What fails by brand: which part breaks, from 18,670 repairs

Every appliance reliability ranking answers the same question — which brand should I buy — and almost nobody is asking it at the moment they search. They are standing in front of a machine that has already failed. This page answers the other question, from 18,670 completed repairs where both the brand and the appliance were recorded: on the machine you already own, which part is the one that fails. The records come from a repair company's own index of 65,924 completed jobs across nine appliance types in six US states.

The methodological point matters more than any single figure. Each percentage below is a share of that brand and appliance combination's own completed repairs — not a share of all calls. Brand popularity therefore does not distort the numbers, which is exactly the trap that makes most published reliability rankings useless: a brand with twice the installed base produces twice the service calls while being no more likely to fail. The same publisher runs that calculation too, and labels it as what it is. A dash in a table is its own marker for a cell it does not report.

Refrigerators: one number stands away from the rest

Share of that brand's completed refrigerator repairs. The dash is the publisher's marker for a cell it does not report.

The LG compressor figure is the outlier in the whole dataset. Compressor work sits between 2.3% and 3.1% of refrigerator calls for Samsung, GE, Whirlpool and Frigidaire — a spread of less than one point across four brands with nearly five thousand jobs between them. LG is at 9.8%, more than three times the top of that band, and Kenmore at 6.6% is the only other brand above it. The publisher's own reading of that number is the sentence worth keeping: nothing an owner does meaningfully changes whether a sealed system fails, so it is the one figure on this page you cannot maintain your way out of — which is why it matters more than the rest. If a compressor is on the table, the test that separates it from a $372 relay and the age rule for the call are in refrigerator compressor failure, and the diagnosis that leads there is in refrigerator not cooling.

Ice makers split the field in half. They are 21.7% of Samsung refrigerator repairs and 21.1% of Whirlpool's, with KitchenAid at 19.2% and Kenmore at 15.0% — against 11.3% for GE and 12.4% for Frigidaire, roughly half the rate. That is a maintenance story rather than a build story: the ice path is water, filter and airflow, and it is the one refrigerator failure that responds to what an owner does. The evaporator fan column is flat across every brand that reports it, between 0.9% and 3.1%, which is the profile of a wear part rather than a design difference.

Washing machines: the control board divides them

Share of that brand's completed washing machine repairs.

GE and Whirlpool send their largest share of work to the control board, at 5.5% and 4.3%, while the same part is 1.5% of Samsung washer repairs. Read the other way, LG's two reported columns are a drain pump at 5.7% and a door seal at 4.3%, and those are the two failures on a front loader that an owner can actually influence: what goes into the drum and whether the door and its boot are cleaned and dried. This is the table to hold against a quote. If a board is named on a Samsung washer it is the rare finding in that brand's records, and the reasonable question is what was tested before it was named; on a GE washer it is the ordinary one. The faults behind those columns are on a washer that will not start and a washer that will not drain.

Dishwashers: the pump is the brand tell

Share of that brand's completed dishwasher repairs.

Whirlpool's wash pump at 7.0% is the highest single share in the dishwasher set and more than three times Samsung's 2.1%, while Samsung's drain pump at 5.0% is the highest drain figure there. Two different pumps on two different brands is a design difference, not a user one — and it is the kind of thing a page ranking brands by survey cannot see. GE is the control board outlier here too, at 4.3% against Bosch's 1.2%.

Dryers and ovens: the heating element splits, the igniter does not

Share of that brand's completed dryer repairs.

Share of that brand's completed oven and range repairs.

Heating elements run from 1.6% on LG dryers to 8.5% on Samsung's, a fivefold spread on a part that costs about the same everywhere — and a heating element is the classic victim of a restricted vent, which is the maintenance case made in the dryer fire data. Whirlpool's drive belt at 6.3% against Samsung's 3.0% is the same kind of gap on a mechanical wear part.

The gas igniter is the exception, and it is the most useful number on this page. It accounts for 9.9% to 11.2% of oven and range calls across Samsung, Whirlpool and GE, with only Frigidaire lower at 7.3% — the most brand-independent failure in the publisher's records, and an unusually tight band for a part that is replaced in three different manufacturers' ovens at nearly the same rate. A gas oven that will not light is therefore the one diagnosis where the brand tells you almost nothing, and the fault guide is the whole answer: oven not heating, and where the igniter is the confirmed cause, a burner that will not light. Control boards behave the other way in ovens, at 5.7% for both Whirlpool and Frigidaire against 3.1% and 3.9% for Samsung and GE.

The ranking you should not use

The same publisher runs the calculation that most "most reliable brands" articles are actually built on: which brands generate the most service calls. Across 64,973 repairs the order is Samsung 21.8%, Whirlpool 21.1%, GE 18.9%, Frigidaire 10.2%, Maytag 8.0%, LG 7.3% of branded repair calls. That is a market-share table. It says how many of each machine are out there and how many were sold, and it warns against itself in as many words — the publisher labels the figure "do not confuse this with reliability". Its own records also show the five most common brands landing within $32 of each other at the median repair, so the badge moves the bill less than the fault does.

What the two tables together do give an owner is a way to evaluate a quote rather than a brand. Hold the named part against the brand's own column. A control board quoted on a Samsung washer, a compressor on a Whirlpool refrigerator, or a heating element on an LG dryer are each the uncommon finding within that brand's own repairs — not impossible, but worth one question about what was tested before the part was named. A compressor on an LG or a Kenmore, an ice maker on a Samsung or a Whirlpool, or an igniter on any gas oven are the ordinary finding, and the quote should look ordinary too. the appliance-level medians to compare against are on what a typical repair costs by appliance, the fault-level bands are on the guide for the symptom, and the question of how often a brand fails at all — which needs a denominator this dataset does not have — is on what three brand datasets actually measure.

Where these numbers come from

One repair company's own job records, published as the Bozmanfix Repair Index: 65,924 completed appliance repairs across Florida, Georgia, Texas, North Carolina, South Carolina and Pennsylvania, of which 18,670 carried both a brand and an appliance attribution and form the base for every percentage on this page. Brands are attributed from the model plate and the manufacturer's registration data rather than from the customer's recollection, and each figure is the share of that brand-and-appliance combination's own completed repairs, so a popular brand is not penalised for being popular. The publisher marks the data set free to cite with attribution, which is why the tables are reproduced here as published rather than paraphrased. Dollars are deliberately absent from this page: it is about which part fails, and the medians for repairing those parts are on the pages linked above, each with its own sample. Nothing here is a prediction about your machine — a brand-level share is not a diagnosis, and the fault guides exist because the symptom is what you actually have in front of you.