Home Services and Contractors: Market Evolution

What the trades are actually standing on

Before asking how discovery is changing, it is worth checking whether there is still a market underneath the discovery. There is. The Harvard Joint Center for Housing Studies projects total US homeowner spending on home improvement and repairs will reach roughly $518 billion by the end of 2026, with growth easing from about 2.1% at mid-year to about 1.6% by year-end as the housing market softens. That is deceleration, not contraction. On the labor side, the Bureau of Labor Statistics recorded roughly 33,000 new construction jobs in January 2026, about 25,000 of them in nonresidential specialty trade contracting, though the agency's own later revisions suggest the early-2026 pace was softer than first reported.

The two data points together describe a market that is still growing, just more slowly, and still hiring, just less than the first headline suggested. That matters because it is easy to read the discovery-channel disruption described in the rest of this study as evidence that the home services market itself is shrinking. The government data available says otherwise: the volume of work is holding up. What is changing is the path a homeowner takes to find someone to do it.

The click that stopped happening

The single most concrete number in this study is also the starkest. SparkToro's analysis of Similarweb clickstream data found that US Google searches ended without any click 68.01% of the time in the first four months of 2026, up from 60.45% in 2024. On searches where an AI Overview appears, click-through rates fall by roughly 60% relative to searches without one. This is a single-vendor clickstream methodology, so we hold it at emerging rather than established, but the direction and the size of the move are hard to argue with: a large and growing share of the research a homeowner does before hiring anyone now happens entirely inside the search results page, with no visit to any business's website at all.

For a contractor whose marketing has been built around ranking for a click, that is the uncomfortable part of the picture. It does not mean the search itself stopped mattering. It means the moment where a business earns attention has moved earlier, into whatever gets summarized on the results page, and away from the moment a homeowner lands on a website.

More than two out of three US Google searches now end without a single click, up from three in five two years ago.

But the local pack still wins the moment that matters

The most important nuance in the entire dataset comes from Whitespark's controlled study of 540 queries across Houston, Phoenix, and Denver, spanning six local-service industries including plumbers. It did not find AI Overviews displacing local search generally. It found the two features splitting the work by query type. On narrow, transactional queries like plumber near me, the local pack appeared 93% of the time and the AI Overview only 15%. On informational queries, that flips almost completely: AI Overviews appeared 92% of the time versus 6% for the local pack. Hybrid queries leaned even further toward AI Overviews, at 97%. Averaged across every query type in the study, AI Overviews appeared 68% of the time and the local pack 39%.

Read as a single average, that looks like AI has already won. Read by query type, it says something more specific and more useful: homeowners still use a different mental mode, and get a different answer, when they are ready to hire someone tonight than when they are researching a project in general. This is a single vendor's study, three cities, six industries, run before the newest AI Mode rollout, so we hold it at emerging. It is also the most rigorous evidence available on the specific question this study set out to answer, and it argues for treating the AI-answer layer as additive to local search rather than a replacement for it, at least for the bottom-of-funnel query that actually books a job.

Ask an engine to explain plumbing, and it answers. Ask it to find a plumber near you tonight, and it still, for now, hands you a map.

Reviews are doing more work, not less

If AI answers were quietly replacing the need to research a business at all, review reading should be declining. It is not. BrightLocal's 2026 Local Consumer Review Survey found 97% of consumers read reviews when evaluating a local business, and 41% now say they always do so before choosing one, up sharply from 29% the year before. Trust in those reviews has kept pace: 49% of consumers now say they trust online reviews about as much as a personal recommendation from a friend or family member.

That is a market leaning harder into review evidence at exactly the moment its search behavior is otherwise fragmenting across new surfaces. For a contractor, it means the review profile is not a legacy channel to maintain out of habit. It is one of the few pieces of a business's reputation that a homeowner is provably still reading directly, in growing numbers, whichever surface brought them there.

The review layer gets a filter, and a federal rule

Two forces are converging on the review itself. The first is summarization: BrightLocal found 82% of consumers now read AI-produced summaries of a business's reviews rather than, or in addition to, the reviews themselves, and 23% say they would rely on that summary alone. That compresses months or years of individual reviews into a handful of sentences an engine chooses to write, which raises the stakes of any single bad or unresolved review and lowers the value of simply accumulating a large volume of average ones.

The second force is regulatory. The FTC's final rule banning fake reviews and testimonials took effect nationwide on October 21, 2024. It prohibits purchased or fake reviews, undisclosed insider reviews, fake company-controlled review sites, review suppression, and fake indicators of social influence, including fake reviews produced with AI tools, with civil penalties up to $51,744 per violation. Taken together, the review a homeowner reads is now more likely to arrive pre-digested by an engine and more legally dangerous to fabricate or manipulate than it has ever been. The response to both is the same: earn reviews that are true, and respond to the ones that are not flattering, because both the summarizer and the regulator are now watching the same corpus.

What we actually know about reviews and revenue

The deepest academic evidence in this study is also the oldest and the most narrowly scoped. Michael Luca's Harvard Business School research on Yelp found that a one-star increase in a business's rating was associated with a 5% to 9% increase in revenue, an effect concentrated in independent, non-chain restaurants, alongside evidence that chain market share fell as Yelp penetration rose, consistent with online reviews substituting for the reputational signal a chain's brand name used to provide on its own. This is well-established, peer-adjacent research, which is why we hold it at established rather than emerging.

It is also about restaurants: a low-consideration, frequent, low-price purchase category. No equivalent peer-reviewed study exists for a high-consideration, infrequent, high-price purchase like a roof replacement or a kitchen remodel. The mechanism Luca describes, reviews substituting for reputation when a buyer chooses between businesses they do not already know, plausibly applies to a homeowner picking a contractor too. But applying his 5% to 9% figure directly to trades work would overstate what the evidence actually supports. Reviews likely move revenue in home services as well, and nobody has yet run the study that says by how much.

A one-star swing moved five to nine percent of revenue for the independent restaurants Michael Luca studied at Harvard. Nobody has run the equivalent study for the trades.

Contractors are betting on the customers they already have

ServiceTitan's 2026 Residential State of the Trades report, fielded by the independent research firm Thrive Analytics across 1,000 residential contractors, found a real shift in where effort is going: 53% of contractors are prioritizing retaining existing customers over acquiring new ones, which only 31% named as the priority, even though 66% still call revenue growth their top overall goal. That is a market responding to rising acquisition friction by leaning on the relationships it has already paid to build.

The same survey found contractors are cautiously optimistic about AI as a tool for their own operations rather than just a channel disrupting how they get found: 74% see AI as key to business efficiency, but only about 25% are currently using it. Among those who already have, 48% report increased productivity and 45% report real time savings. Adoption is still early, but the direction is consistent with everything else in this study: the trades are adapting to a discovery environment they cannot fully control by investing harder in the parts of the relationship they can, existing customers and their own operating efficiency.

What a lead actually costs, in the one place we could verify it

Lead economics is where the marketing-agency compilation content around this topic is thickest and the verifiable data is thinnest. The one figure that survived direct sourcing in this study comes from SearchLight Digital, which tracked $6.72 million in Google Local Services Ads spend across 888 home-services contractors and 126,650 leads in a single month, finding an average cost per lead of $53. That is a real, dated, named benchmark, but it covers one paid channel, Google's own Local Services Ads, aggregated by one vendor.

We went looking for the broader claim repeated across the industry, that a single homeowner lead from a marketplace like Angi or Thumbtack is often sold to several competing contractors at once for a cost in the tens to low hundreds of dollars, and that a majority of contractors report declining lead quality. We could not find a single named, dated, checkable survey behind either figure; it traced back only to a chain of marketing blog posts citing each other. We removed it rather than repeat it. The evidence is that Local Services Ads cost per lead is measurable and roughly $53, and that the broader question of what a booked job actually costs across every paid lead channel remains unmeasured, not settled.

The next homeowner is already researching, just not on Google yet

This Old House's 2025 homeowner survey, as reported by Fortune, found 60% of millennial and 56% of Gen Z homeowners had renovation plans for that year, compared with just 32% of baby boomers, with younger owners more focused on personal style and customization than the upkeep-and-repair priorities of older homeowners. The same survey pointed to social platforms, TikTok and Instagram named specifically, as increasingly shaping how younger owners plan and research a project.

If that holds, the research a homeowner does for a renovation is starting to happen well before anything that looks like a transactional search, in a format none of the search-visibility or review-trust data in this study was built to measure. That is a directional signal from a single secondary source, not an established fact, and it points at exactly the kind of upstream attention a Visibility Corpus needs to capture: not just how a business ranks or gets summarized at the moment someone searches, but where a market's attention is already sitting months earlier, on a platform most local marketing reporting never looks at.