Research
Jacksonville Local Search Research, Counted
We ran 85,853 searches, lookups and checks for two studies of Jacksonville-area businesses in October 2026: how accurate their directory listings are, and what a lead from a Google ad costs in 10 local industries. This page counts that work from our run logs, re-runs included, and says what we left out.
What's in the 85,853?
- 40,344
- Google and Bing searches we ran
- 42,149
- Google search terms screened in Keyword Planner
- 1,710
- listings and directory pages checked by a person
- 1,201
- Google Business Profile lookups
- 449
- rows checked with search and some website automation
- 85,853
- in all, counted November 1, 2026
Which studies is it from?
Business listing accuracy
We checked 600 businesses, drawn at random from 10 industries, against their own Google Business Profiles. 42% had at least one directory listing that contradicts Google, and 16% had none we could find.
Google Ads cost per lead
We screened 42,149 search terms to find the buying searches in 10 industries, then priced the top 100 from Google's own bids. A plumbing lead from a Google ad comes to $473 to $1,680 by our estimate.
What do we count?
Every search we ran, including runs we threw away and ran again, because each one was work. Each line has its run ids in our records.
| Work | Count | What's in it |
|---|---|---|
| Searches for business listings | 30,632 | Six runs measuring the 600, four measuring the map pack 400, two map pack tries we threw away and redid, a re-check, a repair of the directory site searches, and 49 of those re-run for Angi and Thumbtack listings. 1,084 are Bing searches from a test in the first run. |
| Google Maps searches to draw the samples | 3,770 | Four runs of the sample frame (20 searches from each of 29 points; the first three were discarded) and the draw for the map pack 400. |
| Searches to measure demand | 5,942 | The top 100 buying searches from all 29 points, run twice (the second time with the town in each search), and two smaller location tests. |
| Google Business Profile lookups | 1,201 | One for every business measurement, the redone tries included. |
| Search terms screened | 42,149 | Every term Keyword Planner returned for 94 services in the five counties: 29,474 kept as buying searches, 12,675 dropped with a reason. |
| Checked by a person | 1,710 | 1,652 listings checked in a browser across nine sheets, and 58 directory pages counted by hand. |
| Checked with some automation | 449 | One sheet of failed searches, re-run with search and some website automation. Counted on its own line, never as a check by hand. |
What we left out
- 1,201 Yelp lookups. Our data vendor refused every one, so nothing came back.
- A trial of another vendor, and passes that re-read pages without running a new search.
- Google's monthly search estimates. They measure the market, and we didn't run those searches.
Why 244,593 searches a month, not 541,440?
Adding Google's monthly estimate for all 29,474 buying searches we kept gives 541,440. Google gives every close variant its main search's volume: "pedi dentist" and "pedia dentist" each carry "pediatric dentist"'s 1,072, so the sum counts the same people several times. Counted once per group of variants, the 10 industries see 244,593 buying searches a month across 13,643 distinct searches in the five counties, by Google's estimates. That's the figure we use.
How did we pick 600 businesses?
- 29search points across Duval, St. Johns, Clay, Nassau and Baker counties
- 8,783Google Maps results for each industry's two biggest buying searches, run from every point
- 3,924different businesses in those results
- 2,424passed the checks: in the industry, with a phone and a street address, inside the metro
- 600drawn at random, 60 per industry
To compare the businesses Google shows first with the rest, we drew a second group of 400: 20 per industry from the top three map results and 20 from places 10 to 20. 98 of them were also in the 600, so the studies measured 902 different businesses over 1,201 measurements.
Why check by hand?
Our checker reads search results, and search results leave things out. Once people had checked 1,652 listings in a browser, the share of the 600 with a wrong listing went from 27% to 42%, and the share with none found fell from 30% to 16%. Every figure we publish uses the corrected counts.
Businesses with at least one wrong listing
Businesses with no listing found
Who ran it?
Narsil Creative, an AI SEO agency in Atlantic Beach, Florida, in the Jacksonville metro. We ran both studies ourselves with Narsil SEO, our own software, and paid for every search. We sell review management and listings work, so we have an interest in what the studies say. That's why the method, the counts and the data are here to check.
FAQ
- How many businesses did the Jacksonville studies measure?
- 902 different businesses over 1,201 measurements: 600 drawn at random across 10 industries, and a group of 400 drawn by their place on Google Maps, 98 of whom were also in the 600. The lead cost study measured searches, not businesses: 42,149 search terms screened and the top 100 buying searches priced.
- Do the counts include repeated searches?
- Yes. Every search we ran is counted, including runs we threw away and ran again, because each one was work we did. The market figures repeat nothing: 244,593 buying searches a month counts each group of close variants once.
- Can I use the data?
- Yes. The listings study's business-level data is a free download on its method page, with no business named. Cite it as Narsil Creative, Jacksonville Business Listings Study, October 2026. The lead cost spreadsheet, with every search, bid and rule, is free for your email address.