The Catch-Up Bill Blamed on the New Heat Pump

Why this matters

When a shop installs new equipment and the customer's bill jumps two months later, the shop is the first suspect, and it usually accepts the role. It sends a tech, the tech tests the system, the system is fine, and the customer hears "we checked everything" as "we cannot explain it". That is the worst position to be in: the customer still has the bill, and now also has the impression that the shop installed something it does not understand.

This case is reconstructed from the paperwork after it was settled, because the paperwork is where both the answer and the shop's mistake are visible. The answer was on the customer's bills the whole time. The mistake was that nobody at the shop read them until the second call.

The record, in the order it was written

Pulled from the customer's account history, the shop's tickets and one letter the customer kept:

• Nov 3. Shop replaces a 14-year-old heat pump. Install ticket: system commissioned, charge verified, thermostat programmed. Nothing about the meter, which is normal; nobody writes down the meter on an install.
• Nov 10 to Feb 10. Three bills, below.
• Jan 20. Letter from the utility to the customer: two consecutive reads missed, please provide access to the meter.
• Feb 13. Customer calls the shop: "My bill more than doubled since you put the new system in." Office books a diagnostic, ticket reads "new heat pump, high usage, check system".
• Feb 14. Tech visit, 1.5 hours on site. Ticket: pressures, temperatures, amps all normal, aux heat staged correctly, outdoor lockout set. "System operating as designed." No bills on the ticket.
• Feb 17. Customer calls again, angrier, with the bill photographed. Office manager pulls the three bills and reads the meter columns.

BILL 1  Nov 10 - Dec 9     29 days
  Prev 41,220 (A)  Curr 42,040 (E)     820 kWh
BILL 2  Dec 9 - Jan 11     33 days
  Prev 42,040 (E)  Curr 42,910 (E)     870 kWh
BILL 3  Jan 11 - Feb 10    30 days
  Prev 42,910 (E)  Curr 44,820 (A)   1,910 kWh
  (A) actual read    (E) estimated read

Each bill's kWh is its current read minus its previous one: 42,040 minus 41,220 is 820; 42,910 minus 42,040 is 870; 44,820 minus 42,910 is 1,910. Bills 1 and 2 both end on an estimate. Bill 3 is the first actual read since November.

What the customer was comparing

The customer put Bill 3 next to Bill 2: 1,910 against 870 kWh, 2.20 times. Per day it looks worse, 1,910 / 30 = 63.67 kWh against 870 / 33 = 26.36, which is 2.42 times. She was not wrong about the numbers on the page. She was comparing an estimate that was too low with an actual read that had to make up the difference, and that ratio exaggerates the change in both directions at once (the mechanism is owned by Estimated Versus Actual Meter Reads and the Catch-Up Bill; this case is about what it costs a shop not to know it).

Measuring from actual to actual

The only kWh figure this record can vouch for is the one between two actual reads. The last actual before Bill 3 is 41,220 at the start of Bill 1; the next is 44,820 at the end of Bill 3.

44,820 minus 41,220 is 3,600 kWh, over 29 plus 33 plus 30 = 92 days. That is 3,600 / 92 = 39.13 kWh a day across the whole span, and all 92 days are on the new heat pump, because it went in a week before Bill 1 began.

Neither the low estimates (28.28 and 26.36 a day) nor the catch-up bill (63.67 a day) is the house's real use. 39.13 is.

How much of Bill 3 belongs to December

The 3,600 total is fixed. How it splits across the three bills is not, because the real reads are missing. The simplest split assumes even use, 39.13 a day, which puts 1,135 kWh on Bill 1, 1,291 on Bill 2 and 1,174 on Bill 3, and makes the estimates short by 315 and 421, together 736 kWh.

Even use across a cold January and a milder November is not a good assumption for a heated house, so the office manager split it by weather instead. From the house's October and May bills, the baseload (use per day with no heating or cooling) is about 15 kWh a day. Over 92 days that is 15 x 92 = 1,380 kWh, leaving 3,600 minus 1,380 = 2,220 kWh of heating. The span had 2,310 heating degree days (a degree day counts how far and how long the outdoor temperature sat below 65 F), so the house used 2,220 / 2,310 = 0.961 kWh per degree day.

Baseload is 15 times the days; heating is 0.961 times the degree days. The three expected figures add to 3,600.

Bill 1 was estimated at 820 against an expected 1,002, short by 182. Bill 2 at 870 against 1,360, short by 490. Together 672 kWh. Bill 3's 1,910 is its own expected 1,238 plus that 672, which is 1,910. So 672 / 1,910, about 35 percent of Bill 3, is November and December use billed in February. Bill 3's own January use works out to 1,238 / 30 = 41.3 kWh a day, which is an ordinary January for this house.

The even split said 736 kWh and the weather split says 672. The difference is the honest uncertainty in where the use landed. The total does not depend on either.

Was the new system using more than the old one?

This is the question the customer was actually asking, and the record can answer it narrowly. Her account history showed the same three periods last winter at 3,780 kWh on actual reads, on the old heat pump, with 2,240 degree days.

• Last winter: 3,780 / 92 = 41.09 kWh a day. Heating: 3,780 minus 1,380 baseload = 2,400 kWh, or 2,400 / 2,240 = 1.071 kWh per degree day.
• This winter: 0.961 kWh per degree day, from above.

0.961 / 1.071 = 0.90. Adjusted for weather, the house used about 10 percent less electricity per degree day of heating this winter than last. This winter was also colder, 2,310 degree days against 2,240, so the raw per-day totals (39.13 against 41.09) understate it: about 5 percent less per day raw, against about 10 percent less per degree day.

That is the full extent of what the shop told her about efficiency. One winter, a baseload taken from two shoulder-season bills, and a heating figure that also carries whatever changed in the household; it is evidence that the new system is not using more, not a measured efficiency gain, and it is not a promise about next winter. The weather method is in How to Weather-Normalize Two Winters of Bills.

A read the shop took itself

One more check closed the loop, and it is the only number in this case the shop produced rather than read off paper. On Feb 21 the office manager asked the tech, already in the area, to photograph the meter register: 45,280. From the actual read of 44,820 on Feb 10, that is 460 kWh in 11 days, 41.8 kWh a day.

Those 11 days had about 330 degree days, 30 a day. At the house's own rate the expected use is 15 plus 0.961 x 30 = 43.8 kWh a day. The measured 41.8 is close to it, which says two things: the Feb 10 actual read was real, and the house was using an ordinary amount of electricity for the weather in the weeks after the catch-up bill.

Why the estimates happened

The utility's Jan 20 letter explained it, once someone asked the customer whether she had received anything. The meter was on the back wall inside a fenced yard. In November, after a delivery driver left the gate open and the dog got out, she padlocked it. The meter reader could not reach the meter, the utility estimated, and after two misses it sent the letter. She arranged access, which is how Bill 3 got an actual read.

Nothing in the shop's record could have shown this, and the shop did not need it to settle the bill. It still matters, because an account with a locked gate keeps getting estimated, and the next catch-up bill is already accumulating. The customer was told to keep the gate accessible on read days or ask the utility about other read arrangements.

The visit that should not have happened

Reading the record backwards, the Feb 14 visit is the expensive entry. A tech spent 1.5 hours on site, plus the drive, proving that a correctly installed system was correctly installed. The ticket was accurate and useless to the customer, because her question was about a bill and nobody on that ticket had looked at one. Worse, "system operating as designed" with no explanation for the bill reads to a customer as a shop that either does not know or is not saying.

The E flags were printed on the bill she was holding when she made the first call. Thirty seconds of the office reading the meter columns would have ended the call on Feb 13.

What the shop changed

The high-bill intake script now opens with the bill, not the equipment:

1. "Do you have the bill in front of you? Next to the meter readings, does it say E, EST or estimated on this bill or the one before it?"
2. "How many days does this bill cover, and how many did the last one?"
3. Only then: what changed in the house, and what the equipment is doing.

If either of the last two bills is estimated, the office does not book a diagnostic. It explains the catch-up in plain terms, asks for the bills, and does the actual-to-actual arithmetic itself, the way it was done here. A tech goes out only if use per day, measured between actual reads and adjusted for weather, is genuinely up.

The second change is on install tickets: the installer now photographs the meter register at commissioning. It costs nothing and gives the shop an actual read of its own on the day its equipment started, independent of whether the utility's reader gets through the gate that month.

References

• See related: Estimated Versus Actual Meter Reads and the Catch-Up Bill
• See related: How to Take a High-Bill Call at the Office
• See related: How to Weather-Normalize Two Winters of Bills
• See related: How to Diagnose a High-Bill Call, in Order
• See related: Why a Bill Rises: The Causes That Are Not the New Equipment