How to Diagnose a High-Bill Call, in Order

Why this matters

A high-bill call is the one service call where the cheapest evidence is also the best evidence, and most shops collect it last. The usual sequence is a tech in the house with a clamp meter, testing whatever looks old, before anyone has looked at the bills. That tech finds something, because an old house always has something, and the shop sells a repair that does not move the bill. Three weeks later the customer calls back angrier, now with a receipt.

The order below runs from evidence that costs nothing (paper the customer already has) to evidence that costs a truck roll. Each step either explains the bill or narrows where to look next. Most calls end before step 7, and that is the point.

The order, and why it runs this way

The steps are ordered by what each piece of evidence costs to collect, cheapest first, and each one is a filter for the next. Paper rules out billing and calendar effects. Weather rules out the season. Interval data says when the extra energy is being used, which says what kind of load it is. Only then does anyone touch equipment, and by then they know what to test.

Step 1: Get twelve months of bills, not two

Ask for the last twelve to fourteen months of bills, or the usage history table most bills print, or the utility's online usage portal export. Two bills cannot tell you whether this month is abnormal; twelve show the house's own seasonal pattern.

Skip this and you end up comparing January to December, which is comparing a heating month to a shoulder month. Every winter, that comparison makes a normal house look broken.

Step 2: Convert every bill to use per day

Read the billing days on each bill (the service period, usually printed as two dates and a day count). Divide the usage by the days. A 33-day bill against a 29-day bill is 14 percent longer before a single extra kilowatt-hour is used. See related: How to Compare Two Bills With Different Billing Days, which owns this step in detail.

What breaks if skipped: the total-on-the-bill comparison, which is the one the customer is making, overstates every increase on a long bill and hides increases on a short one.

Step 3: Compare against the same months last year

Put this month's use per day next to the same billing period one year earlier. Same season, same daylight, roughly the same weather. That is the fair comparison, and it is often the one that ends the call: plenty of "my bill doubled" calls turn out to be a winter bill that is within a few percent of last winter.

Step 4: Read the flags and the rate lines

Look at every bill for three things:

• Read type. An estimated read (often marked E, EST or "estimated") means the utility guessed. A low estimate followed by an actual read produces a catch-up bill that looks like a spike and is really two months of use on one page. See related: Estimated Versus Actual Meter Reads and the Catch-Up Bill.
• Rate lines. A rate change, a new supply rate from a third-party supplier, a seasonal or tiered rate crossing into its upper block, or a budget billing true-up can raise the bill with no change in use at all. See related: Anatomy of a Residential Electric Bill, Line by Line.
• Meter number. A changed meter number means a meter swap. A new meter is not usually wrong, but a swap is a moment when reads get estimated or misread.

Steps 2 to 4 separate a USE problem from a BILLING problem. If use per day is flat and the bill is up, the house is fine and the conversation is with the utility, not with a tech.

Step 5: Weather-normalize

If use per day is genuinely up against last year, find out how much of that the weather explains. Subtract the baseload (the use per day in a month with neither heating nor cooling), then divide the remaining heating or cooling use by the heating or cooling degree days in each period. A degree day is a measure of how far, and for how long, the outdoor temperature sat below (heating) or above (cooling) a base temperature, usually 65 F. The method is in How to Weather-Normalize Two Winters of Bills; do not skip it because the customer insists it "was not that cold."

Step 6: Pull interval data

A smart meter records use every hour, and often every 15 minutes. That record shows WHEN the extra energy goes in: a flat overnight rise points to something running continuously, a spike every cold morning points to heat recovery, a block in the afternoon points to cooling or a pool pump timer. See related: How to Get a Customer's Interval Usage Data Without Their Password.

Interval data turns "the house uses more" into "the house uses about 12 kWh more between 6 and 8 a.m. on cold days," which is a sentence a technician can do something with.

Step 7: Walk the forgotten loads

Before testing the main equipment, walk the house and the property for loads the customer has stopped thinking about: a second refrigerator or freezer in the garage, a heated aquarium, a dehumidifier in a damp basement, heat tape on a pipe, a well pump, a hot tub, a pool pump, a block heater, a space heater in a cold room, a grow light. See related: The Loads a Customer Forgets They Own. Match each to the interval shape from step 6. A hot tub does not make a 6 a.m. spike; a heat pump recovering from a night setback does.

Step 8: Test the suspect equipment

Now, and only now, test. Measure the draw of the load the evidence points to and confirm its runtime. See related: How to Measure Real Watts Instead of Trusting the Nameplate, and How to Confirm Backup Heat Runtime From the Thermostat and a Clamp. Electrical measurement inside equipment or at a panel is for a qualified person with the shop's electrical safety program behind them; opening a panel, disconnect or equipment cover exposes energized parts, and nobody else does it.

The test is a confirmation, not a search. If the measured load times its runtime does not add up to the excess you computed in step 5, the cause is not found yet, whatever the equipment looks like.

Worked example: an all-electric house with a heat pump

The customer calls in early February: the January bill "nearly doubled." The house has a heat pump with electric resistance backup (heat strips), electric water heater and range.

Step 1 and 2. The bills:

BILL            DAYS   KWH     KWH PER DAY
December        29     1,450   50.0
January         33     2,640   80.0
January (last)  31     1,860   60.0
May (last)      30       660   22.0  <- baseload

The customer's comparison, total against total, is 2,640 against 1,450, which is 1.82 times. Per day, January against December is 80.0 against 50.0, which is 1.6 times. Part of "doubled" was four extra days and a colder month.

Step 3. Against last January, 80.0 against 60.0 kWh per day: 1.33 times. Still a real increase.

Step 4. Both January bills show actual reads. Same meter number. The per-kWh rate went up slightly at the start of the year, which matters to the money but not to the kWh, and the kWh is what rose. So this is a use problem.

Step 5. May shows a baseload of 22.0 kWh per day. Degree days, from the weather station the utility uses: last January averaged 30 heating degree days (HDD) per day, this January 32.

• Last January heating use: 60.0 - 22.0 = 38.0 kWh per day, divided by 30 HDD = 1.27 kWh per HDD.
• This January: 80.0 - 22.0 = 58.0 kWh per day, divided by 32 HDD = 1.81 kWh per HDD.

So 1.81 against 1.27 kWh per HDD: the house used 1.43 times as much heating energy per degree day. The weather was 32 against 30 HDD per day, 1.07 times. Weather explains a small part; most of it is the house.

If the house had heated as efficiently as last year, this January would have used 1.267 x 32 + 22.0 = 62.5 kWh per day. It used 80.0. The excess is 17.5 kWh per day (17.47 before rounding), about 576 kWh over the 33 days.

Step 6. The interval data shows the overnight hours at 2 to 4 kWh per hour, normal for a heat pump at that temperature. On cold mornings, the hours from 6 to 8 a.m. read 11 to 14 kWh each. That shape is backup heat.

Asked what changed in the house, the customer remembers a new programmable thermostat installed in early December, set to drop to 62 F overnight and come back to 68 F at 6 a.m.

Step 7. The walk finds a garage freezer and a dehumidifier, both present last January too, and neither one makes a morning spike. Nothing new.

Step 8. The HVAC tech reads the thermostat's runtime history: auxiliary heat averaged about 1.75 hours a day over the billing period, nearly all of it in the morning recovery. The heat strips, clamped by a qualified tech at the air handler during a recovery, draw close to 10 kW. 10 kW x 1.75 hours = 17.5 kWh per day, which matches the excess from step 5. The cause is a setback recovery that calls for strip heat, not a failed component.

The fix is a thermostat setting (a smaller setback, or the thermostat's heat-pump recovery mode that holds off the strips), and it is an HVAC tech's call because staging and lockouts are configuration. The customer is told the likely range of the change and that the next two bills, read per day and against degree days, will confirm it. See related: Savings Claims Without Overpromising: The Honest Language.

What changes the order

• No smart meter, no interval data. Replace step 6 with a whole-house baseline test at the meter, with loads switched off one at a time. See related: How to Run a Whole-House Baseline Test at the Electric Meter.
• Delivered fuel. Propane and oil arrive by truck, so steps 1 to 3 run on delivery tickets, not monthly bills, and step 4 becomes the price per gallon on each ticket. See related: How to Estimate Delivered Fuel Use From a Run of Tickets.
• A water bill. The meter leak check moves to the front, because it takes five minutes, it costs nothing, and a running leak keeps costing while you analyze. See related: How to Walk a Customer Through a Water Meter Leak Check by Phone.
• Any safety sign. A burning smell, a warm panel, a breaker that keeps tripping or a gas odor stops the bill analysis entirely. A gas odor means everyone out of the house, no switches, no phone use inside, and a call to the gas utility's emergency line or 911 from outside. Electrical heat signs go to an electrician the same day.

How the order fails in the field

The tech starts at step 8. He finds a weak capacitor on a heat pump whose real problem is a thermostat schedule, replaces it, and the bill does not move. Catch it by requiring the step 5 excess figure in the job notes before any test is written up.

The customer's two bills become the whole dataset. Asking for twelve months feels like homework to the customer. Ask anyway, or pull the history from the portal with them on the phone; without it you cannot tell a seasonal bill from a problem.

The test finds "a" cause, not "the" cause. A load that explains 3 kWh a day does not explain a 17.5 kWh a day excess. The closing check is arithmetic: measured kW x confirmed hours should land near the excess. If it is short, keep going and say so to the customer. See related: How to Write a High-Bill Findings Note for the Customer.

References

• See related: How to Weather-Normalize Two Winters of Bills
• See related: Utility Bill Spike - Source Isolation Decision Tree
• See related: Reading Energy Data Without Misleading the Customer
• See related: The Bill That Doubled From Emergency Heat
• See related: How to Take a High-Bill Call at the Office