Short answer
We checked 1,358 car, motorbike and boat rental businesses across 14 Florida markets. 55% cannot be booked online at all, and among car rentals it is about two in three.
Florida's boat tours mostly sell online through ticketing tools; its independent car and scooter rentals mostly still take bookings by form, text or phone.
On this page
The data is ours, taken from public Google Maps listings. The method is the one we used for Spain two months earlier, and the whole of it is published, so anyone can repeat it.
What we found
Of 1,358 independent car, motorbike and boat rental businesses across 14 Florida markets:
| State | Businesses | Share |
|---|---|---|
| Cannot be booked online | 750 | 55.2% |
| Bookable on their own site | 559 | 41.2% |
| Bookable on their own site or a linked marketplace | 608 | 44.8% |
| Runs a named third-party booking engine | 384 | 28.3% |
| No website at all in the Maps listing | 160 | 11.8% |
The 95% confidence interval around 55.2% runs from 49.6% to 60.8%. And 55.2% is a floor: the rule for "bookable online" is generous. A booking step that asks for dates and a specific unit is enough to count, so part of the 45% still ends in a confirmation that someone types by hand.
Cars and boats are two different markets
| Fleet | Businesses | Bookable online |
|---|---|---|
| Boats, including tours and fishing charters | 852 | 50.6% |
| Motorcycles and scooters | 145 | 39.7% |
| Cars | 361 | 33.1% |
About two in three Florida car rentals take bookings by form, text, WhatsApp or phone.
The boat figure is high because of its mix. Google Maps files captained tours, sunset cruises and fishing charters under the same "boat rental" searches, and those businesses sell seats and time slots through ticketing tools like FareHarbor and Peek. That one fact explains why 28% of Florida businesses run a named booking engine, against about 5% in Spain and in Campania. It is the tours, not the rental fleets.
These are estimates. The sample was stratified by what a first, cheap check had detected, not by fleet, so each fleet's share applies the verified rates to that fleet's own mix.
By town
| Town | Businesses | Bookable online |
|---|---|---|
| Key West | 90 | 65.5% |
| Panama City | 32 | 52.5% |
| Destin | 95 | 51.5% |
| Fort Lauderdale | 102 | 48.4% |
| Tampa | 68 | 45.7% |
| Panama City Beach | 49 | 45.3% |
| St. Petersburg | 64 | 42.3% |
| Naples | 67 | 41.9% |
| Miami | 142 | 38.5% |
| Orlando | 81 | 38.5% |
| West Palm Beach | 60 | 38.2% |
| Jacksonville | 57 | 37.1% |
| Miami Beach | 108 | 35.0% |
| Fort Myers | 38 | 31.9% |
The towns that sell boat tours lead. Key West, Destin and the Panhandle are ticket-selling places. The towns that rent cars and scooters to visitors (Miami, Miami Beach, Orlando, Fort Myers) trail. Town figures are estimates built the same way as the fleet figures, so do not read much into a gap of a few points.
How it was measured
Find every listing
We searched Google Maps for six kinds of rental (car, rideshare car, motorcycle, scooter, boat and jet ski) in 14 markets, from Key West to Panama City Beach. We kept Florida addresses and removed national chains, franchises and marketplaces.
Check every website once
We downloaded each homepage and looked for the fingerprint of a known booking engine, a date picker or a cart. That check is cheap and wrong in one direction: it cannot see anything that JavaScript loads.
Verify a sample in a real browser
We drew 150 sites, weighted toward the groups where the cheap check was least reliable, and opened each one in a browser. We recorded the booking engines seen in network traffic, followed the booking links, and read the form at the end. 62 sites that no automatic rule could settle were judged by hand against the written rule above.
Carry the rates back to everyone
We applied the verified rate in each group to its whole population with the standard stratified estimator. The confidence intervals come from the same calculation.
The cheap check had missed 48 online booking flows in the sample and invented one.
We also found two faults in our own instrument along the way:
- A disguised browser looked like a dead site. Our browser announced itself as desktop Chrome, and some firewalls drop that connection.
- A review badge looked like a booking channel. A TripAdvisor review badge counted as a marketplace.
Both were fixed, and the Spanish and Italian rounds were re-checked for the first one. Their published figures stand.
The data, for anyone who wants to redo it. The aggregates are published as CSV under CC BY 4.0. No business is named.
What it means if you rent out cars, bikes or boats
Start with a two-minute check. Open your site on a phone, signed out, and see whether a renter can pick one specific car for specific dates and have it held without anyone answering them. If someone has to reply at some point, you are in the 55%.
That is not a failure in itself. A six-car business with regulars and a phone that rings works fine that way. Three things change the arithmetic in Florida:
- The late-night booking. A request answered at nine the next morning competes with a rental that was booked at eleven last night somewhere else.
- Season. From December to April the cost of answering by hand grows with occupancy, exactly when there is least time for it.
- The deposit. Holding the deposit on the renter's card when they book, rather than at the counter, is what stops the no-show.
The rentals that already take bookings online face a different question: what happens after the booking. A shop or calendar plugin captures the reservation and stops there. The agreement, the deposit hold, the pick-up photos, the SunPass tolls and the invoice are still done by hand.
What this study does not say
- It does not measure revenue or fleet size. A Maps listing carries neither.
- It leaves out anyone not on Maps. The population is the rental businesses findable on Google Maps in those 14 markets.
- It is a snapshot. Listings and sites were checked on 13 November 2026.
- Fleet and town figures are estimates, not independent samples.
- And we sell rental software, so the obvious bias in this study is ours. That is why the whole method is published, including the two faults we found in our own tools.
Journalists and trade associations can write to us for the stratum breakdown or the sample design. It is ready to share.