📍 Clearwater, FL · Roadside Assistance

Case Study: Pinellas Roadside Rescue

A 90-page roadside assistance website, and the traffic it actually produced between 13 July and 30 October 2026. The business takes around 40 calls in a typical month. Below is what we built, screenshots of the live site, the numbers we measured, the method we used to measure them, and an honest account of how much of that call volume we can and cannot pin on the website.

Disclosure: we own this business.

Pinellas Roadside Rescue is owned and operated by Gary Granzow, the founder of Double G Consulting. It is not a third-party client. We are publishing it as a case study because it is the one project where we control both the website and the books, so we can show the build and the measurement honestly instead of quoting a number a client gave us over the phone.

The Build

What we built

90 pages, structured so that a search for a specific problem in a specific city lands on a page about that problem in that city.

Six services × twelve cities

Flat tire repair, battery jump start, mobile battery replacement, vehicle lockout, emergency fuel delivery and general roadside assistance — each written for Clearwater, Dunedin, East Lake, Gulfport, Largo, Oldsmar, Palm Harbor, Pinellas Park, Safety Harbor, Seminole, St. Petersburg and Tarpon Springs. That is 72 service-area pages plus six service hubs and the supporting pages.

Call and text first, forms second

Roadside work is an emergency purchase. Every page leads with a tap-to-call and a tap-to-text button rather than a contact form, because somebody with a flat tire is not going to fill in a form and wait for a reply.

Prices on the page

Flat rates are published rather than hidden behind “call for a quote.” This loses the price shoppers early, which is the point — the calls that come through have already accepted the rate.

The site mirrors the Google Business Profile

The services and categories on the Google Business Profile match the services on the site, one to one. The profile is where most roadside customers actually find the business; the site is what they check before dialling.

The Numbers

What we measured, 13 July – 30 October 2026

The call figure is the owner’s count. Everything else is measured from the site’s own logs — real Florida visitors on consumer internet connections, after bots and datacenter traffic are removed. The method is written out below so you can judge it.

~40

calls a month in a typical month, all channels

pages built and live

116

distinct Florida visitors in 15 weeks

of them arrived from ChatGPT

5.0

Google rating, 43 reviews (Oct 2026)

The call figure is the owner’s own count across every channel — the Google Business Profile, the website, repeat customers and word of mouth. It is not a website conversion number, and we explain below why we are not presenting it as one.

Distinct Florida visitors per month

Month	Distinct visitors	Page views
July 2026 (from the 13th)	10	33
August 2026	34	57
September 2026	25	45
October 2026 (to the 30th)	52	88

Where those visitors came from

Source	Distinct visitors
Direct — typed the address, a saved link, or a Google Maps listing that strips the referrer	54
Google Search	36
ChatGPT	30
Bing	8
DuckDuckGo	7
Yahoo	4
Google Ads	4

The ChatGPT figure is the one worth dwelling on. Roughly a quarter of the people who reached this site in those fifteen weeks were sent by an AI assistant, not by a search engine. Those visits arrive tagged utm_source=chatgpt.com, so they are counted directly rather than inferred. Nineteen of them landed on the homepage and five landed on a city page, which means ChatGPT is citing specific service-area pages and not just the domain.

The service-by-city pages earn their keep. After the homepage, the most-visited pages were mobile battery replacement in St. Petersburg, roadside assistance in St. Petersburg, flat tire repair in St. Petersburg and mobile battery replacement in Largo — all pages that only exist because the site was built city by city.

How these numbers were produced

Every page on the site carries a first-party logger we wrote. It records the visit server-side, so it does not depend on the visitor accepting cookies or running an ad blocker.
Raw traffic for the period was 1,282 visits. We removed everything that was not a real person in Florida on a consumer connection, which left 223 visits from 116 distinct addresses. About 82% of the raw number was discarded.
What was discarded: known bot fingerprints; everything from Google, Microsoft, Amazon, Apple, Ahrefs, Fastly, OVH and similar datacenter and crawler networks; and everything outside Florida. The single largest block was 600 hits from Google-owned infrastructure in California and Oklahoma, which is AI and crawler traffic rather than customers.
Distinct visitors are counted by IP address. A household on one connection counts once; a person who switches from phone data to home wifi counts twice. It is an approximation, and it is the same approximation in every row.
The window starts on 13 July 2026 because that is the day the logger was installed, not the day the site launched.

The Honest Part

What this case study does not claim

Most web design case studies end with a revenue figure and an arrow pointing up. Here is what we are deliberately not telling you, and why.

We are not claiming the website produced the 40 calls. The business takes roughly 40 calls in a typical month, and that is the owner’s own count rather than something the website measured. Roadside customers dial from the Google Maps listing at least as often as from the site, and a ringing phone does not say which one sent it. There is no call tracking on this business, so splitting those 40 calls between the profile, the website and word of mouth would be a guess dressed up as a measurement. What we can say is that the website is one of the inputs feeding that number, and that it reached 116 Florida people in fifteen weeks while doing so.
We are not claiming search rankings. This site does not rank on page one for its main commercial terms. Its visibility comes from the Google Business Profile, from direct and saved traffic, and increasingly from AI assistants. Anyone selling you a roadside website on the promise of ranking first for “flat tire repair” is selling you something we have not managed to do ourselves.
We are not claiming the website produced the reviews. The 5.0 rating across 43 reviews is earned by turning up at 2am and fixing the car. The website's job is to make sure the person who is about to call can see that rating, the price and the response time before they dial.
We are not claiming this is a big number. 116 people in fifteen weeks is a small local business, measured honestly. We would rather show you a real small number with the method attached than a large one you have no way to check.

Everything above is checkable. The site is at pinellasroadsiderescue.com and the Google Business Profile, with its review count, is public.

What Carries Over

What we would do the same for your trade

Build the service-by-city grid, not one “service areas” page. A page that lists twelve city names ranks for nothing. Twelve pages that each describe the job in that city is what produced the traffic above.
Put the price on the page. It costs you the bargain hunters and keeps the buyers.
Match the site to the Google Business Profile exactly. Same services, same categories, same wording. For a local trade the profile is the front door and the site is the room behind it.
Write pages an AI assistant can quote. Plain statements of what you do, where, and for how much. A quarter of this site's visitors came from ChatGPT, and it is quoting the specific city pages.
Measure it properly, then tell you the real number. We will set the same logger up on your site and show you the filtered figure, not the raw one.

See the same approach applied to roofing, HVAC, plumbing and contracting generally, or browse the rest of our work.