How to size and split driver groups
Hall B was quiet. Hall A was flooding. The forklifts in B could not see Hall A jobs because someone had drawn territories on the map. One driver called in sick in A. The line waited. B stayed idle.
What a driver group is supposed to optimize
In twinzo AOS (commonly known as FGS), a driver group is the set of operators who may see and claim a class of work orders. The offer is a broadcast inside that set. First accept wins. Live location does not currently remove farther members from the list. Group design is therefore the main scope control plants have today.
The ordering system's real job is a balanced fleet with almost no production idle from late delivery. That means capacity can move to where demand is hot. Territorial splits by wing, dock vs line, or "expected" designation fight that goal. They keep quiet trucks from helping busy areas. Physical limits are the opposite case. A truck that cannot lift the load, cannot enter the aisle, or is a tugger train instead of a forklift must stay out of that pool. Capability is a hard filter. Geography is usually a soft habit.
Split on what the machine can do, not where it "belongs"
Best results come from groups based on physical limitations. Heavy or high loads that only certain trucks can handle. Mast height. Attachment type. Aisle width. Vehicle class that cannot run the same hop, such as milk-run tugger delivery versus direct forklift pallet bring. Those splits protect safety and equipment. They also match how milk-run orders vs direct forklift orders already differ in stage and load unit.
Area walls and paper designations do the opposite. When you limit where a capable forklift may operate, you stop quieter zones from covering peaks elsewhere. A longer empty or loaded route is often cheaper than a line stop. Peak coverage without disruption beats short territorial hops that leave half the fleet locked in a quiet wing. How templates bind capability classes is under request triggers in the workflow editor. Offer scope stays group broadcast, not nearest-truck routing, as under group broadcast vs nearest-truck myths.
Shared fleet is the resilience model
The end state is a shared fleet that can claim any job its machine can physically do. Saturation then balances across the hall: busy stations draw free capacity, quiet stations stop sitting on protected idle. When one driver is out for illness, vacation, or a headcount cut, the rest of the capable group absorbs the list. Saturation on the remaining trucks rises a bit. Production still gets fed. Under territorial rules the same missing driver leaves a dead zone. Others could help but the group design forbids it. That single-point failure is exactly what AOS group broadcast is meant to remove.
Milk runs still keep their own multi-stage path through the supermarket onto tugger trolleys. Direct forklift pallet work stays in forklift-capable groups. Do not merge tugger trains and pallet forks into one claim list. Do merge forklift territories that only exist on the org chart. When milk-run laps run quieter than theoretical consumption, freed people and planning headroom should thicken the shared forklift pool that can answer real pull, not draw new walls. That mix is under milk runs on theoretical consumption vs dynamic pull across lines. Fleet load patterns sit under balancing forklift fleet saturation.
A typical territory split that looked tidy and failed
A plant carved forklift groups by hall wing so each lead "owned" a crew. Wing 1 peaked every afternoon. Wing 2 ran light. Drivers in Wing 2 never saw Wing 1 call-offs. One Wing 1 absence stacked aged jobs and line waits. After logistics collapsed the wings into one forklift-capable group, with separate pools only for over-height or over-weight jobs the standard trucks could not take, the afternoon peak pulled capacity from the quiet side. Average travel per job rose slightly. Starves and idle waiting for late material fell. The map still showed longer paths. The ordering goal was never shortest path alone. It was balanced saturation and continuous delivery.
When movers are tracked at medium precision around 1–3 m (3–10 ft), acceptance place and paths land in the spatially oriented dataset. Continuous improvement and correlation and causality analytics can then ask whether territorial groups still create idle clusters beside aged jobs, or whether the shared pool is already moving capacity before the line stops.
How teams decide group design is healthy
1. Split only on physical limits - Load weight, height, attachment, aisle fit, tugger vs forklift. Not wing ownership.
2. Prefer one shared pool per capable vehicle class - Longer routes that cover peaks beat idle trucks behind a territory line.
3. Stress-test one absence - If one sickness starves a zone, the group is still territorial.
4. Watch saturation balance and aged jobs together - Hot and cold trucks in the same capable class mean the pool is still too sliced.
Group and workflow setup sits under material order automation and internal logistics optimization, with depth on the features overview. Pricing is under pricing.
Get in touch if you want to walk whether your driver groups still protect territories, or already share a capable fleet so peaks get covered without production idle from late delivery.