Echelon Report 017: A service-search rebound is not a reason to discount every job

Finding

What changed since the last measurement: eight of the ten core service phrases gained national search interest during October 18–24. Furnace repair rose 21.4% from the preceding week, while HVAC replacement also rose 21.4%. Their year-earlier comparisons differed: furnace repair remained 13.6% lower, while HVAC replacement was 3.0% higher. These are Echelon Reports Database readings of relative search interest, not counts of customers or jobs.

For an HVAC owner, the useful decision is to separate repair inquiries from replacement inquiries before changing the offer. A shared weekly increase does not mean the two services have the same opportunity, customer economics or crew requirements. Check which inquiries become suitable appointments and profitable estimates before increasing spending or cutting prices.

Dated evidence table

Weekly and year-earlier changes in national relative search interest for furnace repair, HVAC replacement, roof repair and roof replacement. — U.S. relative search interest for October 18–24, 2026. Each phrase is compared with itself, using the October 26 retrieval. These changes are not local search volume, inquiries or jobs. Open the full-size chart.

Ten core services

Each row uses its own normalized 0–100 scale. Read across a row, not down the column. A larger score for one phrase does not establish more searches than another phrase. The comparisons below use the October 26 retrieval throughout, because a fresh retrieval can rescale earlier readings.

Service phrase	Oct 18–24	Oct 11–17	Oct 19–25, 2025	Weekly change	Year-earlier change
ac repair	34	32	37	+6.2%	-8.1%
bathroom remodel	73	63	87	+15.9%	-16.1%
foundation repair	49	36	51	+36.1%	-3.9%
furnace repair	51	42	59	+21.4%	-13.6%
hvac replacement	34	28	33	+21.4%	+3.0%
kitchen remodel	69	63	81	+9.5%	-14.8%
roof repair	60	63	68	-4.8%	-11.8%
roof replacement	53	56	47	-5.4%	+12.8%
water heater replacement	28	26	27	+7.7%	+3.7%
window replacement	46	43	49	+7.0%	-6.1%

Source: Echelon Reports Database. U.S. web-search interest, retrieved October 26, 2026. The year-earlier comparison is exactly 52 weeks earlier. National relative interest does not establish local demand, lead quality or advertising cost.

Four door services, outside the index

All four supplemental door phrases increased from the preceding week. Entry-door replacement rose 27.8%, but remained 62.3% below its year-earlier reading. Storm-door replacement rose 4.8% for the week and 7.3% against the year-earlier week. Those differences are a reason to keep service-specific records, not to treat doors as one uniform market.

Service phrase	Oct 18–24	Oct 11–17	Oct 19–25, 2025	Weekly change	Year-earlier change
door replacement	42	41	42	+2.4%	+0.0%
entry door replacement	23	18	61	+27.8%	-62.3%
storm door installation	64	48	83	+33.3%	-22.9%
storm door replacement	44	42	41	+4.8%	+7.3%

Window replacement stays in the original index basket. The four door phrases remain separate coverage and do not alter the composite. Small normalized readings can produce large percentage changes; they do not establish the number of prospective customers in a contractor’s territory.

Analysis and counterargument

Our analytical read is that HVAC owners should investigate the difference between repair and replacement interest before adopting a broad promotion. The two phrases moved together this week, but their comparisons with the corresponding 2025 week point in different directions. An owner offering both services needs to know whether inquiries differ in suitability, appointment availability and expected contribution after direct costs.

The strongest counterargument is geographic: national searches can move differently from a local service territory. Weather, homeowner research and changes in search behavior may influence the readings, and this evidence does not identify a cause. The weather component covers October 11–17, an earlier week than the search comparison, so it does not establish a weather explanation for this week’s rebound.

A competitor may respond with a lower advertised price, quicker appointments or more prominent financing. Matching a discount without reviewing the reason customers decline an estimate could reduce margin without addressing the obstacle. Treat those potential responses as planning assumptions, not observations of a named competitor.

Roofing also shows why a broad rebound is an incomplete description. Both roofing phrases declined from the preceding week. Roof replacement remained 12.8% above the year-earlier week, while roof repair was 11.8% below it. Owners should use their own service mix when interpreting the national basket.

Owner decision

Run a two-week inquiry and estimate review from October 26 through November 8, with repair and replacement recorded separately. Compare that period with the preceding two weeks using the same definition of a suitable inquiry: work the business offers, inside its territory, with a realistic opportunity to serve the customer. Keep spending and the offer stable while identifying the main constraint.

The owner, dispatcher and marketing manager need the inquiry log, estimate outcomes, appointment calendar and direct job-cost estimates. Record the requested service, broad location, source, response time, appointment outcome and stated objection. Keep customer details in the business’s private systems.

Observed constraint	First action	Measurement	Continue, revise or stop
Suitable repair inquiries do not become appointments	Review response delays and what the request page promises	Suitable inquiries, response time and appointments	Revise a repeated obstacle; do not assume a discount is the answer
Replacement estimates stall	Compare stated objections with scope, payment information and scheduling	Estimates, accepted work and reasons for declining	Improve the explanation where records support it; retain unknown reasons
Bookings exceed available capacity	Reconcile crew availability before expanding promotion	Promised dates, backlog and cancellations	Stop expansion that displaces better work or creates unrealistic commitments
Suitable bookings fit capacity and cover costs	Preserve the productive approach	Booked contribution and completed collections	Continue while economics hold; revise if contribution falls

Contribution worksheet

Contribution is the amount left after direct job and acquisition costs, before general overhead. Use the business’s actual prices and costs rather than an assumed industry ticket. Expected booked sales are not collected cash; reconcile them after completion.

Record for each service	Repair	Replacement
Suitable inquiries	____	____
Appointments and estimates	____	____
Jobs booked	____	____
Expected booked sales	____	____
Direct labor, material and subcontract costs	____	____
Attributable acquisition and handling costs	____	____
Expected sales minus both cost rows	____	____
Crew capacity for promised dates	____	____

This recommendation requires staff time and consistent records, not an additional advertising budget. If a later expansion test is undertaken, stop when expected sales no longer cover direct job, acquisition and handling costs, or when the work exceeds available capacity. Set any additional required contribution floor using the business’s own overhead and profit needs. This is an analytical test, not guaranteed profit. A small sample remains inconclusive and does not establish improved conversion.

Next observation

At the November 9 review, compare suitable inquiries, appointments, estimates, bookings and contribution by service. Check whether the national furnace and replacement comparisons persist, but let the business’s results govern spending. If inquiries rise without suitable bookings, investigate the request and estimate process before purchasing additional traffic.

Scope method source notes

The complete six-component calculation is 93.6 against its component-specific references. That is a descriptive composite, not a forecast of jobs or revenue. Each component receives one-sixth weight. Values below 100 mean below the stated reference, not necessarily a decline from the preceding edition.

Component	Index	Observation	Reference
Search interest	78.2	Oct 18–24	Preceding 52 complete weekly basket readings, median
Existing-home turnover	100.9	September 2026	Original September 2023–2025 releases, mean
New-housing permits	95.6	September 2026	September 2023–2025 observations, mean
Mortgage financing	93.4	Oct 22	Corresponding 2023–2025 weekly mean, inverted
Weather exposure	108.1	Oct 11–17	Corresponding three-year station-first median
Matched review counts	85.1	Sep 13–Oct 26	Sep 6–13 rate on the same 929 identities

September existing-home turnover uses NAR’s original national release at 3.99 million seasonally adjusted annual sales, compared with the original September releases for 2023, 2024 and 2025. Their values are 3.96, 3.84 and 4.06 million. This keeps later revisions separate from the original-release comparison.^[1]^[3]^[4]

The 2025 reference comes from NAR’s original release syndicated by GlobeNewswire. An annual sales rate describes the pace after seasonal adjustment, not the number of homes sold during one month.^[5]^[1]

Mortgage financing uses the October 22 observation of 7.32%. Permits use the September observation of 1.400 million units at an annual rate. Mortgage observations are weekly, while permits are monthly; neither becomes an October weekly event because this report appears on Monday.^[7]^[8]

Weather uses all ten approved stations and 70 distinct station-days for October 11–17, with three corresponding earlier-year weeks. The October 18–24 target contains only 50 of the required 70 station-days, so the accepted maximum-one-week lag applies. NOAA supplies the underlying daily maximum and minimum temperature records.^[9]

The review panel retains the same 929 matched business identities. Its September 13–October 26 interval spans 43 days and shows 7,647 positive count changes and 1,303 downward changes, net 6,344. The comparison uses a seven-day-equivalent rate, not a reconstructed weekly path. The September 6–13 reference remains a single provisional interval, not an established seasonal norm. Reviews are not jobs or revenue.

The method retains the October 5 historical comparison break. Earlier published prints are not restated. No license-database counts, company totals, regional forecasts or local demand estimates are used in this edition. Search and review observations are attributed to the Echelon Reports Database, with source receipts retained privately.

Relevant next step

Decide which work deserves more attention

KKS website and advertising management can help connect service-specific offers, request paths and follow-up with the work an owner wants to win. Start with the repair or replacement category where the records show an actionable constraint, rather than a general promise of more traffic.

Discuss the work you want more of with KKS

Read the previous Echelon Report, the report archive and the ERDI methodology.