How to Handle a Rejected Rebate Claim

Why this matters

A rejection is not the end of a claim, it is a countdown that started before you opened the envelope. Most rejections are curable, most cures are cheap, and most of the claims that die do so because somebody sorted the pile by how hard each one looked instead of by how much time each one had left. The other half of the damage is conversational: a customer who hears "it was denied" and a customer who hears "it came back for a missing document and I have until the end of the month to fix it" have two completely different opinions of your shop, and the underlying facts are identical.

Read the clock before you read the reason

Open the notice and find one thing first: the cure window, and the date it started. Not the rejection reason. The reason tells you what the work is; the clock tells you whether the work is worth starting.

Three properties of that clock catch shops out. It usually runs from the rejection date, not from the date you received the notice, so mail transit and an unopened portal notification come out of your time, not theirs. It is usually shorter than the program's original filing deadline, sometimes much shorter. And on some programs it is the only remaining deadline, meaning a lapsed cure window closes the claim permanently even when the original filing deadline is months away.

Write the cure end date on the claim's log row the day the notice arrives, before you do anything else with it. A rejection sitting in a pile without its end date recorded is the single most common way a curable claim becomes an uncurable one.

Four classes, and what each costs to cure

Once the clock is written down, classify. The class determines the cost and, more usefully, tells you whether the cure is even in your control.

Clerical. A field left blank, a superseded form version, an unsigned line, a transposed digit, an illegible scan. The information exists and is correct; the document is wrong. Cheapest class by a wide margin, typically well under an hour, and the most common. These are also the ones people put off, because they feel trivial.

Evidentiary. The document exists somewhere but was not submitted or was not acceptable in the form submitted: a permit record you have to pull, a bundled invoice that has to be reissued itemized, a photo that is in a technician's phone rather than on the job record. Moderate cost, and the cost is mostly in retrieval and handoffs rather than in the fix.

Reconstructable. The evidence does not currently exist in usable form and has to be legitimately recreated: a return visit to photograph a nameplate, a letter from the manufacturer confirming a rating, a corrected identity block after a data plate was never captured. Expensive, often several times an evidentiary cure, because it involves a truck, a customer's schedule, or a third party's response time. Legitimately recreating a document means recovering a fact that is true. It never means creating one that is convenient: no back-dating, no serial numbers from a parts list, no adjusted completion dates, no signing for a customer.

Uncurable. The claim fails on something no document fixes: work outside the program period, equipment not on the qualifying list as of the install date, a premises already claimed, funding exhausted, installer credential status wrong at the time of install, or a cure window that has already closed. Cost here is only the cost of closing it out properly and making one honest call.

Triage by closing window, not by ease

Sort the batch by cure end date, soonest first, and work it in that order. That is the whole rule, and it is counterintuitive enough that shops reliably do the opposite.

The natural instinct is to knock out the easy ones first, which feels productive and clears the pile fast. It also means the hard claim, which is hard precisely because it needs a third party or a truck, gets started with the least time remaining. A reconstructable cure that needs a customer to be home and a manufacturer to answer needs the most calendar, not the least.

One exception overrides the sort: an uncurable claim is closed out immediately regardless of its dates, because it consumes nothing but a customer call and leaving it in the pile makes the pile look worse than it is. Do that first, then work the rest by clock.

Fix every defect, not the cited one

Processors run their checks in sequence and stop at the first failure, so the rejection you received names the first defect found, not the only one. Refiling against just the cited reason is how a claim comes back twice, each cycle burning calendar you did not have.

Before you refile, re-run the whole check sequence against the corrected claim yourself: form completeness and version, filing and cure deadlines, claimant identity against the program's payee rule, product identity against the qualifying list as of the install date, date ordering, invoice itemization and legibility, installer credentials, duplicate and premises. That review takes minutes on a claim you already assembled once, and it is the difference between one cure cycle and two. See the companion reference on what a processor actually checks for the sequence itself.

A batch of nine rejections

Nine rejections arrived across one week on one program. Effort below is office hours.

Three readings come out of that table, and only the first is obvious.

One claim consumed about 42% of the batch's cure effort. The single reconstructable claim, at 3.5 hours out of 8.4, cost more than all four clerical cures and one evidentiary cure combined. It needed a return visit for a nameplate photo that was never captured at the job. That is not a rejection-handling problem, it is a capture problem arriving late, and the honest conclusion the shop drew was that the fix belonged on the job close-out card rather than in the coordinator's process.

Seven of the nine, about 78%, were paid after refile. Eight of the nine were curable in principle. One of the eight lapsed.

The one that lapsed was an evidentiary claim, and it lapsed because it was worked last. It looked routine, a permit record that needed pulling, so it went to the bottom of a pile sorted by apparent difficulty. Its cure window happened to be the shortest in the batch, and the permit office took longer than expected to respond. Sorted by clock rather than by ease, it would have been started first and would almost certainly have paid. That single claim is the entire argument for the triage rule, and it cost the shop more than the 3.5-hour reconstruction did, because the reconstruction got paid and this one did not.

The uncurable one was a premises already claimed inside the program period, on a home that had changed hands. No process on the shop's side would have prevented it, because the program exposes no premises lookup and the customer did not know the prior owner had claimed it. Worth naming plainly: that one belongs in a different bucket from the lapsed claim when you review the season, and lumping "we lost two" together hides the only failure that was actually yours.

What you tell the customer, and when

Call on the day you classify, not on the day you resolve. The customer has been told a claim is in flight and they are entitled to know it came back, before they hear about it from the program.

Say four things. What the program returned it for, in plain language rather than the rejection code. What class it is and therefore what has to happen. What you need from them, if anything, with a date attached. And what the realistic outcome is, honestly, including the case where it is uncurable.

For a curable claim: "It came back because the invoice bundled the equipment into one line and they need it itemized. I am reissuing it and refiling this week, and nothing changes on your end." For an uncurable one, do not soften it into ambiguity, because ambiguity means they will call three more times: "It was denied because the program's records show an incentive already claimed on this address in this period. That is not something a corrected form fixes. Here is who you can call if you want to hear it from them directly."

Never characterize a rejection as the program being unreasonable, even when you privately think so. You will be filing with them again next month, the customer may call them, and a shop that badmouths a processor sounds like a shop making excuses.

Who eats it when the cure window closes

Write this policy down before the first rejection, not during it. Deciding case by case produces three different answers for three customers who talk to each other, and the answer always drifts toward whoever complains loudest.

A defensible split, and the reasoning behind each side:

• The shop makes the customer whole when the failure traces to a link the shop owned and the window has closed. Identity evidence never captured, an invoice format you chose, a submission that missed a deadline, a notice that sat in an unshared inbox. You controlled it, you failed it, and the customer's loss is measurable in exactly the amount the program would have paid. Book it against the job so the cost lands where it was caused rather than disappearing into overhead.
• The shop does not pay when the failure traces to the customer or to the program. A premises already claimed, an account name the customer changed, funding exhausted, terms revised, a signature they would not return in time. You explain it, you document what you did, and you hold the line, kindly.
• Ambiguity goes to the customer once, and then into the process. Where you cannot cleanly assign it, pay it, and open the actual question, which is which link was ambiguous and who owns it going forward. A rule that never costs you anything is a rule nobody believes.

The number to watch is not any single make-whole. It is the rate. Rejections traceable to your own links should be falling season over season, and if they are flat while volume grows, the problem is upstream at capture or submission and no amount of skilled rejection handling will fix it.

References

• The current published cure window, appeal process and filing deadlines of the specific utility, manufacturer or state program, which govern every clock in this article
• See related: What a Rebate Processor Actually Checks; How to Capture Rebate Documentation at the Job; How to Track a Submission You Do Not Control; The Rebate Season That Cost More Than It Returned