Carsharing Discovery: Location Search, Filters, and Post-Trip Ratings

Empty maps kill conversion faster than high prices. If a passenger cannot narrow “cars near me for Saturday afternoon” into a shortlist with photos, config, price, and social proof, they bounce — even when your yard is full. After the trip, if nobody rates anything, next week’s shortlist is still a coin flip.

Ptaxi’s carsharing line — carsharing system as car-plus-driver rental, not a keyless P2P protocol list — documents passenger find and filter flows, vehicle cards with images, price, configuration, and ratings, and post-trip order / vehicle / driver evaluation. Fee stacks, unpaid half-hour cancel, and return inspection are covered in earlier posts. This one is discovery and the rating loop.

Find by place and time first

Documented passenger search starts with use location (vehicles within N kilometers of the address) and pickup/return time. That is the inventory question that matters: what can I actually have for this window near this pin?

Implementation and ops notes:

• N kilometers is a policy dial. Too small and downtown looks empty; too large and you show deadhead fiction.
• Time windows must respect how drivers pick up from owners and position to the passenger — car-plus-driver supply is not an idle keyshare sitting on a street corner.
• Show honest emptiness. A blank result with a clear “nothing in range for that time” beats a junk list of cars that cannot serve the window.

Filters that match how people shop cars

After the spatial-temporal cut, passengers can filter by site, vehicle type, brand, seats, pickup address, pickup/return time, rental count, distance, and price. That list is a merchandising toolkit. Use it:

• Fleets with mixed vans and sedans need type and seats early.
• Brand-sensitive corporate renters filter brand before price.
• Price and distance sorts resolve the last-mile hesitation when two cars both “work.”

Do not ship twelve filters with three cars in the city. Filters without density feel broken. Launch with the filters your actual inventory can populate, then expand.

Vehicle cards: proof before pay

Selecting a vehicle shows images, price, configuration, and ratings, then the renter picks time and confirms price. Ratings on the card are why post-trip evaluation is not a vanity feature — they are inventory merchandising.

Card hygiene for ops:

• Photos that match the real unit beat stock glamour shots that trigger “not as described” complaints.
• Configuration fields should match what drivers and owners can actually deliver (seats, type).
• Price on the card must reconcile with the confirm step’s base / service / insurance / driver fee stack — surprise fee stacking after a cheap card is a trust bug.

Unpaid orders still auto-cancel after about half an hour; discovery that converts into unpaid holds should clear so the next renter can see the car. That cancel timer is inventory hygiene for the same map you just searched.

Post-trip ratings: order, vehicle, driver

After the order ends, passengers evaluate the order, the vehicle, and the driver. Three targets matter because car-plus-driver failure modes differ:

• Bad order orchestration (late positioning, messy handoffs) → order rating.
• Dirty or misdescribed unit → vehicle rating.
• Driver service quality → driver rating.

If you only collect a single star, you cannot tell whether to fix the yard, the roster, or the workflow. Feed ratings back onto vehicle cards and into driver ops reviews. Admin already has driver and passenger management tools; ratings give those tools signal.

Cancel policy still sits beside ratings: free cancel inside a window, paid cancel after rules. A passenger who felt trapped by cancel fees will punish you in stars even when the car was fine. Align cancel copy with the rating ask.

How discovery connects to the rest of the car-plus-driver machine

Search and ratings do not replace:

• Passenger qualification and violation prepayment gates before someone can rent.
• Driver qualification review before the order hall is useful.
• Realtime vehicle location for ops watching supply (admin map), which is not the same UI as passenger N-km search but answers the same “where is the fleet?” question for staff.
• Return inspection and owner confirm at the end of the trip.

Think of discovery as the front door. Qualification is the lock. The driver hall and trip lifecycle are the hallway. Ratings are the note on the door for the next guest.

Competitor-shaped outline, our facts

Many fleet-software blogs pitch “search and book in minutes.” Keep the outline if you want — map, filters, book — but fill it only with confirmed pieces: N-km find, the filter dimensions above, card fields, fee confirm, half-hour unpaid cancel, and three-way ratings. Skip invented map vendors, tips, and P2P hardware claims.

What this is not

• Not store-rental smart-lock self-drive merchandising.
• Not a guarantee of ML ranking — docs describe find, filter, and ratings, not a proprietary ranker brand.
• Not WeChat-only login packaging for overseas pages.
• Not invoice/payout deep dives (separate finance article).

Ops dashboard vs passenger map

Admin realtime vehicle location tells staff where supply sits right now. Passenger N-km search tells renters what they can book for a window. Do not demo one as the other. A live ops map full of pins can still yield an empty passenger result if times, qualification, or driver availability do not line up. Train support to check search parameters before telling a customer “the map shows cars.”

Likewise, ratings without moderation workflows become a vandalism surface. Decide who reviews extreme one-star streaks on a vehicle or driver, and whether that unit stays filter-visible while you investigate. Discovery quality is an ops queue, not only a UX widget.

Soft next step

Pick one city pin and one Saturday window, count how many units truly appear inside your N-km setting, and check whether vehicle cards show ratings you actually collect after trips. When search density and the rating loop look demoable, continue on the carsharing system page or book a discovery-and-feedback walkthrough with ops and a driver scheduler in the same call.