Mobility platforms, built for the two numbers that fight each other.

Market, licensing and the vehicle model

Mobility is regulated locally rather than nationally, and the rules differ between cities in the same country. Operating licences, driver permits, vehicle categories, insurance minimums and whether your model is even permitted are the first constraints, and they shape the product rather than sitting beside it.

The vehicle model follows from that. A platform serving licensed taxis, private hire vehicles and peer-to-peer car sharing needs three different compliance profiles, three insurance positions and three answers to who is responsible when something goes wrong.

The output is a market entry position: which cities, under which permission, with what your platform must prove about every driver and vehicle before dispatch.

Matching and dispatch

Nearest available driver is the intuitive rule and it is close to the worst one. It strands drivers who happen to be near a low-demand area, ignores where the trip will end, and produces a distribution of wait times that is fine on average and terrible at the tail.

Better matching considers where the driver will be afterwards, whether a slightly longer pickup produces a materially better trip, and whether waiting three seconds for a closer driver to become free beats assigning now. Batched matching over a short window outperforms greedy assignment, and the window length is a tuning decision per city.

Location itself is a hard problem underneath this. Positions arrive irregularly, tunnels and dense buildings break them, and a matching engine that trusts every reported coordinate will dispatch to a driver who is actually a street away.

Pricing, surge and driver earnings

Pricing in mobility is a supply instrument, not a revenue setting. Raising prices when demand exceeds supply does two things at once: it reduces requests and it attracts drivers into the area. Both are necessary, and a platform that cannot do it will simply fail to serve peak demand.

It is also the most scrutinized mechanism you will operate. Surge during emergencies, price differences between neighbourhoods and opaque driver earnings all attract regulatory and press attention, so the rules need to be explicable and the caps deliberate.

Driver earnings deserve equal engineering weight. A driver who cannot reconstruct why a trip paid what it did will assume they were shortchanged, and that belief spreads through driver communities faster than any communication you send.

Trip lifecycle, payments and disputes

A trip is a state machine that runs in the physical world, which means it disagrees with reality regularly. Riders cancel after a driver has travelled. Drivers end trips early or late. Phones lose signal in the middle. The fare has to be computable and defensible in all of those cases.

Payment adds its own timeline. Cards are pre-authorized then captured on completion, cash remains significant in many markets and creates a settlement flow in the opposite direction, and wallets need topping up and refunding. Tolls, waiting time, cleaning fees and tips each attach differently.

Disputes are frequent and low value individually, which means they have to be resolvable from trip data rather than by investigation. Route, timing and fare breakdown retained per trip is what makes that possible.

Safety, verification and insurance

Mobility carries physical risk to real people, which puts safety above every other consideration in this sector. Driver identity, licence validity, vehicle roadworthiness and insurance are all checked before dispatch and re-checked on a schedule, because every one of them expires.

In-trip safety is a product surface of its own: share-trip, an emergency path that works with one hand, and the ability to identify a vehicle before entering it. These are not features to add after an incident.

Account sharing is the failure mode that matters most. A verified driver lending their account to an unverified person defeats every check you performed, so periodic re-verification against a live capture is the control that actually holds.

Supply balance and city economics

A mobility marketplace works or fails on liquidity in a defined area, and liquidity is local. Coverage across a whole city with thin density everywhere is worse than density in three neighbourhoods, because wait times in a thin area are long enough to lose riders permanently.

This is where incentives get designed rather than improvised. Guarantees, peak bonuses and repositioning nudges all cost money and all change driver behaviour, and their effect has to be measurable per zone and per hour rather than assessed by feel.

Launch and day two

One city, and within it a small number of zones with real driver density before any rider acquisition spend. The most common way to waste a launch budget in this sector is to advertise across a city your supply cannot serve.

Day two is safety incidents, fare disputes and driver earnings queries, all of which arrive immediately and none of which can wait for a release cycle. Sixty days of dedicated support, six months of priority bug resolution, twelve months of updates.