Customer Disputes Production Estimate After First Bill SOP

Purpose

A customer's first utility bill after going live is the highest-stakes production conversation the shop has, not because the system is more likely to be broken then than any other time, but because the customer has no history to compare against and reads the whole relationship off this one number. Run it as a full shortfall investigation and it burns days pulling data on a system with barely two weeks of production behind it. Wave it off as "bills run low the first month" and the one call in ten where the number really is wrong gets missed. This procedure is the fast triage in between: correct for the two things that make almost every first bill look worse than it is, a partial first billing cycle and a proposal figure never actually quoted for that calendar month, and route only a genuine gap to a full investigation.

Scope

Covers the shop's response to a customer disputing production or bill impact on their first utility bill after permission to operate, through a written close or a handoff.

Does not cover the full data-driven shortfall methodology, irradiance normalization, module-level screening, inverter fault-log attribution, owned by the production shortfall investigation SOP, which this hands off to when the quick screen fails. Does not cover net metering tariff structure or true-up mechanics themselves, covered by the net metering and billing reference and any state-specific tariff reference for that utility. Does not cover a dispute raised months into service with production history already on file; that is a shortfall claim from the start.

Roles and responsibilities

Procedure

Step 1 - Take the claim in the customer's own words, then split it into what it actually is. Ask whether the complaint is that the system made less power than promised, a production claim measured in kWh, or that the bill's dollar impact was smaller than expected, a rate-design or true-up question, since the two need different first moves and are frequently confused even when only one is really wrong. Acceptance: the intake record states which claim type, plus the billing period's exact start and end dates. Wrong looks like treating "my bill is way higher than you said" as a production shortfall when the system produced exactly to plan and the mismatch is in how net metering nets against usage, not generation. Hazard: none at this step, it is an intake call.

Step 2 - Cross the billing period against the permission-to-operate date before comparing anything. Count the days inside the billing cycle that actually had solar production against the cycle's total length. Acceptance: a stated production-day count and the cycle's total length, both written down before any kWh comparison happens. Wrong looks like comparing a full month's proposal figure against a bill that only covers two weeks of live production; a partial first cycle is the single most common driver of a first-bill call, and skipping this step guarantees the wrong verdict. Hazard: none, this is arithmetic against a stated calendar.

Step 3 - Pull the proposal's own modeled figure for that specific calendar month, never the annual total divided by twelve. Annual production is not flat across the year, and a customer who recalls "the proposal said about this many kWh a month" is almost always remembering the naive annual average, not the modeled figure for the month in question. Acceptance: a monthly modeled figure for the billing period's calendar month, sourced from the proposal's own monthly breakdown or rebuilt in a public PV performance model against the as-built system where the proposal never broke one out. Wrong looks like quoting the annual-average figure back to the customer as the standard for every month; that figure is exactly what created the dispute. Hazard: none.

Step 4 - Reconcile the monitoring platform's reported production against the utility bill's metered figure before trusting either one for the comparison ahead. Acceptance: platform and bill figures agree within 2 percent of the bill figure. Wrong looks like running the case off the platform alone; a production CT is more failure-prone than the utility's own revenue meter, and the bill is what the customer is actually charged against, so it governs regardless of what caused a mismatch. Stop rule: a gap beyond 2 percent gates the case; dispatch a physical check before quoting either number as settled. Hazard: reaching the CT means opening the inverter's AC output enclosure, which stays energized until proven dead per the live-dead-live sequence at NFPA 70E-2021, 120.5, under 29 CFR 1910.333(b)(2) (1926.417 in construction); where the enclosure is the utility's own sealed meter socket, the tech does not break that seal, the utility is contacted instead, since a broken utility seal can itself trigger a service disconnection.

Step 5 - Restore and confirm the reading only after the enclosure is proven dead, and record which figure governs going forward. Where the CT is found loose or unseated, reseat it, close the enclosure and restore power before leaving; the current dispute still runs on the bill's own figure regardless of the CT finding, since that is the number the customer is billed against, while the corrected CT protects every future case at that site. Acceptance: enclosure closed, power restored, a clean reading confirmed, and the governing figure stated in the file. Wrong looks like walking away with the fix undocumented, leaving the next tech to rediscover the same loose CT. Hazard: this is the resume step for what step 4 opened, and it gets its own check, not just a closed cover: confirm no fault or nuisance trip on restoration, since a connector reseated under load can announce a separate fault the moment power returns.

Step 6 - Prorate the monthly-specific expectation to the verified production days and run the quick screen. Multiply the calendar-month modeled figure by the ratio of production days to the cycle's total length, then compare the verified bill figure against that prorated number. Acceptance: the screen passes at or above 80 percent of the prorated expectation; anything below hands off. Wrong looks like screening against the un-prorated full-month figure, which fails almost every first bill by construction regardless of how the system is actually performing. Stop rule: below 80 percent, the case does not close here as a literacy issue; it hands off to the production shortfall investigation SOP with these verified figures attached, rather than restarting from a phone call. Hazard: none.

Step 7 - Where the claim included a bill-impact question, address it separately from the production number. Walk the customer through how net metering nets exports against draw for this cycle and what the true-up cycle does with any banked credit, using the net metering and billing reference and the utility's own program rules rather than improvising the mechanics on the call. Note that one month essentially never shows the full annual offset. Acceptance: the dollar-side question answered against the customer's own rate structure and true-up cycle, not a generic explanation. Wrong looks like promising a specific future bill amount to end the call; that promise is what turns into next month's dispute. Hazard: none.

Step 8 - Close in writing, and where the mismatch traces to an annual-versus-monthly proposal gap, feed it back to sales the same week. Send the customer the corrected comparison, prorated expectation against the verified figure, in plain terms; where the proposal never showed a monthly breakdown and the customer built its own annual-average number, flag that presentation gap to sales so the next quote shows the month the customer will actually open first. Acceptance: written close-out sent, sales feedback logged where warranted. Wrong looks like a verbal reassurance with no written comparison left behind; that is the record a repeat call arrives into with nothing to point to. Hazard: none; a first-bill dispute can coincide with a contract's own satisfaction or cancellation terms, so what this written record establishes matters beyond the one call, and any question about what those terms require routes to the shop's own contract and counsel, not a promise made on the phone.

The record this produces

One first-bill case record: claim type, billing period dates and the PTO date, production-day count against cycle length, the calendar-month modeled figure and its source, the monitoring-versus-bill reconciliation with any CT finding, the prorated expectation and the verified figure it was measured against, the screen result and its routing, the bill-impact explanation where relevant, and the close-out letter. Where the case hands off, the full investigation starts from these figures rather than a blank intake.

Worked pass

A 7.6 kW inverter system, permission to operate granted October 1. The utility billing cycle runs September 15 to October 14, 30 days. The customer calls after the first bill: the proposal said 11,000 kWh a year, over 900 a month, and the bill only shows 310.

Step 1: claim is production, not bill impact; billing period September 15 to October 14 recorded. Step 2: only 14 of the 30 cycle days fall after the October 1 PTO date; the other 16 had no array in service at all. Step 3: the proposal's own monthly breakdown models October at 790 kWh for a full month; the customer's remembered "over 900" is the naive annual average, 11,000 divided by 12 is about 917, which was never the October-specific figure.

Step 4 fails. The monitoring platform shows 322 kWh for the period; the bill shows 310. The gap is about 3.9 percent of the bill figure, beyond the 2 percent tolerance. The stop rule runs: a tech is dispatched before either number is quoted as final.

Step 5: at the inverter's AC output enclosure, proven dead per the live-dead-live sequence, the production CT is found seated with a slight gap in its jaws rather than fully closed around the conductor. It is reseated, the enclosure closed, and power restored with a clean reading and no fault on restoration. The case still runs on the bill's 310 kWh, the authoritative billed figure; the reseated CT is what keeps the next case accurate.

Step 6: prorating October's 790 kWh full-month figure to 14 days gives about 369 kWh expected. The verified 310 kWh is 84.1 percent of that, which clears the 80 percent screen, and the case closes here rather than handing off.

Step 7: the bill-impact question is a smaller true-up concern, walked through against the utility's annual reconciliation cycle and why one autumn month never carries the full annual offset.

Step 8: the written close-out shows the prorated comparison, 369 kWh expected against 310 kWh verified for 14 production days, not the 900-plus figure the customer was comparing against. Sales is flagged, since the proposal's monthly breakdown was never shown at the sale and the customer built its own annual-average number instead.

References

• NREL PVWatts or an equivalent public PV performance model, for a calendar-month figure where the proposal never broke one out.
• NFPA 70E-2021, 120.5 for live-dead-live verification and 29 CFR 1910.333(b)(2) for electrical work practices (29 CFR 1926.417 in construction).
• See related: the net metering and billing reference for true-up and rate-design mechanics, the production shortfall investigation SOP for the full data-driven method this procedure hands off to, and any state-specific interconnection or tariff reference for the customer's utility.
• Your own contract terms and counsel for what a first-bill dispute record establishes under any cancellation or satisfaction clause.