Every electric utility in the country measures its own reliability with the same two figures, publishes them, and almost no customer has ever looked them up. They are called SAIDI and SAIFI, and between them they answer the question any backup-power decision actually rests on: how often does the power go out here, and for how long.
Reading the Scorecard
SAIDI is the System Average Interruption Duration Index. It is the total outage time across every customer in a year divided by the number of customers, which gives the average minutes a customer spent without power. SAIFI is the System Average Interruption Frequency Index, the same arithmetic applied to the number of interruptions rather than their length. Divide one by the other and you get CAIDI, the average length of a single outage.
Scale matters for reading them. The Energy Information Administration puts the US baseline SAIDI near 2 hours a year. Include major event days, meaning storms and the other things utilities are allowed to exclude, and the figure climbs to somewhere between 3.5 and 8 hours. That gap between the two versions is the most informative part of the whole exercise, because it separates routine reliability from what happens in bad weather.
This is where the exclusion rule matters. Standard reporting leaves out major event days, so a utility in hurricane or wildfire country can post a respectable baseline number while its customers experience something quite different. When you look up your own territory, look for both figures, and treat the with-major-events number as the one that describes your actual planning problem.
The other thing the numbers reveal is territory shape. Rural systems, including most electric cooperatives, run many more miles of line per customer than an urban utility does, often through trees. More exposure per customer generally means higher SAIFI, and longer drive times to each fault mean higher CAIDI. That is geography rather than mismanagement, but it lands the same way on your evening.
The Hardware Stack:
Sense Home Energy Monitor: Logs when power is actually lost at your address, which builds a record specific to your house rather than to the territory average.
EcoFlow Delta Pro 3: A portable power station sized for the short, frequent interruptions that a high SAIFI and low CAIDI profile produces.
Generac Guardian 24kW: A standby generator sized for the opposite profile, fewer but much longer outages, where days of runtime matter more than instant transfer.
Vulnerability Score
Unknown baseline: A household that has never looked up its territory's numbers is sizing backup power from memory, which overweights the one bad storm and forgets the rest.
Excluded events: Reliability reported without major event days describes a normal year, not the year that actually drives the decision to buy backup.
Wrong-shaped backup: Frequent short interruptions and rare long ones call for different equipment, and buying for the wrong pattern means either paying too much or running out of runtime.
Where to Find Yours
Most state utility commissions publish annual reliability reports by utility, and many utilities put their own figures on their websites. Searching your utility's name with the word reliability, or checking your state commission's site, usually gets you there in a couple of minutes.
What to do with the answer is straightforward. A high frequency and short duration profile points toward something that switches over instantly and covers a few hours: a battery, or a well-chosen power station. A low frequency and long duration profile points toward runtime and refueling: a larger battery with solar, or a standby generator. Both are reasonable purchases. Which one is reasonable for you is a question your utility has already published the answer to.
Written by Mason Vance
