Warranty work and callbacks: handle them well, learn from them ruthlessly
The callback call has a particular sting. "That thing you fixed — it's doing it again." Every trade gets them; no trade enjoys them. A callback is a job you're doing twice and billing once, a customer whose confidence is wobbling, and — if you're paying attention — a data point about your own business that most owners never collect. How you handle callbacks decides whether they cost you a morning or a reputation. How you track them decides whether they keep happening.
Here's the system for both halves: handling the individual callback like a professional, and mining the pattern like an operator.
First, establish what you're actually looking at
Not every "it's broken again" is a warranty claim, and sorting them fast protects everyone. There are really four species: a genuine defect in your workmanship (yours to fix, free, promptly), a failed part (the supplier's warranty, your labour policy per your terms), a new fault the customer understandably associates with the old one (billable work, explained kindly), and damage or misuse since your visit (billable, documented carefully).
Telling them apart in seconds is what the job history is for: the original job on screen shows exactly what was done, which parts went in, the photos of the completed work, and what was recommended but declined — that last note matters enormously, because "the cheap option you chose has failed as advised" is a very different conversation from a warranty claim, and only a written record makes it graciously winnable.
Handle the genuine ones like they're marketing — because they are
When it's genuinely yours, speed and grace are the whole play. A warranty visit handled the same week, without quibbling, converts a wobbling customer into a loyal one — research on service recovery keeps finding the same uncomfortable truth: a problem fixed well often builds more loyalty than no problem at all. Practically: the callback gets scheduled like real work, not squeezed apologetically into a someday-gap (the schedule board makes room the same way it absorbs any urgent job); the tech arrives with the full original job on their phone; and the fix gets photographed and documented on the same record, closing the loop. What kills reputations isn't the defect — it's the dodge: the unreturned calls and the "that's not ours" reflex. Your Google reviews are written in exactly these moments.
Link the callback to the original job — always
The single most valuable habit in this whole area is mechanical: every callback gets created as a job linked to the original. Not a fresh job with no history — a visible child of the work that spawned it. This does three things at once. The tech walks in with context. The customer's file tells the true story if the dispute escalates. And — the operator's payoff — callbacks become countable, which is where the real money is.
The callback rate: quality data wearing a cost disguise
Individually, callbacks feel random. Counted, they almost never are. Your callback rate in the weekly numbers, sliced by job type, by crew and by parts used, surfaces patterns with direct fixes: one job type generating repeat visits (the install process needs a checklist step), one supplier's component failing young (change supplier, claim the pattern back against them — your linked records are the evidence), one crew's work bouncing more than the others' (a training conversation backed by data instead of vibes, and the flip side of the load-balancing view). Each percentage point of callback rate is pure margin: unbillable hours, travel, and parts, all recoverable through fixes the data points straight at.
Put your warranty in writing before anyone needs it
The awkwardest callbacks are the ones where nobody knows what was promised. A short written warranty policy — what's covered, for how long, workmanship versus parts, what voids it — attached to every quote and invoice turns future arguments into lookups. It also, counterintuitively, sells: a stated 12-month workmanship warranty reads as confidence, and confident reads as competent. Just make sure the promise and the job documentation that backs it grow up together; a warranty you can't verify against records is a liability wearing a smile. And know your local consumer law floor — in most countries, statutory guarantees on trade work apply regardless of what your policy says, so write the policy to sit on top of them, not pretend beneath them.
Frequently asked questions
Should warranty visits be logged with time and materials like paid jobs? Absolutely — that's how the true cost of callbacks becomes visible in job costing, and how a failing part's pattern gets proven to a supplier.
What's an acceptable callback rate? It varies by trade, but the trend matters more than the benchmark: measure it, then make it fall. Most businesses that start counting find a cluster they can eliminate within a quarter.
How long should I keep job records for warranty purposes? At least your stated warranty plus your legal liability horizon — for building work that can be a decade. Cloud records make long retention free; paper makes it fiction.
Run the whole job in one place
Schedule the crew, run jobs from the van, manage site safety and invoice the moment a job is done.
Start free trialKeep reading
A customer database for trades: why job history is your most undervalued asset
Every past job is future revenue. How a trade CRM with per-customer job history speeds up repeat work, powers…
GrowthThe 7 numbers every trade business should check weekly (and where to find them)
Seven KPIs that tell you if a trade business is actually healthy: quote win rate, jobs per crew, unbilled…