Somewhere on the bill, next to the meter reading, there is a single letter. Most people have never noticed it. It says whether the number beside it was measured or guessed.
An estimate is not misconduct. A meter reader cannot get in, a communication module drops off the network, a locked plant room is a locked plant room. Every state commission permits estimation under defined conditions and limits how long it may continue before an actual read is required. The utility issues a bill based on a projection, and when a real reading arrives, the difference is corrected.
That correction is where the interest lies, because a demand tariff does not true up the way an energy-only tariff does.
Energy trues up cleanly
Energy is cumulative. The register counts up and does not care whether anyone looked at it. Two estimated months followed by an actual read produce a total that is simply right: whatever the estimates assumed, the actual reading contains the true cumulative figure, and the true-up redistributes the difference across the estimated periods.
The only complications are ones of price rather than quantity. If the estimated periods spanned a season boundary, or sat in different time-of-use buckets, or fell either side of a rider adjustment, then how the corrected energy is allocated between periods changes what it costs. The tariff or the commission's billing rules say how. It is worth checking that the allocation was done on that basis rather than by a simple average, particularly where a summer rate is involved.
Demand does not
Demand is a maximum, not a total, and a maximum cannot be reconstructed from a later cumulative reading.
If the meter was read only at the end of a three-month estimated stretch, the register may hold the highest demand across the whole stretch, or it may have been reset, or the meter may retain a monthly maximum for each period — it depends on the meter and on how the utility configures it. On an advanced meter that lost communication but kept recording, the interval history is usually recoverable and the correct monthly maxima can be extracted. On an older meter it may genuinely not be available, in which case the tariff or the commission's rules specify a fallback, commonly the estimated figure or a value derived from the same period in the previous year.
This is worth attention for one specific reason: a demand figure that lands in an estimated period can set a ratchet floor. If the estimate was high, and the tariff has a ratchet, the consequence is not confined to the estimated month. It follows you for the whole look-back period, and correcting the estimate later does not automatically unwind the ratchet unless somebody recalculates it. See ratchet clauses for how far that reaches.
What to check when a true-up arrives
A true-up bill is frequently large, arrives without warning and is presented as a single adjustment line, which is exactly the format in which errors survive.
- Identify every period being corrected and the read type originally used for each.
- Confirm the actual reading is genuinely an actual reading, not a second estimate.
- Reconcile the total: corrected register difference times multiplier should equal the corrected energy across the whole span.
- Check how the corrected energy was allocated between periods, and whether that allocation crosses a season or time-of-use boundary in a way that changes the price.
- For each corrected period, check what demand figure is now being billed and where it came from.
- If a ratchet is in force, recompute the floor using the corrected demand figures rather than the estimated ones.
- Confirm the correction is within the commission's back-billing limit for your state.
- Confirm any interest, late fee or penalty applied to the corrected amount is permitted by the tariff — an undercharge caused by the utility's own estimate is generally not treated as a late payment by the customer.
The sixth line is the one that recovers real money and is almost never done. A ratchet floor set by an estimated peak, left standing after the estimate was corrected, quietly overcharges every subsequent month until somebody notices.
Avoiding the situation
Repeated estimation is a symptom, and the fix is nearly always physical: access to the meter, or a communication module that works. Both are worth pursuing, because estimation costs you more than accuracy.
An estimated period is a period with no interval data, which means no load shape, no peak timing, and a gap in the baseline any future measurement and verification exercise will depend on. If you are planning demand work, a run of estimated months in the baseline year is a genuine problem: measurement and verification explains why the baseline has to be continuous.
Ask the utility for the read-type history on the account. If more than the occasional period is estimated, that is the thing to fix before anything else on this site becomes measurable.
Where this sits
Read type is point ten in the 12-point bill audit, and it belongs there rather than at the top because it is rarer than the tariff mismatches. But it is worth a glance every month, because it takes two seconds and because a run of estimates that nobody questioned is the usual precondition for a true-up nobody expected. Understanding what the corrected determinants actually are starts with billing determinants.
