Original research
The Customer Who Never Reaches Your Website
How local customers find and judge independent service businesses across search, Maps, directories, reviews and AI
- Key findings
- 6
FlashCrafterSign inBook a consultationBook a call
FlashCrafterOriginal research
How local customers find and judge independent service businesses across search, Maps, directories, reviews and AI
Picture a homeowner with water spreading across the laundry-room floor. She searches for an appliance repair company on her phone. Before she reaches any company website, Google may show an ad, a map, nearby businesses, star ratings, a directory page and perhaps an AI-generated answer. She is already forming a shortlist.
In the August 2026 review, company websites were often only one stop in a longer customer journey—and sometimes were not visible in the sampled results. Websites still matter, but they cannot carry the whole burden of helping a business get found and trusted. The practical question is where customers encounter your business before they call, and whether those places give them consistent, credible information.
The evidence
Third-party pages occupied most of the regular organic results measured, while subject websites had no median presence in those result sets.
Consumers commonly used more than one channel in a local-business search.
A number-one Maps result in one check did not establish broad market visibility.
Google began moving selected Local Services Ads accounts into Google Ads as Performance Max campaigns with pay-per-lead goals.
Search Console began reporting visibility in generative AI features for some sites.
AI Overviews appeared in a minority of the sampled local-service results.
| Number | Basis | |
|---|---|---|
| Third-party pages occupied most of the regular organic results measured, while subject websites had no median presence in those result sets. | Median third-party share: 77.1%; median subject-site presence: 0%. | Retained search snapshots for 16 local-service businesses, collected by August 14, 2026. |
| Consumers commonly used more than one channel in a local-business search. | Three-quarters used multiple channels; most who began with an AI tool continued elsewhere to verify the recommendation. | BrightLocal consumer research published July 9, 2026, about respondents’ most recent local-business search. |
| A number-one Maps result in one check did not establish broad market visibility. | Median best measured pack position: 1 among six businesses; median share of measured Maps visibility: 2.5% among 13 businesses. | Retained Maps measurements for separate groups of six and 13 businesses, collected by August 14, 2026. |
| Google began moving selected Local Services Ads accounts into Google Ads as Performance Max campaigns with pay-per-lead goals. | The first U.S. phase began in August 2026 for selected home and storefront service advertisers. | Google Ads Help guidance accessed August 14, 2026; first U.S. migration phase described as beginning in August 2026. |
Third-party pages included directories, review sites, marketplace pages and other sites the business did not own. The measures do not show that third-party sites caused a business website to be absent. Among a separate group of seven businesses, the median best organic position was 2; a best position and broad visibility are different measurements.
BrightLocal's survey supports a multi-surface customer journey. It does not establish which combination produces the most leads for a particular trade or market.
The position and visibility measures came from different eligible groups and cannot be combined into a cause-and-effect claim. They illustrate why one favorable rank check is not the same as dependable reach across a service area. Manual searches and rank trackers can also differ because local results depend on context.
Google said Search and Maps placements and the pay-per-lead model remain, while campaign management moves into Google Ads. Historical reports do not transfer automatically; weekly budgets become daily budgets, and manual maximum cost-per-lead bidding is deprecated. This platform change is not evidence of improved lead volume, quality or return.
Google's reporting views can show appearances by page, country, device and time. The report measures visibility, not booked jobs, and does not explain why a page appeared. Owners should check whether it is available in their account.
This measures whether an AI Overview appeared. It does not identify the cited source, show whether a business appeared in the answer, or measure traffic or leads. Google says foundational SEO still matters for generative Search and recommends unique, useful content; this sample cannot establish how that advice affects ranking.
Method
Reviewed retained search and market snapshots for 16 non-restaurant local-service businesses in a defined set of Google business categories. Coverage ran from July 15 through August 14, 2026; research was conducted August 14, 2026.
Combined only measurements with the same meaning across businesses. The sample is a broad local-service snapshot, not a benchmark for any one trade. Automotive, landscaping and appliance-repair groups were too small to publish separately.
Compared anonymous aggregate observations with current platform guidance and credited outside research. Outside sources helped identify topics to investigate; they did not establish what happened in the client sample. This is a one-month, cross-sectional snapshot and does not establish causes.
Client identity was pseudonymized before aggregation. Public results omit raw client rows and direct or quasi-identifiers. Numeric cells require at least five businesses; percentages and rates require at least ten. Small or rare cells are suppressed.
What it means
Across 16 local-service businesses, third-party pages occupied a median 77.1% of the regular organic search results measured, while the businesses’ own websites had a median presence rate of zero. This is an early, cross-category snapshot—not a benchmark for any one trade—but it shows why customers may form an opinion before visiting a company website.
For owners
What to act on
Check a valuable service in the neighborhoods you serve and note the directory, review, map and service pages customers encounter before your website. Fix inaccurate business information and unanswered trust questions on the pages customers actually see. Measure Maps visibility from several points across the real service area. A favorable pin in one check does not establish reach across the market. If your Local Services Ads account is migrating, export historical reports and record the weekly budget before conversion. Compare it with the daily budget afterward, verify account settings and track booked jobs as well as lead volume. Check Search Console for its generative-AI report. If available, record the first month as a baseline; treat appearances as visibility, not booked-job attribution. Organize marketing around one customer journey: discovery, credibility checks, contact and booked work. Start with the service that matters most financially, then track qualified calls, estimates, booked jobs and gross profit rather than treating channel averages as targets. If testing paid search, use the $13.23 median planning estimate only as an early pressure check, then replace it with actual cost per qualified lead. Separate high-intent services and locations so you can see which searches produce calls or bookings. Keep major directory profiles accurate, include review generation in normal follow-up, and create useful service and location pages that answer customers’ questions before they call.
For customers
What to notice
A business may appear in directories, review sites or Maps before you see its own website. Check whether its contact details, hours, services and recent reviews agree across the places you use. Compare information across channels before choosing: three-quarters of respondents in BrightLocal's July 2026 survey used multiple channels in their most recent local-business search, and most people who began with an AI tool continued elsewhere to verify the recommendation. A business that ranks first in one Maps check may not appear throughout its service area. If location matters, check results from where you need the service. An AI-generated answer or a Search Console appearance does not establish that a business is suitable or that it will provide a good service. Verify the business through its details, service information and reviews.
The August 2026 edition covers retained snapshots from July 15 through August 14, 2026, and was researched August 14. It includes 16 non-restaurant local-service businesses from a defined set of Google business categories. The sample is a broad snapshot, not a benchmark for any one trade. Automotive, landscaping and appliance-repair groups were too small to publish separately.
Businesses were included when a retained client research snapshot was collected no later than August 14, 2026, its Google primary category belonged to the defined local-service group, the measurement had the same meaning across included businesses, and the anonymous total met the publication threshold.
Only measurements with the same meaning across businesses were combined. Client identity was pseudonymized before aggregation. The public projection contains no raw client rows or direct or quasi-identifiers. Numeric cells require at least five businesses; percentages and rates require at least ten. Small or rare cells are suppressed rather than described as anecdotes.
Excluded from publication were restaurants, banquet venues, blank primary categories and categories outside the allowlist; client names, domains, addresses, place IDs, account IDs, campaigns, queries, screenshots and row-level observations; Local Services Ads performance because comparable account-level measurements were unavailable; and ChatGPT, Perplexity and other assistant rankings because no reproducible assistant probe was available.
The observations are cross-sectional. They do not establish ranking causes or predict future performance. The combined group cannot establish a benchmark for any one trade. Market CPC is planning evidence, not observed client campaign cost or return on ad spend. AI Overview detection shows whether that result type appeared, not whether it sent traffic or cited the business. AI-assistant visibility was not measured; crawler guidance is separate from performance evidence.
The median Google market-planner CPC estimate was $13.23 among eight businesses with comparable planning data. It is a planning estimate, not a client’s actual cost or a useful universal budget rule. The middle business in the 16-business sample had no paid-result block in its retained search snapshot; ads vary by search and location, so that does not mean ads were absent from the market.
This was the first monthly edition, so there is no month-over-month comparison for paid search, Local Services Ads, Maps, organic search or AI Overview presence.
This combined group cannot establish a benchmark for any one trade.
The observations are cross-sectional and do not establish ranking causes or predict future performance.
Market CPC is planning evidence, not observed client campaign cost or return on ad spend.
AI Overview detection shows whether that result type appeared, not whether it sent traffic or cited the business.
AI-assistant visibility was not measured; crawler guidance is kept separate from performance evidence.
The Maps position and visibility figures came from different eligible groups and cannot be combined into a causal claim.
BrightLocal's survey does not establish the best channel mix for an individual trade, market or business.
Comparable Local Services Ads account results were not available for the client group.
Google's generative-AI Search Console report was initially available to only a subset of websites and measures visibility, not booked-job attribution.
The AI Overview appearance rate does not identify citations, traffic or leads.
The median CPC figure is a planning estimate, not observed campaign cost or return.
In the August review, third-party pages—directories, review sites, marketplace pages and other sites the business did not own—occupied a median 77.1% of the regular organic results measured across 16 local-service businesses. The median presence of the businesses’ own websites in those sampled result sets was zero.
That is not a universal benchmark, and it does not mean every directory deserves your time. It describes a recognizable customer experience: people can encounter a business on pages the owner did not write and cannot fully control. A wrong phone number, abandoned profile, thin reviews or confusing service description can lose a job before the homepage gets a chance.
BrightLocal reported in July 2026 that three out of four consumers used more than one channel in their most recent local-business search. Among people who began with an AI tool, most continued to Google, social media or another source before deciding. Discovery is spread across channels, while trust is assembled through multiple checks.
Return to the homeowner with the leaking washer. If the directory page shows old hours, the Google profile describes the service differently and recent reviews go unanswered, the business creates doubt at each step before the phone rings. Search for one valuable service in the neighborhoods you serve, note the profiles customers actually encounter and repair those trust breaks first.
Owners often check a favorite search from the office, see the business near the top of Maps and assume the market is covered. One manual check cannot establish coverage across a service area.
The retained Maps measurements illustrate the gap. Some businesses reached a best observed position of number one, while broader visibility measurements remained thin. The median best measured pack position was 1 among six businesses with comparable position evidence; median share of measured visibility was 2.5% among 13 businesses. These figures came from different eligible groups, so they cannot be combined into a cause-and-effect claim. They show why a trophy screenshot is not the same as dependable reach across a service area.
The better question is: for which valuable services, and in which parts of the market, is the business present often enough to be considered? That question can guide useful service pages, accurate business information, deliberate review work and more honest reporting. Whitespark notes that manual searches and rank trackers can differ because local results depend on context; these observations support checking multiple locations, but do not prove any specific ranking factor.
Google began the first U.S. phase of moving selected Local Services Ads accounts into Google Ads as specialized Performance Max campaigns with pay-per-lead goals. The Search and Maps placements and pay-per-lead model remain, but management moves into Google Ads, weekly budgets become daily budgets, and historical Local Services Ads reports do not transfer automatically. Google says manual maximum cost-per-lead bidding is deprecated after migration.
For an affected owner, the administrative step is to export historical reports and record the weekly budget before Google converts it to a daily amount. That preserves a before-and-after baseline. After the move, verify the account settings. Google describes Local Services Ads ordering as an auction that considers bid, likelihood of producing a lead, responsiveness, relevance, profile quality and search context. Whitespark’s 2026 expert survey also highlighted budget and bidding, reviews, service selection and response times; that survey is expert opinion, not disclosure of Google’s formula. Comparable account-level results were not available for this business group, so this guidance was not tested against its client sample.
Google also began rolling out dedicated Search Console reporting for visibility in generative AI features. The report is available to a subset of sites and can show which pages appeared, where, on what devices and when. It does not turn AI visibility into booked-job attribution, but gives owners a baseline they can preserve and compare.
AI Overviews appeared in only a minority of the local-service searches reviewed: the median sampled appearance rate was 9.03% across 16 businesses. That percentage alone is not the story. The useful development is that owners can begin measuring whether their own pages appear in Google’s generative results while continuing the fundamentals customers use to verify a business elsewhere. Google says foundational SEO still matters for generative Search and recommends unique, useful content; this sample does not prove how that advice affects ranking.
If your Local Services Ads account is migrating, export historical reports and compare the converted daily budget with the weekly amount you recorded before the move.
Check Search Console for the generative-AI report. If it is available, record the first month as a baseline. If it is not, keep measuring normal search visibility and leads instead of buying a speculative AI package.
It would be easy to turn this into six new workstreams—ads, Local Services Ads, Maps, organic search, AI Overviews and ChatGPT. Most owners do not need six projects. They need one view of how a qualified customer discovers the business, checks its credibility, makes contact and becomes a booked job.
Channel averages are not operating targets. A plumber, dentist, mover and detailer can tolerate very different lead costs because jobs, close rates and repeat value differ. Judge a channel against the value of the work it produces: qualified calls, estimates, booked jobs and gross profit.
Start with the service that matters most financially. Follow its search journey across the places customers actually see. Repair the trust breaks, then measure qualified calls, estimates and booked work. That sequence follows the customer journey and connects visibility to the work the business wants.
Across the eight businesses with comparable planning data, the median estimated market cost per click was $13.23. This is a planning benchmark, not what a client actually paid. Use it to pressure-test an early budget, then manage campaigns using actual qualified leads and booked work.
The middle business in the 16-business sample saw no paid-result block in its retained search snapshot. Paid ads vary by search and location, so that does not mean ads were absent from the market. This is the first monthly edition, so there is no earlier paid-search benchmark to compare. Clicks can be expensive enough that broad, unfocused campaigns waste budget quickly. Use the estimate only as an early budget check, replace it with your own cost per qualified lead, and separate high-intent services and locations to see which searches produce calls or bookings.
Google describes Local Services Ads ordering as an auction that considers bid, likelihood of producing a lead, responsiveness, relevance, profile quality and search context. Raising a budget alone may not improve visibility if calls go unanswered, services are mismatched or a profile is incomplete. Confirm that services, hours, service areas, license details and business information are accurate. Respond to leads quickly and track booked jobs—not just lead volume—before changing bids.
The first U.S. migration phase began in August 2026 for selected home and storefront service advertisers. Google says campaign management moves into Google Ads, while pay-per-lead goals and Search and Maps placements remain. Historical reports do not carry over automatically, weekly budgets become daily budgets, and manual maximum cost-per-lead bidding is deprecated. This platform transition is not evidence of improved lead volume, lead quality or return. Comparable client LSA performance was not available for this first edition.
The median best measured pack position was 1 among six businesses with comparable position evidence; median share of measured Maps visibility was 2.5% among 13 businesses. The groups differ, and the results do not show that one rank caused a particular level of coverage. They show that a business could rank first in one Maps check and appear across only a small part of the measured search area.
The ranking seen at the office may not match what a homeowner sees five or ten miles away. Measure Maps visibility from multiple points across the real service area. Prioritize weak neighborhoods and relevant services instead of celebrating one strong rank check. This is the first edition, so a month-over-month Maps change cannot yet be calculated.
Directories, review sites and other third-party pages occupied most of the regular search results measured. Across 16 businesses, the median third-party share was 77.1%, while median presence of the local business website was zero. A separate group of seven businesses had a median best organic position of 2; a strong best result and broad visibility are different measures.
These figures do not mean third-party sites caused business websites to be absent. They do show that customers often discover local businesses on directories and review sites, making reputation and listing accuracy part of search visibility. Keep major directory profiles accurate, make review generation part of normal customer follow-up, and create useful service and location pages that answer questions people ask before calling. This first edition establishes a baseline; there is no earlier organic-search share to compare.
The median sampled Google AI Overview item-type presence was 9.03% across retained results for 16 local-service businesses. This records whether an AI Overview appeared; it does not measure traffic, identify a cited source or show whether a business appeared in the answer.
Google says foundational SEO still matters for generative Search and recommends unique, useful content. The sample confirms that AI Overviews appeared in local-service searches, but cannot establish how that guidance affects ranking. Keep core SEO strong, make business details easy to verify, and track AI Overview appearances separately from website traffic and lead results. Because this is the first edition, it cannot show whether appearances increased or decreased.
Google began rolling out dedicated Search Console views showing how often a site’s URLs appeared in generative AI features in Search and Discover. The reports can break down appearances by page, country, device and time, and were initially available to a subset of websites.
The report measures visibility, not booked jobs, and does not explain why a page appeared. Check whether it is available in your account. If it is, record a baseline; continue to track leads separately. The report does not replace lead tracking.
Customers may use an AI assistant to build a shortlist, then confirm a choice in Google, Maps, reviews or a business website. Reliable assistant measurements were not available for this edition, so no visibility result or comparison is claimed. Allow OpenAI’s search crawler if you want public pages to be eligible for ChatGPT search, and keep business facts consistent across your website, Google profile and trusted directories. Crawler access does not guarantee visibility or leads.
Choose the service whose next ten jobs would matter most. Trace how a customer finds it, what they see before calling, where trust breaks and which contacts become booked work. Keep that scorecard for the next month.
When September arrives, you will have a record of whether the journey became easier for the customers you want to reach.
Questions
Keep reading
Vendor pages change. Figures are as published on the capture date; confirm before deciding.
| Search Console began reporting visibility in generative AI features for some sites. | Available to a subset of websites. | Google Search Central announcement published June 3, 2026 and accessed August 14, 2026. |
|---|
| AI Overviews appeared in a minority of the sampled local-service results. | Median sampled appearance rate: 9.03%. | Retained Google results for 16 local-service businesses, collected by August 14, 2026. |
|---|