing a router that is dark, not a fault worth troubleshooting.

Point five is the one that sends people down the wrong path. A phone showing a dead Wi-Fi network during an outage is behaving correctly. There is nothing to reset, because the thing to reset has no power.

Why the app fails first

An outage app is not weak software. It is software that happens to sit at the end of the longest dependency chain of any reporting path available.

Three of those dependencies fail together in a storm, and they fail for reasons that have nothing to do with the app's code:

• Home broadband goes down with the power. A router, a cable modem and a fibre terminal all run on mains electricity. Battery backup for that equipment exists as an option in some installations, but it has to have been fitted and kept charged in advance. In practice the phone falls back to cell data the moment the lights go.
• Cell service is being restored at the same time as the grid. This is not speculation about congestion. The Federal Communications Commission runs the Disaster Information Reporting System, described by the agency as "a voluntary, web-based system through which the Commission collects operational status and restoration information from communications providers during major disasters and subsequent recovery efforts." The system exists because wireline, wireless, broadcast, cable, interconnected VoIP and broadband networks are themselves damaged in these events.
• The map is the heaviest thing the utility serves. An outage map is a live geographic dataset requested simultaneously by everyone in the affected area plus everyone watching from outside it. The phone system carries one short transaction per customer.

One common claim is worth declining. It is often said that a text message will get through when a call will not. No federal source consulted for this article publishes a general rule to that effect, and whether a given message succeeds depends on network conditions no customer can observe. The defensible reason to prefer a call or a text over the app is simpler and does not require a claim about radio behaviour: fewer things have to be working.

The smart meter question, and why a report can still be needed

Modern meters do change the picture. According to the U.S. Energy Information Administration, U.S. electric utilities had roughly 119 million advanced metering infrastructure installations as of 2022, about 72% of all electric meter installations. EIA defines AMI as metering that "measure[s] and record[s] electricity usage at a minimum of hourly intervals" and provides that data "to both the utility and the utility customer at least once a day," with the more capable units offering two-way communication.

EIA has also written that this technology lets utilities obtain "detailed outage information in the event of a storm or other system disruption, helping utilities restore service to customers more quickly."

What that does not establish is that every meter alarms the moment power is lost, or that an alarm reaches the utility while the local communications path is itself damaged. AMI, as EIA defines it, is interval metering with a daily data obligation. Outage notification capability sits above that baseline and varies by deployment. Treating "there is a smart meter on the wall" as "the utility already knows" is an assumption, not a fact about the meter.

Two situations make a customer report worth filing regardless:

• The outage stops at the property line. If neighbouring homes are lit and this one is not, the fault is narrow. Narrow faults are the ones least likely to stand out in a system-wide picture built from thousands of simultaneous events.
• No restoration estimate has appeared for the address. An account with no estimate attached to it after the utility has begun publishing estimates elsewhere is worth a call, if only to confirm the address is on the list.

Before calling in the narrow case, check the main panel. A tripped main breaker produces exactly the same experience as a utility outage and is entirely on the property side of the meter.

What the national numbers say about storm days

EIA collects distribution reliability metrics from utilities and publishes them by year. The two headline measures are SAIDI, the System Average Interruption Duration Index, defined by EIA as the minutes of interruption the average customer experienced over the year, and SAIFI, the System Average Interruption Frequency Index, the number of non-momentary interruptions the average customer experienced. A third, CAIDI, is described by EIA as the "average number of minutes it takes to restore non-momentary electric interruptions."

Each is published twice: once counting every event, and once with major event days removed under the IEEE classification EIA applies. The gap between those two columns is the entire subject of this article.

U.S. total, EIA	2023	2024
SAIDI, all events	366.6 minutes	662.6 minutes
SAIDI, major event days excluded	123.9 minutes	131.6 minutes
SAIFI, all events	1.348 interruptions	1.531 interruptions
SAIFI, major event days excluded	1.022 interruptions	1.065 interruptions
CAIDI, all events	271.8 minutes	432.7 minutes

Three readings follow directly from those figures.

• Ordinary years are stable; storm years are not. With major event days removed, SAIDI moved from 123.9 to 131.6 minutes between 2023 and 2024. Counting everything, it moved from 366.6 to 662.6.
• Storm days carry most of the outage time. In 2024 the difference between the two SAIDI columns is 531 minutes, close to nine hours per average customer in a single year.
• Average restoration time roughly moves with the severity of the year. CAIDI counting all events was 432.7 minutes in 2024 against 271.8 minutes in 2023 — about seven hours versus four and a half.

None of this predicts any individual restoration. It does explain why a restoration estimate behaves differently on a storm day than on an ordinary Tuesday, and why the app is under a load it never sees the rest of the year.

Reading a restoration estimate without over-reading it

An estimated restoration time cannot exist before the damage behind it has been assessed. That is a constraint rather than a policy choice: a number cannot be produced for work that has not been surveyed. On a large event, the practical consequence is that an early estimate is either absent or broad, and that it is revised as assessment catches up.

What is not claimed here is how any particular utility calculates or updates that number. Methodology and update frequency are set utility by utility, generally under state commission rules, and are not published to a single national standard. Any article that tells you a specific company's restoration policy without citing that company's own filing is guessing.

When the communications network is itself part of the outage

If calls fail as well as data, the problem has moved past the utility. During major disasters the FCC activates DIRS and, in the agency's words, "announces DIRS activations through public notices and emails to DIRS participants," with announcements that "list the counties covered in the activation and provide reporting and contact information."

Two details of that system matter to a customer. Participation is voluntary for most providers — the FCC notes that reporting is mandatory only "for all Stage 2 recipients of the Uniendo a Puerto Rico Fund and the Connect USVI Fund" — and the status information filed is "presumed confidential," with the FCC compiling aggregated data for federal emergency management officials. So a DIRS activation covering the county is a signal that communications infrastructure in the area is being tracked and restored, but it is not a per-address status page and should not be treated as one.

Practically: if voice, text and data all fail from one location, the next step is a different location or a different network, not further attempts on the same handset.

Set the non-app path up before it is needed

• Save the outage number as a contact. Take it from a paper bill or the account page, not from a search result during a storm.
• Record the account number somewhere offline. Automated lines that cannot match a caller ID will ask for it.
• Enrol in text reporting in advance if the utility offers it. Enrolment typically requires a working data connection, which is the one thing an outage removes.
• Charge a power bank, not the router. Keeping a router alive on battery is pointless when the upstream line has no power either. Phone battery is the resource that matters.
• Know where the main panel is and what its normal state looks like, so the property-side check takes seconds.

Safety limits, in official wording only

Two hazards during outages are not troubleshooting questions and should not be treated as such. CDC guidance states plainly: "Never touch a fallen power line," and avoid driving through standing water where downed lines are present. For generators, CDC advises never running a gasoline-powered engine indoors and keeping a generator at least 20 feet from windows, doors and vents, with a battery-powered carbon monoxide detector in the home; symptoms of CO poisoning include headache, dizziness, weakness, nausea and chest pain, and the guidance is to call 911 or Poison Control at 1-800-222-1222 immediately.

On food, CDC gives a refrigerator roughly 4 hours unopened and a full freezer about 48 hours, with perishables discarded after four or more hours without power.

When This Doesn't Apply

• Municipal utilities and rural cooperatives may have no app at all. Where the only reporting channel is a phone line, there is no ordering problem to solve. The prep checklist still applies; the comparison does not.
• Planned de-energisation is not an outage report. Where a utility has publicly de-energised lines ahead of a fire or wind event, filing an outage report adds nothing. Status for those events is normally published separately from the storm outage map.
• Prepaid electricity accounts behave differently. Where service is prepaid, loss of power can reflect an account balance rather than a network fault, and the reporting path is the billing channel, not the outage line.
• A whole-building outage in an apartment is the building's problem first. If the meter serving the property belongs to the building rather than the unit, the report goes to building management, and a customer-level report will not match any service address the utility recognises.
• If the property line check shows a tripped main breaker, none of the above applies. That is inside the premises, and repeated tripping is an electrical fault to be examined by a licensed electrician rather than reported to the utility.
• Life-support equipment changes the priority order entirely. Where a household depends on powered medical equipment, the relevant procedure is the utility's medical or critical-care registration and the local emergency number, not a reporting-channel comparison.

The short version

The outage app is not broken in any way that a reinstall would fix. It is the reporting path with the most dependencies, and during a storm several of those dependencies fail at once — starting with the home router that lost power at the same moment the lights did. Call or text the report in, note the confirmation, and use the app afterwards for status rather than beforehand for filing. The national reliability figures explain why the difference only shows up on the days it matters: with major event days stripped out, the average U.S. customer's annual interruption time in 2024 was 131.6 minutes. With those days counted, it was 662.6.