How We Rate: The Full ApplianceWatchDog Methodology

Every score on this site comes from the same formula. It's published here in full — weights, anchors, edge-case rules, and two worked examples with the arithmetic shown — so you can check our math and argue with specific numbers instead of vibes.

Overall score = (Coverage & Plans × 0.30) + (Pricing & Value × 0.20) + (Service & Claims Experience × 0.20) + (Customer Reputation × 0.20) + (Transparency & Flexibility × 0.10)

Each sub-score runs 1.0–5.0. The weighted total is rounded to one decimal. Coverage carries the most weight because a warranty that won't cover your failure is worthless at any price; Reputation and Service together carry 40% because a contract is only as good as the company's habit of honoring it.

For each sub-score, here is what a 5 looks like versus a 2. Brands land between the anchors; the anchors don't move.

1. Coverage & Plans — 30%

What's actually covered, at what caps, with which exclusions and add-ons.

• A 5 looks like: caps high enough to replace real appliances ($7,000+ per item or equivalent), broad standard coverage, meaningful add-on menu, pre-existing-condition or rust coverage, few silent exclusions.
• A 2 looks like: caps that can't replace what they "cover" ($500-per-item territory), key appliances missing from base plans, depreciation clauses that shrink payouts with age, exclusion lists longer than the coverage list.

2. Pricing & Value — 20%

Premium plus service fee versus the protection you actually get. Cheap isn't value if the cap is hollow; expensive can be value if caps are real.

• A 5 looks like: total annual cost (premium + realistic fee usage) at or below the ~$73/mo 2026 market average for clearly above-average caps; predictable flat fees; honest list pricing rather than promo bait.
• A 2 looks like: premium pricing for ordinary caps, $125+ fees stacked on top, teaser promos that reset high at renewal, or multi-year prepay pushed hard at signup.

3. Service & Claims Experience — 20%

What happens between "it broke" and "it's fixed": claim intake, dispatch speed, technician network, repair completion, payout speed.

• A 5 looks like: 24/7 intake that works, technicians dispatched in days, first-visit fix rates customers praise, cash-outs paid in weeks not months, consistent answers between reps.
• A 2 looks like: week-plus waits for dispatch, repeat visits without resolution, cash-out checks that take a quarter to arrive, contradictory claim-status answers, or documented patterns of engineered denials.

4. Customer Reputation — 20%

The verdict of people who already paid: BBB complaint files and customer scores, Trustpilot, Google, ConsumerAffairs — read the way we describe below, not averaged blindly.

• A 5 looks like: BBB A-range with a near-empty complaint file for the company's size, 4.5+ third-party scores without polarization, praise themes that match the marketing.
• A 2 looks like: thousands of complaints over three years, BBB customer scores under 2/5, polarized review distributions, or any recent regulatory action for claims handling (see the hard rule below).

5. Transparency & Flexibility — 10%

Whether the company tells you the truth before you buy, and lets you leave.

• A 5 looks like: sample contracts downloadable before purchase, caps and exclusions in plain English, published fees, easy pro-rated cancellation, no games with review platforms.
• A 2 looks like: contract only visible after signup, caps buried, aggressive retention on cancellation, sales reps overpromising coverage the contract doesn't contain, or paid placements dressed as organic ratings.

BBB complaint volume. We report raw counts (3-year and 12-month) but score the pattern: volume relative to company scale, whether complaints get substantive responses, and what the themes are. A giant with 19,000 complaints and a small firm with 150 aren't scored on raw count alone — but past a certain density, scale stops being an excuse. We also weight the BBB customer review score over the corporate letter grade; a B+ company with a 1.8/5 customer score is a 1.8/5 experience wearing a B+ badge.

Trustpilot polarization. An average hides the shape. A 3.8 built from 60% five-star and 20% one-star reviews describes a company that's fine until it matters — small claims sail through, big claims hit a wall. We read distributions, check for heavy review solicitation, and flag platforms where the brand holds paid-partner placements. Polarized distributions cost Reputation points even when the average looks respectable.

Regulatory actions — the hard rule. A government enforcement action about claims handling or sales practices (settlement, consent order, or judgment) within the past 24 months caps Customer Reputation at 2.0 and excludes the brand from ranked recommendations, regardless of overall score. Denying wrongdoing in the settlement doesn't lift the cap; completed reforms plus clean subsequent data can, at a scheduled re-review. This is the rule that removed Choice Home Warranty from our rankings.

Parts stores, marketplaces, and on-demand services are scored on the same five dimensions, translated to what each category promises:

Cross-category scores are comparable as "how well does this company keep its specific promise" — a 4.4 parts store keeps its promise about as reliably as a 4.4 warranty keeps its bigger one. The comparison table's category badges keep the apples labeled.

Angi reaches more local pros than anything else on this page, and the experience of being routed to them is the weakest in our lineup. The formula holds both truths:

It ranks last because a marketplace hands you a stranger, not a contract: reach is worth something, but nothing here is guaranteed by Angi itself. The gap between its Coverage 4.0 and Reputation 2.9 is the review in one line: it will find you someone, and who that someone turns out to be is the risk you carry.

Choice's sticker prices are competitive (~$46.67–$55/month, $100 fee) and availability is wide. But: ~11,500 BBB complaints in three years, a low-1s-to-2s BBB customer score, a polarized 60/20 Trustpilot split — and a $11.8M Arizona AG consumer-fraud settlement (February 2026) in which state evidence showed a covered "normal wear and tear" diagnosis rewritten to justify denial. The hard rule caps Reputation; the documented denial engineering and multi-year prepay sales tactics drag Transparency:

A 2.9 with a regulatory cap means: fully reviewed, not ranked, no referral links. The full review explains what re-inclusion would require.

Primary, checked directly and access-dated on every page: company contracts, pricing pages, and state availability lists; BBB complaint files and customer scores; Trustpilot, Google, and ConsumerAffairs distributions; state regulator records (e.g., the Arizona Attorney General); SEC/company filings for ownership facts; published 2026 cost benchmarks (HomeGuide, Angi, HomeAdvisor, This Old House) and the NAHB component-life study for lifespan data. Paraphrased customer quotes carry platform and approximate date. What we don't use: brand press kits as fact sources, affiliate-network "performance" data, or any metric a brand can buy.

Full re-verification of every score and volatile figure quarterly; out-of-cycle updates for regulatory actions, ownership changes, and major price moves. Every page shows its "as of" stamp. Between reviews, reader reports via /contact feed the next cycle.

Most competitor methodologies disclose that "compensation may influence placement" — read their fine print. The observable differences here: our formula and anchors are public; complaint counts appear in reviews, not behind links; our #1 pick's negative data (a BBB pattern-of-complaints note) is disclosed in its review; and a top-paying brand sits excluded at 2.9. We may earn commissions like everyone else (disclosure) — the difference is that our scores are falsifiable. Check the math above; if we're wrong somewhere, tell us and we'll fix it in public.