Isla Tech field guide · Puerto Rico

How to Start a Scooter Rental Business in Puerto Rico (2026)

Written from inside a working fleet on Culebra — four standard e-bikes and eight Cake Makka Flex e-mopeds. Every price, multiplier and threshold below is the live configuration of a shop that takes real bookings, not a software vendor's estimate.

Key takeaways

  • The number that decides whether this works is revenue per available unit-day — not price, and not "rentals per day."
  • Most published price ladders contain at least one inverted tier, where the customer pays more for less time. We found two in our own and cut a $50 tier to $38 on 2026-07-09.
  • Insurance bills in two incompatible shapes — flat per vehicle or per day. Get it wrong and you leak on every order, in one direction or the other.
  • Your bottleneck is turnaround, not fleet size. On a four-unit fleet, a vehicle sitting unmarked in the rack is a sold-out day.
  • We are not publishing a startup-cost table. We do not hold receipts we would stand behind, and every guide that hands you a low/high column is guessing on your behalf.

Every "how to start a rental business" guide you have read was written by a rental-software company about a business it does not operate. This one is written from inside the fleet. The prices below are the ones our checkout actually charges, and where we do not have a verified number — startup cost, maintenance cost, permit fees — we say so instead of filling the gap with a plausible range.

01Start with revenue per available unit-day

Competitor guides open with a profitability table. We open with an equation, because the table is where the lying happens. The number we run on is revenue per available unit-day: revenue divided by (units in service × days in the period) — including the days a unit earned nothing.

Our standard fleet is four e-bikes; our premium fleet is eight Cake Makka Flex e-mopeds. Four units is 28 available unit-days a week, and at our 24-hour price of $60 that is the theoretical weekly ceiling — which nobody touches.

What moves the number is never $60 → $70. It is the days a unit sat idle because it was not marked returned fast enough, because a discount was still live on an almost-sold-out date, or because the ladder priced 12 hours above 24 and the customer walked.

02The startup-cost table we refuse to write

Here is where every competing guide puts a low column, a high column and a bolded total. We will not, and that is the point of this page: we do not hold purchase receipts we would publish. Our own fleet paperwork carries an invoice number and no dollar figure. Price these yourself instead, from written quotes:

  • Vehicles — out-the-door, including freight to Puerto Rico, which is not the mainland number.
  • Spares — tubes, tires, brake pads, chargers, a spare battery per model. On an island your spares shelf *is* the supply chain.
  • Storage and charging — secured space, and the power to charge the whole fleet overnight.
  • Titling and paper provenance — MCOs, registrations, a physical file. Cheap, until it is the only thing you have.
  • Insurance — written quotes, and read §7 first: two quotes can be structured so differently the premiums are not comparable.
  • Card processing, booking software (can it express your ladder without you typing prices twice?), and location — see §13.

The capital risk nobody prices for you

Cake, the manufacturer of our eight premium e-mopeds, went out of business. Ours are 2022 Cake Makka Flex — 3.9 HP, 400 lb GVWR, 48V/3000W in two motor variants, all on one invoice, MCO'd 2023-11-08 — and every MCO image in our fleet register is still marked *pending*. With the manufacturer gone, the paper file is the fleet's only provenance and every parts decision is an owner decision. Price that risk before you price a boutique-brand vehicle.

03The price ladder — build it, then prune it

Our live standard-fleet ladder, exactly as our checkout charges it.

Todo Culebra standard e-bike ladder, live 2026-08-31
DurationPriceNote
4 hr$22Same-day / walk-up only
6 hr$30Top seller
8 hr$38Was $50 until 2026-07-09
12 hr$48Was $58 until 2026-07-09
24 hr$60The backbone
48 hr$99
3–4 days$45 / bike / day
5+ days$40 / bike / day

The reprice: two tiers that were quietly dead

On 2026-07-09 we rebuilt the middle of the ladder because it was incoherent. The diagnosis, written into our pricing code at the time: "+$2 bought 12 extra hours at 12hr→24hr, +$20 bought 2 hours at 6hr→8hr."

The mix confirmed it: the 12-hour tier was 3.6% of all-time revenue, while 24-hour plus multi-bike couples' bookings were 34%. We cut 8hr $50→$38 and 12hr $58→$48 and left the working tiers alone. Both figures are pinned to that date, not today's mix.

Odd hours are generated, not typed

Only four rungs are owner-priced: 4hr $22, 6hr $30, 8hr $38, 12hr $48. Every whole hour between them is linear interpolation — live: 5hr $26, 7hr $34, 9hr $41, 10hr $43, 11hr $46. Reprice an anchor and they move with it, so they cannot drift out of order.

We do not offer 13–23 hours at all: interpolated, every one would land above the 24-hour price of $60, so we pruned the ladder instead of filling it. A hole beats an inversion.

04The multi-bike floor and the anti-overpay rule

Two or more standard bikes on any day-based price at a flat $35 per bike per 24 hours, prorated: two bikes for 24 hours is $70 (the Couple's Deal), two for 48 hours $140, three for 24 hours $105. The floor is exempt from day-of-week pricing, group discounts, promos and surge.

The rule that comes with it is the one to steal: whenever two or more bikes quote on an *hourly* tier and the total exceeds $35 × bikes, the system must state the 24-hour flat price in the same reply and say the full day costs less. Worked example: a Saturday, 12-hour, two-bike rental prices at $104 dynamically against $70 flat.

05Dynamic pricing that does not lie to anyone

Three layers sit on top of the ladder. All three are live; all three have hard exemptions.

Day-of-week multipliers — a ±15% spread
SunMonTueWedThuFriSat
1.100.850.850.900.901.001.15

Group discounts by bike count: 2 bikes 5%, 3–4 bikes 10%, 5–6 bikes 15%. The premium fleet carries a flat +50% multiplier, applied *before* day-of-week and discounts so every tier scales proportionally — 6 hours $45, one day $90.

Occupancy surge, and where it must never fire

Surge keys off free units for the start date: 2 or fewer free → ×1.15, 4 or fewer → ×1.10. It is exempt from the multi-bike floor and Couple's Deal (a live ad promises "$70 every single day — no surge"), from the 4-hour tier, and from every extension and reprice path. And it fails open — any error returns 1.0 with a loud warning.

06Deposits, money, and the cash question

Card-only since 2026-07-17. A $10 deposit online reserves the vehicle; the balance auto-charges about two days before the ride via a 10 AM daily job, or the customer pays any time from a link in the confirmation email. No cash. Free cancellation 2+ hours before start refunds the $10; inside two hours, or a no-show, forfeits it.

No cash is also our answer to cash-versus-record drift: there is no drawer to reconcile, because there is no drawer.

Our damage deposit figure is $75 — but the automatic hold has been switched off since 2026-06-03. Damage charging is a manual, owner-only decision at return time. Photos get taken either way, every time.

Add-ons, and the two that are free on purpose

Snorkel set $10 · beach chair $10 · cooler backpack $10 · bike trailer $15 · beach blanket $5 · Bluetooth speaker $5 · rain poncho $5 · cup holder $5 · phone mount $3. Helmet $0. Lock $0 — safety gear is deliberately not an upsell. Delivery anywhere on the island is a flat $25 per order, added once regardless of vehicle count.

07Insurance: per vehicle or per day?

This mechanic is worth more than any startup-cost table. Protection products bill in two structurally different shapes, and they are not interchangeable.

From another Puerto Rico operator's published storefront and order lines, de-identified: their three-wheeled autocycle carries flat per-vehicle protection at $40, regardless of rental length — a four-day rental billed a single $40 charge. The Jeep on the same storefront carries per-day protection at $15/day: a thirteen-day rental billed 13 × $15 = $195.

Copy the wrong shape and you either insure thirteen days of exposure for one day's money, or price a one-hour rental out of its own tier. The shape is not a detail; it is the product.

Deposits scale with risk, not with price

On that same published ladder, security deposits track vehicle risk rather than rental price: $100 on scooters and e-bikes, $250 on mopeds and the autocycle. Extra-day rates follow — $50 for an e-bike or e-scooter, $85 and $95 for the two moped classes, $349 for the autocycle. Four-wheel inventory carries a separately-billed under-25 surcharge of $15/day.

08The ladder-inversion check you can run in ten minutes

Sort your published prices by hours and compute dollars-per-hour down the column. Any row where a longer rental costs *more* than a shorter one is an inversion — a tier that either never sells or insults whoever books it.

Not theoretical. On the same de-identified Puerto Rico ladder, the 10-hour "9am–7pm" tier sits below the 24-hour tier on every vehicle listed — except the e-bike, where 9am–7pm is $110 and 24 hours is $95. The customer pays $15 more for 14 fewer hours.

It is the same defect class we found in our own middle tiers in July 2026. Nobody catches it by intuition; you catch it with a sorted column.

09Utilization math, and what a red day does by itself

We compute pressure from free units for a date: RED at 2 or fewer free, YELLOW at 3–4, GREEN at 5 or more. On a four-unit fleet, RED triggers at 50% booked. A failed read degrades to GREEN on purpose, so pressure logic can never block a sale.

Nobody flips a switch when a day goes RED — the day flips itself:

  • Discount promos turn themselves off for new bookings on that date.
  • Premium vehicles surface as a +50% overflow option, hidden on normal days.
  • Sold-out visitors are offered a standby list.
  • Marking a vehicle Returned auto-emails standby that one freed up. The sweep runs every 5 minutes.

Which gives the lesson that generalizes past our island: because the standby email fires off return-marking, fast return-marking is the most valuable thing a staff member does on a busy day. Latency between a physical return and system-visible availability is lost revenue.

10Seasonality, honestly labeled

Directionally — labelled a general market observation rather than a published figure, because the numbers that corroborate it are not ours to publish — Caribbean and Puerto Rico rental demand peaks roughly December through April, December strongest, and troughs in September–October during hurricane season.

Plan the trough as a cost, not a surprise: carrying cost, insurance, storage and financing keep running against near-zero revenue. It is also the only sane window for a reprice — a ladder change during peak is one whose results you cannot read.

12The failure modes that actually cost money

SMS is a trap

Outbound texting is disabled shop-wide and inbound texts reach nobody; the business line is call-only, and customer contact is email plus in-app chat. Never build a process that depends on a text message arriving. We learned it expensively — pizza-order notifications were built as SMS, so the procedure now is to poll an admin panel every 20–30 minutes and phone the kitchen.

Missed calls

A missed call at a rental counter is not a missed call; it is a booking that went to whoever answered. It is the most expensive item on this list, and why our phone is answered by an AI 24/7 instead of a voicemail box.

Never quote from memory

Our walk-up kiosk calls the same price calculator as checkout, so the kiosk price is the charged price to the cent. That parity is the only reason the rule works: never quote a price from memory. Checkout sessions also expire roughly 30 minutes after creation — the remedy is re-quoting, which takes about ten seconds.

A gap in our own counter flow

Our walk-up kiosk quote flow does not include the digital waiver step the online booking flow has; the interim fix is routing walk-ups through the regular flow on their own phone. We publish our own defect because it is the one you are most likely to reproduce: check whether your counter path inherited your waiver step.

Overtime you should not charge

30-minute grace, then $25 per vehicle — but since 2026-05-16 the system only flags it and builds a one-tap charge link; nothing charges automatically. When *not* to charge: an apologetic guest just past grace; one who already bought a paid extension; a chat self-return whose flag had not cleared; weather or a failure that was ours.

Oversells and maintenance windows

What breaks: flats, dead batteries, brakes. Pull the vehicle from the rack physically, then set a maintenance window — dates plus a reason like "flat tire" — and the site sells one fewer unit for those dates without disturbing existing bookings. If your software cannot do that without cancelling someone, you will oversell on your busiest day.

13Should you even do this?

An honest close, since nobody else writes one. Do it if you are willing to be operationally present: a four-unit fleet hits RED at 50% booked, so the difference between a good week and a flat one is return-marking latency and answered calls. Our shop opens 7 AM to 8 PM daily, a flat twelve-minute walk from the ferry dock and deliberately not downtown — worth more than a price increase, and not reversible in season.

Do not do it if you need the ceiling number to work; you will never see it. Do not do it if half your fleet is a boutique brand whose parts supply you have not checked. Do not do it if you are counting on cash walk-ups for cash flow; we removed cash entirely and would again, but every dollar arrives on a processor's schedule.

And do not do it if you are unwilling to publish your own defects to yourself. The shops that fail are not the ones with the wrong prices — they are the ones that never found out.

Frequently asked questions

How much does it cost to start a scooter rental business in Puerto Rico?

We will not give you a number, and you should be suspicious of guides that do. We do not hold purchase receipts we would publish for our own fleet, and we have not verified Puerto Rico permit fees, insurance premiums or lease terms to a standard anyone should spend money on. What we can give you is the checklist to price yourself — vehicles landed in PR, spares and a spare battery per model, storage and charging capacity, titling and paper provenance, written insurance quotes, card processing, booking software, and location — plus the capital risk most people miss: if your manufacturer goes out of business, as ours did, the paper file becomes the only provenance your fleet has.

What should I charge for scooter and e-bike rentals?

Build the ladder from a few owner-priced anchors and generate everything between them rather than typing every rung. Ours, live as of 2026-08-31: 4hr $22 (same-day only), 6hr $30, 8hr $38, 12hr $48, 24hr $60, 48hr $99, then $45/day for 3–4 days and $40/day at 5+. The whole hours in between are linear interpolation between the anchors, and 13–23 hours are not offered at all because they would price above the 24-hour rate. Then sort your ladder by hours and check for inversions — we found two in July 2026 and cut a $50 tier to $38.

Should rental insurance be charged per vehicle or per day?

Both shapes exist in the real market and they are not interchangeable. On one Puerto Rico operator's published storefront, the autocycle carries flat per-vehicle protection at $40 regardless of rental length — a four-day rental billed a single $40 charge — while the Jeep carries per-day protection at $15/day, so a thirteen-day rental billed $195. Copy the wrong shape and you either insure a two-week rental for one day's money, or price yourself out of your own hourly tier. Ask every insurer which shape their product is before you compare premiums.

How many scooters or e-bikes do I need to start?

We run four standard e-bikes and eight premium e-mopeds. On a four-unit fleet our pressure engine goes RED — discounts self-disable, premium overflow appears, and sold-out visitors get a standby list — at two or fewer units free, which is 50% booked. Small fleets hit constraint far earlier than people expect, which is why turnaround speed matters more than unit count: marking a vehicle Returned is what triggers the standby email, so return-marking latency is directly lost revenue.

Is dynamic pricing worth it for a small rental fleet?

Ours is, with guardrails. Day-of-week multipliers run a ±15% spread (Saturday 1.15, Sunday 1.10, Friday 1.00, Wednesday and Thursday 0.90, Monday and Tuesday 0.85), and occupancy surge adds ×1.10 at four or fewer units free and ×1.15 at two or fewer. The guardrails are what make it defensible: surge never touches the advertised multi-bike flat deal, the 4-hour tier, or any extension; it fails open at 1.0 on any error so it can never crash a checkout; and the cap is disclosed to customers as roughly 15% in copy that must move in lockstep with the config.

What breaks first, and what does downtime actually cost?

Flats, dead batteries, and brakes. We will not publish a maintenance cost per unit because we do not have verified parts-and-labor figures we would stand behind. What we can tell you is the process that stops a breakage from becoming an oversell: pull the vehicle physically, then set a maintenance window with from/to dates and a reason, which makes the site sell one fewer unit for those dates without disturbing an existing booking. If your software cannot remove a unit from inventory without cancelling a customer, that is the feature to fix before opening day.

The part where everyone else says "book a demo"

Ready to run a fleet that answers its own phone?

You now have the ladder, the utilization math, the insurance shapes and the failure modes — the same ones our own shop runs on. None of that is the hard part. The hard part is the Tuesday when three people call while you are handing over a bike, and two of them book somewhere else.

We built the software because we needed it for our own fleet first: it quotes from one price calculator so the counter, the website and the AI can never disagree, it flips a red day's promos off by itself, and it answers the phone at 2 AM in Spanish. It is a productized service, hand-onboarded — not a signup form.

Every guide you just compared this one to ends with a demo request. This one ends with a phone number that a machine picks up. Call it. Ask it something hard. It will not transfer you, and it will not invent a price.

Or leave it in writing

Tell us what you are trying to run and we will reply from a real fleet, not a sales sequence.