How we measure our data

Which datasets are behind Leadity's statistics?

The datasets every Leadity statistic comes from, with their size, window and read date

A window is the span the data covers; a read date is the day we counted. The owner-name and email checks cover shorter windows inside Q3 2026, so quote each with its own dates. The Q3 2026 data gives line types two ways: as shares of the checked phone numbers (36.8% mobile across all 245 trades) and as shares of all businesses, including the 15.1% that list no phone (31.2% mobile). Quote each with its base.

Source Leadity

How do we collect the Google Maps data?

The Q3 2026 data counts what real Leadity customer scrapes found. A scrape runs when a customer orders a list: it searches one trade in one area, one Google Maps search at a time, each anchored on a ZIP or postal code. The data holds the businesses those scrapes found between 1 July and 30 September 2026.

Test accounts are left out, and each business counts once. A business is one Google Maps place, however many scrapes or neighbouring searches found it. A business found in two trades counts once in each trade, so the trade counts add up to more than the dataset total.

The data is customer-chosen scraping, not a census. Customers chose the trades and areas, so the figures lean toward the trades and places ordered most. Leadity's Q3 2026 Google Maps data holds only US (2,727,139) and Canadian (55,370) businesses, because the UK and Australia went on sale on 4 October 2026, after the window closed. A country needs 3,000 or more businesses for a figure.

A trade gets its own report page when its data says something no sibling report already says. Every one of the 32 trades shown in detail holds 3,000 or more businesses. The trades without a page sit in one table on business lists by trade, and all 32 are in one table in the annual report, State of Google Maps Business Data 2026.

Source Leadity

What we publish, and what we never publish

• Shares, medians and averages per search, such as 35.9% of businesses with no website or 17.5 businesses per ZIP search, and counts of businesses per trade, state or country.
• Owner-name and email rates with the businesses checked, such as 32.0% of 2,397 HVAC contractors checked between 15 and 30 September 2026.
• No figure for a small group. A state needs 3,000 or more businesses of a trade, a country 3,000 or more businesses, and an owner-name rate 500 or more businesses checked.
• Nothing about one business, person or customer. No business is named and no phone number is shown in any figure.
• No figures about Leadity's own business: how many scrapes, searches, orders or lists customers run, how many customers or accounts there are, credits used, billing, revenue or growth.
• Test accounts are left out of every figure.

Source Leadity

Medians, ranks and state tables

A median here is the middle of the 32 trades, not of the businesses. With 32 values it is the mean of the 16th and 17th printed trade figures, rounded half up to one decimal. A trade report prints the median of each measure beside its own figure.

A rank counts the trades with a higher value, plus one. Rank 1 is the highest value, and a tie is marked on the page.

A state enters a trade's state table once it holds 3,000 or more of the trade's businesses in the window. The table shows at most the 12 biggest. A trade with no such state has no state table, and the free preview offers the same states only (3,000 or more businesses). The state gap is the best state's mobile share minus the worst state's, for the 17 trades that have two or more states.

The reading for 1,000 verified mobiles is 1,000 divided by the mobile share. The result is the businesses the scrape reads, and dividing it by the businesses per ZIP search gives the ZIP searches. You pay for the mobiles delivered, not for the businesses read.

Source Leadity

Census employer firms beside the trade figures

Trade reports with a Census industry code print the Census count of employer firms for context. The figures are from the U.S. Census Bureau's Statistics of U.S. Businesses for 2022, read 19 October 2026. They are not a Leadity dataset, and a trade with no clean industry code prints none.

A Census firm is a business with at least one paid employee. One-person businesses with no payroll are not counted, and many of them have a Google Maps listing. A firm with sites in several states counts once in each, so state counts do not add to the U.S. count.

Google Maps businesses per Census firm compares listings with firms, not companies with companies. Leadity counts only the businesses customers' scrapes found, and the industry code can be wider or narrower than the trade, so the ratio is a rough guide to how far a business count stands from a firm count.

What does each number mean?

Every metric Leadity publishes from its Google Maps and number block data, defined

A share of checked phone numbers and a share of businesses are different figures. A share of businesses keeps the businesses with no phone in its base, so its six phone shares (mobile, landline, VoIP, toll-free, unnamed, no phone) add up to 100%, give or take rounding.

Source Leadity

How does Leadity validate a phone number?

Source Leadity

How is a validated mobile different from the raw mobile share?

The raw share describes the listings; validation decides what you receive. In Leadity's Q3 2026 Google Maps data (1 July to 30 September 2026), 36.8% of the checked phone numbers were mobiles, counted over every business the scrapes found.

With the mobile-only filter on, 100% of delivered numbers are mobiles. A number whose line type comes back landline, VoIP, toll-free or other/unknown is left out and never billed.

The raw share sets how much map an order needs. Roofing contractors are 31.8% mobile, so 1,000 verified mobiles mean the scrape reads about 3,145 businesses, about 133 ZIP searches.

So a raw share never describes a filtered list, and a filtered list never describes the market.

Source Leadity

How does the no-website check work on orders?

The 35.9% no-website share counts one field: the website on the Google Maps listing. The share covers the 2,782,509 businesses in Leadity's Q3 2026 Google Maps data (1 July to 30 September 2026), and a business whose listing shows no website counts as no website.

The no-website filter on an order goes further. It checks the listing, social media and other online sources, and leaves out any business with a website posted anywhere, even when the website is missing from its Google Maps listing. This is the one place Leadity looks beyond Google Maps, because a business that already has a website is not a prospect for web services.

The cross-check has no measured share. No figure on this site counts how many businesses it removes, so a delivered no-website list can be smaller than the listing share suggests. Quote the figure as a listing statistic: 35.9% of the businesses found on Google Maps show no website on their listing.

Source Leadity

How do we count the US and Canadian number blocks?

The source is public. Block assignments are published by localcallingguide from NANPA's filings, free to anyone. Leadity loaded a copy of 1,539,378 rows on 20 September 2026 and read the figures on 22 September 2026. Each row is one area code, one exchange, one thousand-block digit and the carrier that holds it.

A block becomes a line type by its carrier type. A carrier type of W is wireless, I is a landline incumbent, and a competitive carrier is decided from its name against a keyword list.

The dataset counts blocks, not phones. A block is 1,000 consecutive numbers, so the shares describe the number plan, not the phones in use. A number ported since it was assigned keeps its block, so a block says what a number was assigned as, not where it sits today.

The headline figures: 40.5% of blocks are wireless, 31.3% landline and 28.2% VoIP, and 47.5% of 231,623 exchanges hold more than one kind of line. Give every block its exchange's most common type and 327,459 of 1,307,755 blocks (25%) get the wrong one.

NANPA block records are part of the carrier data Leadity's validation uses in step 2, allocation; they are not the whole check. The full counts are in US phone line types.

Source Leadity

How was the cited pages study done?

On 14 October 2026 Leadity ran 19 questions about Google Maps lead tools as AI web searches. The searches returned 178 result URLs, and 47 content pages were picked from them: the ones the answer text leaned on. With the 12 URLs ChatGPT cited in a logged-out baseline, that made 59 pages.

Fetching and counting. Pages were fetched as raw HTML first, one request at a time per host, 2 seconds apart, with robots.txt respected. 53 of the 59 could be measured, and 52 form the published tables, with Google's terms page left out. A script read each saved page with JavaScript off and counted words, headings, tables, dates, schema and FAQ blocks. Word counts are approximate, about 10 percent either way.

Quoted passages. One person hand-coded about 130 passages the answers quoted, so those shares are approximate: about 37% carried a number with a unit.

Limits. One niche and one day. The study shows what cited pages share, not what caused a citation: domain strength, index coverage and topic match decide most of retrieval first. The results are in AI search statistics.

Source Leadity

What are the limits of these numbers?

• Listings are not a census. A business with no Google Maps listing is absent, so 35.9% describes listings, not every local business. Leadity
• Trades and areas follow customers' scrapes. The trades and places are the ones customers chose to scrape, not a sample built to represent the economy, so the figures lean toward the trades and areas ordered most.
• A search returns neighbours. A search for one trade can also return businesses whose own main Google category is a neighbouring one, so a trade's figures include some of them. Leadity
• A line type is not a person. The figures say what businesses publish on Google Maps, not who answers the phone.
• Blocks are not phones. The block count shows what a number was assigned as, and a ported number keeps its block.
• Every figure is a snapshot. Listings change after the window ends, so a figure is true for its window, not for today. A number ported to another carrier since the scrape keeps the line type it had then.
• The owner-name and email checks cover shorter windows, 15 to 30 September 2026 and 16 to 30 September 2026, so quote them with their own dates.
• The cited pages study covers one niche. It asked about Google Maps lead tools only, and one person coded the quoted passages.

Source Leadity

How do we round the figures?

• Shares carry one decimal place, as in 36.8%. In a sentence a trailing zero can drop: 17% and 17.0% are the same figure.
• Whole-number shortcuts are rounded, not measured. 37% and 36 in 100 are 36.8% and 35.9% rounded.
• Each share is rounded on its own, so the six phone shares of one trade can add up to a tenth of a point above or below 100%.
• Businesses per ZIP search keeps one decimal in tables (23.6 for roofing contractors) and is rounded in sentences (about 24).
• Counts are exact. 2,782,509, 2,241,697 and 1,539,378 are counts, not estimates; the word about marks every approximate figure, such as the study's passage shares.

Source Leadity

How often is the data refreshed?

When we have new counts, not on a fixed calendar. A new read replaces the old figure, its window and its read date together, on every page that quotes it.

The window and the read date belong to the data; the Updated date belongs to the page. A page's Updated date moves only when its copy or a figure changes.

The lists you order are not snapshots. Each list is scraped from Google Maps when you order it, so it is as current as the listings are that day.

How should you cite Leadity's data?

Quote a figure with its dataset, its window and a link to the page it appears on. To point readers at the method, link this page.

The figures are aggregates of public listings: no business is named and no phone number is shown. The Google Maps business data page offers its figures as a free CSV, each with its dataset and read date.

Source Leadity

How do you report an error?

Email support@leadity.io with the page, the figure and what you think is wrong.

How we correct errors, and how we disclose that we rank Leadity on our own comparison pages, is in our editorial policy.

Missing a trade? Ask at the same address: we publish a trade's figures once it holds 3,000 or more businesses in the data.