Equipment costs in the trades are up roughly 60 percent since 2019. Carrier posted increases of up to 8 percent for 2026 alone, and the A2L refrigerant transition pushed refrigerant pricing up 50 to 200 percent on some lines. Set that against how most shops actually price: one flat-rate book update a year, usually in January. By July, every quote runs against costs that have moved twice since the book was built.
A stale price book is the quietest leak in a trades profit and loss statement. A missed call announces itself the same day. Mispriced line items show up six months later as a gross margin that drifted three points south while revenue looked fine. This week the tooling that fixes it moved from enterprise add-on to mainstream platform feature, and that changes what a five-truck shop should do about pricing before August.
A stale book leaks five figures a quarter
Run the math on a $2.4 million HVAC shop carrying 55 percent direct costs. That is $1.32 million a year in equipment, materials, and labor flowing through the price book. If costs creep 6 percent mid-year and the book holds still, the shop eats the difference: about $79,000 a year, or $6,600 a month, gone without losing a single job. No dashboard flags it because every job still closes. The leak only surfaces at year-end when the accountant asks why gross margin fell from 52 to 49.
The pressure is not letting up. Service Nation’s 2026 Trend Report, released in early July, called out pricing complexity as a defining challenge for home service contractors this year: equipment costs stacked with rebates, financing offers, and the refrigerant transition. Industry pricing trackers put typical 2026 quote movement at 12 to 18 percent on equipment and 15 to 20 percent on labor. Annual repricing was survivable when costs moved 2 percent a year. At this volatility, cadence is the whole game.
Shops respond to cost creep in one of three ways. Some eat it and find out in December. Some slap a blanket 10 percent on everything, which overprices the competitive items and underprices the specialized ones. And some let techs improvise in the driveway, which turns pricing into a personality trait. A monthly reprice against real cost data beats all three, and the tooling for it just got cheap.
AI pricing benchmarks just went mainstream
On July 15, Housecall Pro launched its first trade-specific packages for HVAC, plumbing, and electrical shops. The headline feature for pricing: real-time AI benchmarks on price book items, built from data across more than 100 million jobs run by its 200,000-plus customers. Open a line item like a 40-gallon gas water heater swap and see where your price sits against the market. The packages also bundle a mechanical KPI dashboard, membership plans techs can sell from the field, credit card surcharging, and commissions set by price book line item.
Benchmarking is not a new category. ServiceTitan sells Pricebook Pro, a managed flat-rate content and vendor catalog service, quote-priced and typically landing in the few-hundred-dollars-a-month range. Profit Rhino sells a maintained flat-rate database starting around $149 a month. What changed on July 15 is placement. Benchmark data now ships inside the platform a large share of sub-ten-truck shops already run, at package pricing, rather than as a separate purchase the owner has to discover, evaluate, and defend.
There is a second reason pricing discipline got urgent: your customers now check. ACHR News reported this month that homeowners are using AI tools to research equipment options, pricing, and repair-versus-replace decisions before they ever call a contractor. The buyer on the other side of your quote increasingly arrives with a number in mind. A book that sits 15 percent over market without a stated reason does not read as premium anymore. It reads as unprepared, and the homeowner has receipts.
Be clear about what a benchmark answers. It tells you whether your book is off by 3 percent or 15, in which direction, on which items. It does not know your costs. Your supplier invoices know your costs. A benchmark without a cost refresh just tells you how confidently everyone else is guessing. Use it as one input in a monthly routine, not as the routine itself.
The 90-minute monthly reprice
The fix is a cadence, not a project. Here is a workflow sized for an owner or ops manager, run once a month on the first Monday. The first pass takes an afternoon because the book is years behind on some items. Every pass after runs about 90 minutes, and the AI does the tedious parts. Work the list in order.
- Pull the top 50 price book items by trailing 90-day revenue. ServiceTitan, Housecall Pro, and Jobber all export this in a couple of clicks. Fifty items usually cover 80 percent of revenue. Ignore the long tail this month; it can stay wrong a little longer.
- Refresh costs from real invoices. Photograph the last three supplier invoices for each key SKU and drop them into ChatGPT or Claude with a one-line prompt: “Extract part numbers, descriptions, and unit prices into a table.” Fifteen minutes replaces the data-entry afternoon that used to kill this habit. Update the cost field on each of the 50 items.
- Check benchmark deltas where you have them. Flag anything more than 10 percent off market. Priced low with healthy demand: raise it. Priced high: decide whether response time, warranty terms, or install quality earn the premium, and say so in the estimate.
- Reload the labor rate. Fully loaded tech cost, wages plus payroll taxes plus truck plus benefits, divided by billable efficiency. A tech costing $52 an hour loaded at 70 percent efficiency costs $74 per billable hour before a dollar of overhead or profit. If wages moved this quarter and the labor rate did not, the book is lying.
- Version the book. Date-stamp every change and log the effective date. Next month’s pass becomes a diff instead of an audit, and a price question from a tech in the field gets answered in one lookup.
Benchmarks are a floor check, not a strategy
Now the tradeoff nobody selling the feature will volunteer. A benchmark is an average of the market, and the market average includes shops that are quietly going broke. Matching the median means importing other people’s mistakes into your book. Your overhead, your warranty exposure, and your answer rate are not average, so your prices should not be either. Treat the benchmark like a smoke alarm. It tells you something needs attention. It does not tell you what temperature to keep the house.
Watch the commission mechanics too. Paying techs by price book line item, another feature in the new Housecall Pro packages, aligns the field with the book and rewards techs for selling memberships and add-ons. It also creates a new incentive to discount for an easy close if you allow field-side price edits. Pair line-item commissions with a rule: no unapproved discounts below book, ever. The book only protects margin if it is the single source of truth.
The decision to make this month is cadence and ownership. Name the person who owns the price book, put the 90-minute reprice on the calendar for the first Monday of August, and pick the data source. On Housecall Pro, that means asking about the new trade packages and turning benchmarks on. On ServiceTitan, it means pricing Pricebook Pro against Profit Rhino. Still quoting from a laminated binder built in 2023? The binder is the decision. Costs already voted, twice this year. The only question left is whether your prices get a vote too.
