Short answer
We checked 1,759 vehicle and boat rental businesses across Spain's five biggest tourist hubs. 69% cannot be booked online at all — reservations arrive by form, email, WhatsApp or phone.
Of those that can be booked, only one in seven runs recognisable rental software; the rest wired a shop or calendar plugin to a car. 209 have no website at all.
On this page
The measurement is ours, taken from public Google Maps listings, and the whole method is published here because the number this industry quotes — the one we had written down ourselves — was wrong by a factor of four.
What we found
Of 1,759 car, motorbike and boat rental businesses in Palma de Mallorca, Ibiza, Málaga, Alicante and Tenerife:
| State | Businesses | Share |
|---|---|---|
| Cannot be booked online | 1,212 | 68.9% |
| Bookable on their own site | 541 | 30.8% |
| Bookable on a marketplace their listing points to | 6 | 0.3% |
| Runs a recognisable booking engine | 83 | 4.7% |
| No website at all in the Maps listing | 209 | 11.9% |
| Publishes WhatsApp as a contact channel | 455 | 25.9% |
The 95% confidence interval around 68.9% runs from 56.9% to 80.8%. And 68.9% is a floor, not a midpoint: the rule for "bookable online" is deliberately generous — a reservation step that asks for dates and a specific unit is enough — so part of that 31% still ends in a confirmation email written by hand.
Two of those rows belong together. Only 4.7% run a recognisable booking engine, while 30.8% take bookings on their own site. So among the businesses that can be booked online, only one in seven does it with rental software. The other six wired a shop or appointment plugin — WooCommerce, a Wix calendar, a Ninja Forms form — to a car or a boat. It captures the reservation and does nothing that comes after it: the contract, the deposit, the handover inspection, the invoice.
How it was measured, and why that matters here
Our first measurement said 8% took online bookings. It came from downloading each homepage once and looking for the fingerprint of a known booking engine. That has two flaws, and they compound:
One download does not run JavaScript
A widget injected by a script, an engine loaded in an iframe, a flow that lives at
/reservar— none of it is in the initial HTML. Repeated in a real browser across a sample of 119 businesses, the cheap method had missed 34 online booking flows and invented none.The label conflated two different states
The classifier marked "I found a booking engine" and "the site did not respond" identically. Of 140 businesses labelled that way, 48 had an engine and 92 were a blank response. Reading it as "has online booking" was wrong twice over.
Verification ran in three passes over a stratified sample of 119: a real browser capturing network requests (the signal a plain download structurally cannot have), then the vocabulary of the reservation step — proceed to payment, card, total against request a quote, send enquiry, we will get back to you — and finally the form itself: which fields it asks for, what the button says, whether a price appears. The 45 businesses no automatic rule could settle were classified by hand against a written rule, and the rates were carried back to the population with the standard stratified estimator.
The methodological finding is as publishable as the number: any study of this shape built on a homepage crawl overstates the gap by roughly four times. When a software vendor signs it, the bias runs in the vendor's favour.
By island and by fleet
| Region | Businesses | Bookable online | Cars / Boats / Bikes |
|---|---|---|---|
| Mallorca | 479 | 33.4% | 127 / 300 / 52 |
| Málaga | 388 | 29.2% | 222 / 136 / 30 |
| Tenerife | 337 | 32.7% | 152 / 98 / 87 |
| Alicante | 284 | 30.3% | 161 / 123 / 0 |
| Ibiza | 271 | 28.6% | 92 / 157 / 22 |
| Fleet | Businesses | Bookable online |
|---|---|---|
| Boats | 814 | 33.3% |
| Cars | 754 | 29.3% |
| Motorbikes and scooters | 191 | 29.0% |
The regional spread — 28.6% to 33.4% — fits entirely inside the margin of error of the national figure, so no island is behind the others. The finding is that the number is high in all five destinations at once. Both tables are synthetic estimates: the sample was stratified by what the first method had detected, not by region, so the regional shares apply the verified rates to each area's own mix.
What it means if you rent out cars, bikes or boats
Start with a two-minute check: open your site the way a customer would on a phone and see whether they can pick one specific vehicle for specific dates and have it held without anyone answering them. If you have to answer at any point, you are in the 69%.
Being there is not a management failure by itself. An eight-car business in Jávea with repeat customers and a phone that rings works fine that way. Three specific things change the arithmetic:
- The out-of-hours booking. A form answered at nine the next morning competes with the one that was booked at eleven last night somewhere else.
- August. The cost of answering by hand grows with occupancy, exactly when there is least time to do it.
- The deposit. Taking or holding the security amount at the moment of booking rather than at the counter is what removes the no-show.
And there is one reason with a date on it that does not depend on any of the above: since 29 July 2025, invoicing software sold new in Spain has had to arrive already adapted to Veri*Factu, even though your own deadline is 1 January 2027 for corporate income tax filers or 1 July 2027 if you are self-employed. If you are thinking of changing tools this winter, it is the same decision — what Veri*Factu requires of a rent-a-car.
What this study does not say
- It does not measure revenue or fleet size. A Maps listing carries neither.
- It excludes anyone not on Maps. The population is "rental businesses discoverable on Google Maps" in those five areas.
- It is a dated snapshot. Listings from 9 September 2026, sites verified on 22 September 2026. A business that added online booking in between reads as manual.
- Regional and fleet figures are synthetic 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, and why the first thing it reports is that our own earlier figure was wrong in the direction that flattered us.
If you are a journalist or work for a trade association and want the per-area detail, the stratum breakdown or the verified sample, write to us: it is ready to share.