Untitled

New free tools for HVAC and appliance repair operators.

Ask ten techs what insurance their state requires and you'll get ten different answers, most of them wrong. The rules are scattered across licensing boards and statutes, and the person who knows the most about your coverage is usually the one selling it to you. That's a lousy spot to negotiate from.

So we built two free tools to move that information back to your side:

1. A 50-state requirement checker — pick your state, get a citation-backed breakdown of what your licensing board and your customers actually require, including where workers' comp and surety bonds come in.

2. An insurance cost estimator — enter your trade, state, revenue, crew size, and vehicles, and get directional ranges for GL, workers' comp, commercial auto, and more, built from published carrier benchmarks (HVAC general liability averages ~$941/yr; appliance repair ~$876/yr).

No ads in the tools, no affiliate kickback, no pay-to-play. We don't sell insurance. If you haven't checked your coverage against an independent benchmark this year, run both — worst case, you confirm you're already fine.

Five refrigerants are competing to replace R-410A under the HFC phasedown, and a 2025 Springer study finally ran them all through the same head-to-head analysis: R-32, R-290 (propane), R-454B, R-454C, and R-513A, scored on energy, exergy, and total environmental impact.

The efficiency winner: propane. R-290 hit a COP of 7.91 at the study's optimal operating point against 7.49 for R-410A, a 5.6% edge, plus the best exergy efficiency of the group at 0.52. R-454C brought up the rear at 5.80.

But the emissions scoreboard flipped the podium. On Total Equivalent Warming Impact, which counts refrigerant leakage AND the emissions from the electricity a system burns, R-454B came out cleanest at 13.3 metric tons CO2e, about 16% below R-410A. R-290 landed at 14.8.

The twist: R-454C finished dead last at 18.4 tons despite carrying the lowest GWP of any blend in the test. Poor efficiency means more power-plant emissions year after year, and that swamps the low number printed on the cylinder. TEWI counts the meter.

The practical read: the OEMs bet well. Carrier and Trane went R-454B, which takes the emissions crown. Daikin and Goodman went R-32, which ran about even with R-410A while cutting GWP by roughly two thirds. And propane's efficiency win keeps its A3 asterisk; charge limits keep it out of US ducted splits, so you'll meet it in window units and European monoblocs instead. R-513A, the lone A1 nonflammable in the lineup, ran mid-pack on every measure, which is about the trade you'd expect for skipping flammability entirely.

None of this changes which cylinder lands on a service truck. The manufacturers decide that. It does change how you explain the transition to a customer staring at a changeout quote, because two of the five candidates beat the outgoing refrigerant on the metrics that matter.

The study: Ashour et al., International Journal of Air-Conditioning and Refrigeration, 2025.

Your Washing Machine Has Been Keeping a Diary

Modern appliances log their own failures, and most will happily tell on themselves if you know where to look. Connected Samsung machines push error codes to the SmartThings app the moment they happen. GE appliances report through SmartHQ, and GE even runs a separate pro app (SmartHQ Service) that lets technicians command individual fans, valves, and compressors from a phone, on models going back to about 2012.

LG's trick is the best one. Smart Diagnosis encodes the machine's fault data as a series of acoustic tones played from a speaker on the appliance. Hold a phone running the ThinQ app up to it, and the app decodes the song into a diagnostic report. It works even on machines that were never set up on Wi-Fi, which is most of them. Your washer can literally sing its symptoms.

Whirlpool (plus Maytag and KitchenAid) takes the opposite approach. The consumer apps handle cycle notifications and basic alerts, but the company keeps its professional tools inside a portal reserved for authorized servicers. Independent techs work the onboard diagnostic modes documented in the tech sheet hidden inside the machine, which store error histories and run automatic test cycles. Less flashy than a phone app, still effective once you know where the sheet lives.

The catch with all of it: an app only reports what the control board believes, and the board only knows what its sensors say. A drifted $15 temperature sensor will produce a confident, wrong diagnosis in any app on earth. Good techs use connected data to decide where to point the meter first, then verify before ordering parts. Apps narrow the search. Meters close the case.

Why Nobody Applies to Trade Jobs Anymore, According to Peer Review

Every shop owner has a theory about the empty applicant pipeline. Now there's a published one. A study by Ibrahim Osman and Hossein Ataei in ASCE's Journal of Legal Affairs and Dispute Resolution in Engineering and Construction traces the post-pandemic skilled labor shortage to four causes that stacked on top of each other.

The pandemic hit the workforce directly. Older tradespeople took early retirement faster than the pipeline could replace them. Immigration, a quiet source of trade labor for decades, fell off. And underneath it all: a generation of defunded shop classes and vocational programs, while school counselors pointed every capable kid at a four-year degree.

None of it reversed when the emergency ended. The shortage is structural, the authors find, with the predictable results: delays, higher costs, burnout, safety concerns, and job postings that nobody qualified answers.

The study is about construction, but appliance repair and HVAC hire from the same pool and lost the same shop classes. The appliance trade alone is estimated to be short roughly one in five technicians.

The interesting part is the fix list, sorted by who controls it. Rebuilding vocational education? A 10-to-15-year payoff that depends on legislatures and school boards. Employer-run training? That pays off next year. The shops with the lowest vacancy rates quit waiting for experienced techs to show up on job boards and started building their own through structured apprenticeships, community college partnerships, and certification sponsorship. There's a recruiting angle hiding in that model too: earn-while-you-learn pathways let a 19-year-old start drawing a paycheck immediately instead of taking on four years of tuition debt, and separate research has found the wage gains hold up over time.

Blackstone just agreed to pay about $2.5 billion for Champions Group, a heating, plumbing, and electrical company. That's roughly 18.5x its $140 million in EBITDA. In May, Redwood Services bought the five-company Sierra platform with its $100M+ in residential revenue. Private equity is paying real money for businesses that fix furnaces.

So what's a normal shop worth? Published 2025 data: small owner-operated shops trade around 2.6-3.5x seller's discretionary earnings. Residential HVAC companies with $500K-1M in EBITDA averaged about 6.3x. At $5-10M in EBITDA, roughly 10.8x. Notice what that means: the bigger business earns 10x more AND each dollar of earnings is worth ~70% more. Size compounds twice, which is the whole engine behind PE consolidation: buy a platform at 11x, bolt on small shops at 3-8x, mark everything at the platform multiple.

What drives the multiple? Recurring revenue above all. Champions reported 150,000 active members; Sierra's announcement led with 19,000+ maintenance agreements. Then technician retention, no customer over 10-15% of revenue, clean books, and the brutal one: owner-independence. If customers ask for you by name and you price every job, a buyer is purchasing a job, not a company.

The useful twist: everything on a buyer's checklist also makes a shop better to own if you never sell. Clean financials, an agreement base, a manager who can run the day, two weeks off without the place burning down.

New NBER research built on a decade of U.S. tax returns (10+ million Americans who earned gig income, 2012-2021) finds that gig workers are about twice as likely to start their own businesses as comparable non-gig workers. And they overwhelmingly start them in the same field they gigged in.

The study, by economists Denes, Lagaras, and Tsoutsoura, found the effect strongest among younger and lower-income workers, with about three-quarters of it coming from first-time entrepreneurs. The authors' interpretation: the gig is on-the-job learning. Customer acquisition, pricing, scheduling, all absorbed one dispatched job at a time.

The firms these founders start launch bigger (roughly 23% more revenue, 39% more employees than comparable startups) but are about 3 percentage points less likely to survive three years. The survivors outperform.

For appliance repair and HVAC, this describes the Thumbtack-to-independent-shop pipeline that owners already watch happen. The tech running weekend platform jobs is learning exactly what customers pay and what leads cost. Once the platform's cut becomes visible, the math on going independent writes itself.

For established shops it cuts both ways: more hungry new competitors, but also future subcontracting partners, acquisition targets, and experienced hires when ventures fold. And if your own tech is moonlighting on platforms, the data says it's a countdown clock, not a loyalty problem.

Combustion analysis: what the numbers actually mean

An analyzer reads what the flame is doing inside the heat exchanger, which your eyes can't. The short course for furnaces and water heaters:

O2: 6-9% on natural gas. Lower = running rich. Higher = excess air stealing efficiency.

CO air-free: the number that matters. Dilution hides CO, so the analyzer corrects it: CO x 20.9 / (20.9 - O2). 100 ppm measured at 14% O2 is really 300+ air-free. ANSI Z21.47 allows 400 ppm air-free as a pass/fail ceiling, but field best practice calls 100 ppm the action line, and a healthy furnace runs under 50. CO that climbs as the blower starts is the classic heat exchanger red flag.

Stack temp: 270-370°F above supply air on natural draft, 170-270°F above on induced-draft 80% furnaces. A condensing furnace should run around 100-130°F at the flue. If a 90%+ unit shows 200°F, it isn't condensing and something's wrong.

Where to probe: 80% furnace: drill the single-wall flue 12-18 inches above the inducer, before any elbow, and cap it after. 90%+: use the factory test port. Atmospheric water heater: get the probe upstream of the draft hood, because dilution air past the hood makes a sick unit look healthy.

Why bother: a tune-up without numbers competes on price. Before-and-after readings justify the burner cleaning, the gas pressure adjustment, and the heat exchanger referral, and a documented reading protects you when the customer declines the repair.

Interpreting high CO: high O2 with high CO points at flame impingement or dirty, misaligned burners. Low O2 with high CO points at overfiring or starved combustion air. Same CO number, opposite fixes, so check manifold pressure (3.5 inches water column on natural gas) and clock the meter before adjusting anything.

What actually happens to the energy bill when a house swaps a gas boiler for a heat pump? Four economists measured it with real utility data instead of guessing. NBER Working Paper 33036 (Bernard, Hackett, Metcalfe, and Schein, 2024) tracked UK households through the conversion using data from Octopus Energy.

The results: gas consumption down roughly 90%. Electricity up 61%. Total household energy consumption down about 40%. Emissions fell 36% in the first year, with lifetime savings around 68% as the grid gets cleaner.

The part worth stealing for your own house: the rate plan experiment. Octopus runs a time-of-use tariff built for heat pump homes, with cheap windows outside the evening peak. Households on it cut peak-hour electricity draw roughly in half by preheating during the cheap windows and coasting through the expensive ones. A smart thermostat handles the scheduling. And the shifting held up on the coldest days of winter, which is when grid operators worry most.

Translation: what a heat pump costs to run depends on your electricity rate almost as much as the hardware. Pair the install with the right tariff and a scheduled thermostat, and you're playing a different game than your neighbor on a flat rate. Most California households are already on time-of-use rates whether they've noticed or not.

Bonus finding: the UK's heat pump subsidy returned about £1.24 in value per pound spent, useful ammunition in the rebate-funding fights happening on this side of the Atlantic too.

The usual caveats: UK homes run hydronic radiators off boilers, the housing stock is older and leakier, and a British cold snap is a mild week in Minneapolis. Exact percentages won't map one-to-one onto a US furnace-and-AC changeout. The direction holds, though, and the rate-plan lesson travels just fine.

The paper: NBER Working Paper 33036.

Flat-rate or time-and-materials? Most repair shop owners never actually choose. They inherit whichever pricing model their first boss used and run it for decades.

The difference matters more than most owners think. A flat-rate book is a menu of task codes, each priced from a standard labor time, a fully loaded hourly rate (wage plus taxes, truck, insurance, overhead, profit), and parts markup baked in. The customer sees one number, quoted before the wrench comes out.

T&M looks simpler and leaks everywhere. Diagnostic time gets under-logged. Drive time goes unbilled. Clock rounding always goes down. And when the customer can see the meter, they negotiate the meter. An illustrative case: a shop billing $140/hr that under-logs 25 minutes per job across eight daily jobs walks away from roughly $14,000 a year per truck. On work the tech actually did.

Both models get gamed by techs, in opposite directions. Flat-rate plus commission pay invites cherry-picking the fat task codes. T&M invites slow-walking, because nobody gets paid more for finishing early. You manage whichever distortion you can measure.

And customers? Homeowners want a fixed number up front. Commercial accounts often want the hourly meter and read flat-rate as padding. That's why a lot of healthy shops run hybrids: flat-rate residential menu, contracted hourly for property managers, disclosed T&M for genuinely unpredictable jobs.

GE Appliances is now running more than 800 AI agents across its factories, logistics, and supply chain. The company announced the deployment with Google Cloud's Gemini Enterprise platform at Cloud Next '26 in Las Vegas on April 22.

The scale is worth sitting with for a second. This is the Haier-owned manufacturer behind some of the most common appliances in American homes, with 15,500 employees and 27 million service parts shipped every year through 700-plus service partners. The AI agents monitor equipment health in real time, generate shift summaries in minutes instead of hours, handle routine communication with more than 600 suppliers, and mine customer feedback for defect patterns.

The number that matters most if you ever need an appliance fixed: GE Appliances credits the system with a 25% reduction in parts backorders in 2025. Parts availability is the hidden bottleneck of appliance repair. When the part shows up faster, your refrigerator gets fixed on the first visit instead of the third, and the repair company stops eating the cost of wasted trips.

"AI is now integral to the way work gets done at GE Appliances," said Mandar Deo, the company's chief digital officer.

There's a catch for the repair trade, though. Faster defect detection means more mid-production fixes, which means two machines with the same model number can carry different parts inside. Service documentation has to keep pace, and historically it hasn't. Technicians are going to be checking service bulletins a lot more often before condemning a board that doesn't match the tech sheet. The revision in their hand might be the factory's fix, not the fault.

The world generated a record 62 billion kg of e-waste in 2022, and only 22.3% of it was documented as properly collected and recycled. That's from the Global E-waste Monitor 2024, the UN's official scoreboard on the problem (ITU and UNITAR, lead author Kees Baldé).

The trend lines are the alarming part. Generation is up 82% since 2010, climbing about 2.6 billion kg a year, and rising five times faster than documented recycling. By 2030 the report projects 82 billion kg annually, with the recycling rate slipping toward 20%. The metals in 2022's e-waste alone were worth an estimated $91 billion, including $19 billion in copper and $15 billion in gold.

For the appliance trade, the category data matters. Heavy, bulky equipment gets recycled at the highest rates; small gadgets are the problem children (20.4 billion kg, roughly 12% recycled). Nobody ships a toaster to a recycler. A dead fridge at least gets hauled somewhere.

Temperature exchange equipment (fridges, freezers, ACs, heat pumps) carries special stakes: roughly 500 million kg of copper sits in that category's compressors and cables, and refrigerant vented during informal processing damages the climate and the ozone layer. By the report's framing, every EPA 608 card in the field is a small piece of climate infrastructure.

The bigger point for working techs: repair sits upstream of recycling entirely. A recycled refrigerator recovers some copper and steel. A repaired one keeps the whole machine in service, embedded materials, manufacturing energy, refrigerant charge and all. Every completed service call is waste deferred, and it's the one line on the e-waste ledger that never appears, because it never became waste.

The report projects that raising global collection and recycling to 60% by 2030 would generate benefits exceeding costs by $38 billion. Repair sits outside that model entirely, avoiding the waste rather than processing it.

Stop Shooting the Control Board

The control board is the most expensive part in most appliances, and it's also the most wrongly accused. Distributors see a steady stream of returned boards that tested fine, swapped in by techs (and DIYers) who fired the priciest part at a weird symptom and hoped.

Here's the thing: a control board is a decision-maker, not a magician. It takes power in, reads sensor inputs, and switches outputs to motors, valves, and heaters. So the honest diagnosis is three checks, in order. Is correct power arriving at the board's own connector? Are the sensors feeding it accurate readings (a drifted thermistor makes a healthy board refuse to run a compressor)? And when the board should be commanding a load, is voltage actually leaving the output terminal? Good power, good inputs, no output: that board is guilty. Anything less is a guess.

When boards really do die, the failures are visible more often than you'd think. Pitted or welded relay contacts. Cracked solder joints under heavy components. A scorched MOV after a power surge. Bulged capacitors that cause random reboots and shape-shifting error codes.

And the great impostor: house wiring. A loose neutral or bad ground at the outlet makes a board reset, scramble, and throw phantom codes that look exactly like electronic failure. One screw tightened in an old receptacle has saved more than one $400 board purchase.

If the board is truly dead, a new OEM part isn't the only road. New boards run $250-600 and up, while mail-in rebuild services run $100-250 and can rescue discontinued boards on older machines that would otherwise head to the landfill. A tech with a soldering iron can often replace a $4 relay in ten minutes and skip the wait entirely.

Connecticut's right-to-repair law is in effect as of July 1. Manufacturers of electronics and appliances sold in the state must now make documentation, parts, and tools (including diagnostic software and updates) available to consumers and independent shops on the same footing as their authorized servicers. Availability runs three or five years depending on product price.

For appliance techs, the headline is that appliances are squarely covered. A shop in Hartford can now request the service manual and diagnostic software for a covered range or washer and cite state law if the manufacturer stalls. Carve-outs include boats, ATVs, generators, power tools, EV charging stations, and farm/construction equipment. Enforcement runs through the Connecticut Unfair Trade Practices Act, so the state carries the stick, not private lawsuits.

Texas follows September 1. HB 2963, the first consumer electronics repair law in a Republican-led state, requires parts, tools, and documentation on fair and reasonable terms, priced at or below what authorized providers pay. Documentation must stay available three years for products priced $50-$99.99 and seven years above $100. One gap: Texas does not prohibit parts pairing. Colorado and Washington, whose laws took effect January 1, do restrict those software locks, and Colorado's explicitly covers washing machines and dishwashers.

The wave isn't slowing. The Repair Association counted more than 33 right-to-repair bills filed across 13 states in early 2026 sessions, including three in Massachusetts and two each in Florida and Missouri.

The practical playbook for shops in newly covered states: request the documentation you've been denied. Cite the statute when refused. Keep records of parts orders that get slow-walked or priced above what authorized servicers pay. Complaint-driven enforcement only works if somebody complains, and a documented refusal builds the record that triggers state action.

Right to repair keeps winning in statehouses. Connecticut's law takes effect July 1, joining California, Minnesota, Colorado, New York, and Oregon. So here's an uncomfortable question: what do manufacturers do about it?

A 2023 Management Science paper by Chen Jin, Luyi Yang, and Cungen Zhu built the first formal economic model of right-to-repair legislation, treating it as what it is in practice: a drop in the cost of independent repair. Then they watched what a profit-maximizing manufacturer does in response.

The answer: it doesn't sit still. At intermediate production costs, the manufacturer first cuts new-product prices to make replacement more tempting than repair (a volume strategy). Push independent repair costs low enough and it flips, raising prices while offering free repair itself to capture the product's lifetime value (a margin strategy).

Both moves compete head-on with independent shops. One puts a cheaper new unit next to your repair quote. The other puts a free OEM repair next to it.

The welfare math is just as tangled. Depending on production costs, right to repair can help consumers and hurt the environment, do the opposite, or hurt both. The authors even identify a possible lose-lose-lose where manufacturer profit, consumer surplus, and the environment all decline together.

That doesn't make the laws bad. The model only isolates pricing, and it ignores things like faster local repair and consumer autonomy. But it retires the assumption that parts access alone decides who profits.

The paper: Jin, Yang & Zhu, Management Science 69(2), 2023.

Home warranty work splits the appliance trade into two camps: guaranteed volume that built my business, versus discounted labor for a dispatcher that treats my schedule as inventory. A former claims adjuster's take in ServiceMag's new guide: both are right, and the difference is the shop's stage of life, not the warranty company.

How the money actually moves: the homeowner pays a trade call fee at the door ($100 or $125 at American Home Shield, $65-125 across other nationals), which the contractor typically keeps. Everything past diagnosis runs on flat approved rates set in the servicer agreement, with shops reporting pressure to discount 50% or more off retail. Parts often ship from the warranty company's own supply chain, so the markup that carries 25-40% of a retail ticket's gross profit compresses toward zero. All-in, a typical warranty ticket lands at 50-70% of the retail equivalent.

Then add the authorization bottleneck. You diagnose, submit, and wait days or weeks for approval while the customer's fridge stays dead. One-visit jobs become two or three, and payment can lag weeks or months.

When it works: new shops. Empty truck slots, zero acquisition cost, dense territory-based routes, and steady volume through thin January weeks. A $200 ticket against an empty afternoon is $200 you didn't have.

When it poisons: shops at capacity, where each warranty call displaces a $375 retail job. Plus non-solicitation clauses that make the homeowner the warranty company's customer, not yours, while their one-star review (for the adjuster's denial) lands on your Google profile.

Before signing, read for the full rate schedule, payment terms in days, the recall window, parts sourcing rules, chargebacks, required acceptance rates, and the non-solicitation clause word by word. And track warranty work as its own P&L line from day one, because blended averages hide the bleeding.

Electrolux is now defending the Frigidaire gas range recall in court. Reato v. Electrolux Consumer Products, Inc., filed May 14 in federal court in Illinois, alleges the recalled ranges pose an unreasonably dangerous burn hazard and that the company sold them despite 'superior knowledge about the defective nature' of their bake burners (ClassAction.org).

The recall behind the suit: about 174,800 Frigidaire gas ranges in the US (plus 5,300 in Canada), sold at Lowe's, Home Depot and Frigidaire.com between June 2025 and January 2026 for $630 to $2,700. Delayed ignition of the oven's bake burner lets gas pool in the cavity before it lights. Electrolux logged 62 malfunction reports, 30 with burn injuries, per Consumer Reports. Affected models include FCFG3083AS, FCRG3083AD and PCFG3080AF, serial numbers VF52200000 through VF54399999.

The part the consumer press keeps skipping: the fix is labor. The remedy is free professional in-home installation of a replacement bake burner, which makes this recall, functionally, 174,800 scheduled service calls routed through Electrolux's authorized network. For authorized shops in the affected retail footprint, that's months of guaranteed volume.

For independents: if a 2025-2026 Frigidaire gas range shows up on an unrelated call, check the serial against the recall range before touching the oven. If it's in scope, document it on the ticket and point the customer to the hotline at 866-291-7633 (CPSC recall 26-333). The customer should stop using the oven until the repair is done; the cooktop burners run on separate ignition circuits and remain safe. And don't field-fix a recalled bake burner with an aftermarket part. It can muddy the owner's free remedy and, with litigation pending, their legal position too.

Electrolux hasn't commented publicly on the lawsuit. Class certification will take months either way, and the burner-swap campaign keeps rolling through the network in the meantime.

The most rigorous field test ever run on home weatherization found that the savings were about half the cost of the work. Not half of what was hoped. Half of what was paid.

In 2018, economists Meredith Fowlie, Michael Greenstone, and Catherine Wolfram published a randomized controlled trial covering more than 30,000 Michigan households in the federal Weatherization Assistance Program: attic insulation, air sealing, furnace work. Real upgrades, real meters, real winters.

The numbers: upfront costs ran about twice the measured energy savings. The engineering models used to qualify the upgrades predicted roughly 2.5 times the savings that actually appeared on utility bills. Average annual rate of return: about negative 9.5%.

And here's the twist. It wasn't the rebound effect. The researchers found no evidence that weatherized homes ran warmer than the control group. People weren't spending the savings on comfort. The models were just wrong about how real houses behave.

The caveats are real too: low-income single-family homes, one state, free upgrades through a federal program. Equipment swaps and owner-funded retrofits can perform closer to projections. But the takeaway stands for anyone quoting or buying efficiency work: modeled savings are a ceiling, not a midpoint.

If a contractor promises a five-year payback off audit software, the right follow-up is "based on the model, or on measured bills from jobs like mine?" The good ones know the difference.

The paper: Fowlie, Greenstone & Wolfram, The Quarterly Journal of Economics 133(3), 2018.