If you're manually fixing the same route decision every cycle, the setting is wrong, not the system

A planner who overrides the same routing decision every cycle isn't just doing extra work. They're quietly erasing the one thing that would show anyone, planner or system, what's wrong. Each manual fix solves that day's route and destroys the record of why the system got it wrong in the first place.

The override fixes the route. It never fixes what caused it.

Manual patching breaks the loop that would otherwise make the system better

A route optimization system trained on delivery history improves because it has a consistent record to learn from, actual outcomes tied to actual decisions. When a planner overrides a route before it runs, or adjusts it after, the system's own account of what it would have done, and what happened as a result, gets replaced by what a person decided instead. Do this every cycle, and there's no consistent history left to learn from, for the system or for anyone trying to work out why the same override keeps happening.

A repeated override is a symptom, not a judgment call

One override might be a genuine judgment call, an unusual order, a one-off disruption. The same kind of override happening cycle after cycle is different. It usually means a setting or ruleset upstream is miscalibrated, not that the system needs a human correction it will never learn to make on its own.

An override that goes from decision to reflex

Picture a planner correcting the same kind of assignment as it comes up again and again.

• The first time, the system proposes adding a new vehicle rather than refilling an existing route mid-shift. The planner overrides it without much thought, a refill is obviously cheaper here.
• The second time, the same proposal comes up on a different route with a different order mix. The planner overrides it again, muttering that the system always tries to add a vehicle when it shouldn't.
• By the fifth or sixth time, the override isn't a decision anymore, it's a reflex. Nobody checks whether the capacity number behind it is still correct, because overriding it has become faster than questioning it.

An override that takes ten seconds and one that took ten seconds five times before it aren't the same event. The first was a correction. By the fifth, it's a habit standing in for a setting nobody's gone back to check.

Sometimes the number is wrong

Among the settings most likely to produce this pattern, vehicle capacity assumptions and the cost weighting between adding a new route and extending an existing one. If a vehicle's usable capacity is set incorrectly, too generous or too conservative, the system will keep recommending refills or reassignments that look wrong to a planner every cycle, and every override re-teaches the plan to repeat what a mis-set number told it to do. If the cost weighting favors establishing new routes over building out what already exists, the system will keep adding vehicles instead of consolidating, cheap to create, expensive to run. Weighted the other way, it will keep overloading existing routes past what a shift should reasonably carry, resisting a vehicle it should have added.

Sometimes the goal is wrong, and that's harder to catch

A route that drives long stretches between a small number of stops looks inefficient on a dashboard, a high cost per stop, a lot of empty road relative to drops. Overriding it to reshuffle stops onto a better-looking route rarely solves the underlying density problem, it relocates it onto whichever route absorbs the leftover stops. Pushed further, the same instinct leads a planner to compress delivery windows to hit a productivity number that was never the goal to begin with, trading a customer commitment for a metric that looks better on a screen.

A related pattern shows up when planning is anchored to a fleet size instead of a cost target, keeping a fixed number of owned vehicles busy, or a contracted minimum of external capacity filled, instead of letting the plan choose the cheapest, most reliable way to serve that day's orders. Protecting a fleet size is a reasonable constraint to configure for. Mistaking it for the plan's goal, instead of a boundary around cost and service, is different, and it produces overrides that look like operational judgment while protecting the wrong number underneath.

Manual judgment rarely beats a correctly configured system, and that's not a knock on the planner

In the time it takes a planner to glance at a proposed reassignment and decide it looks wrong, a correctly configured system has already checked that vehicle's remaining capacity against its next three stops, confirmed the driver's remaining hours against the shift limit, and verified the route doesn't cross into a zero-emission zone the vehicle isn't cleared for. The planner's glance caught one thing that looked off. The system's check ran through all of them, the same way, for every vehicle on the board, not just the one that happened to catch someone's eye. When manual correction keeps feeling necessary, the more likely explanation is that the system isn't currently weighing these the right way, not that a person genuinely should be.

“I would say Mover's TMS has been instrumental in driving down the cost of our operations.”

— Operations lead, grocery delivery operator

This shows up wherever a plan gets overridden by hand, again and again

Heavy goods operations feel the capacity version of this most sharply, since a slightly wrong number compounds across every refill decision made that week. Networks running a deliberate mix of owned and external fleet feel the fleet-size version, when protecting a contract or a depot's worth of trucks takes priority over the cost of serving the day's orders. Operations with wide variation in stop density feel the productivity version, watching a route get reshuffled to look better without the underlying density problem going anywhere. Different symptom, same root, an override repeating cycle after cycle is pointing at something upstream nobody's gone back to check.

The takeaway

The next time the same kind of override happens two cycles in a row, stop correcting the route and ask what's behind it, a number set wrong, like vehicle capacity, or a goal that's taken over, like protecting a fleet size instead of serving the day's orders at the lowest cost. Fixing either one upstream is worth more than fixing the same route by hand for the rest of the year.