Building A Commute Programme That Scales

• BTT Team
• 14 July 2026
• 3 min read

Almost every commute programme works at fifty seats. Very few survive contact with five hundred without being rebuilt. The difference is decided early, by choices that look harmless at small scale.

A programme that scales is not a bigger version of a small one. It is a different design. Below are the four decisions that determine which one you end up with.

1. Route To Nodes, Not To Doors

Door-to-door pick-up is the easiest promise to make and the hardest to keep. Every additional address adds a stop, a detour and a failure point, and the cost grows faster than the headcount does. Route length climbs, the first person collected sits on the vehicle far longer than the last, and punctuality becomes hostage to the least accessible address on the list.

Node-based routing — a fixed set of well-chosen pick-up points, each serving a cluster of homes — scales in a way door-to-door does not. Stops stay constant as headcount grows. New joiners map onto existing nodes instead of adding new legs. And a node can be moved or retired without renegotiating with individuals.

• Place nodes where people can already reach them safely on foot in a few minutes.
• Prefer landmarks that do not move — a junction, a station, a school gate — over private addresses.
• Keep the total number of nodes per route small enough that the ride time for the first passenger stays defensible.

2. Fix The Roster Before The Routes

Routing is downstream of rostering. If shift timings are ragged — starts scattered across forty minutes, teams changing pattern weekly — then no routing plan will hold, because the demand it was built for changes faster than the plan can be revised.

Consolidating start and end times into a small number of fixed slots is the single highest-leverage change most organisations can make, and it is an internal decision that costs nothing. Three clean shift boundaries will out-perform a sophisticated routing tool applied to thirty ragged ones.

You cannot route your way out of a rostering problem. Fix the timetable first, then the vehicles.

3. Right-Size The Vehicle To The Corridor

A common and expensive habit is standardising on one vehicle type. Dense corridors carrying forty people at the same hour want a coach or a bus. Thin corridors carrying four want a car. Running the same vehicle on both means either paying for empty seats or running more trips than you need.

Mixed fleets are harder to administer and cheaper to operate. The test is occupancy: a route that consistently runs below roughly two-thirds full is telling you the vehicle is wrong, the node placement is wrong, or the corridor should be merged with a neighbour.

4. Instrument It From Day One

The reporting you will need at five hundred seats is impossible to retrofit, because the data was never captured. Decide at the start what each trip must record, and make it non-optional:

• Planned versus actual time at every node, not just at the site gate.
• Seats booked versus seats occupied, per trip.
• Vehicle and driver identity, on every leg.
• Exceptions, with a reason code — a late trip with no cause recorded is a late trip you cannot fix.

Occupancy is the number that governs cost, and node-level punctuality is the number that governs trust. Programmes that track only site-gate arrival look healthy right up until employees stop using them.

Scaling Is Mostly Subtraction

The instinct when a programme grows is to add: more routes, more vehicles, more exceptions for special cases. The programmes that scale well tend to go the other way — fewer shift slots, fewer stops, fewer vehicle types, fewer bespoke arrangements. Every exception you decline at fifty seats is one you do not have to operate at five hundred.