Gaylord to doorstep: running a same-day hub sort that doesn't fall apart
The difference between a hub that clears by noon and one that clears at 6pm is not headcount. It's whether the sorter has to think.
- Upload the manifest before the truck arrives — scanning then fills itself in.
- Never let a sorter decide which bin a parcel belongs to.
- Confirm placement with a bin scan or your route data is a guess.
- Dispatch the whole floor in one action, not bin by bin.
A hub sort has exactly one job: convert an undifferentiated pile of parcels into dispatched, drivable routes before the day runs out. Everything that slows that down is either a decision a human should not be making or a piece of missing information.
Before the truck arrives: load the manifest
If your retail partner can send a parcel list — a CSV, an export, a spreadsheet with tracking numbers and addresses — you have already removed the slowest step from the entire day. Every scan on the floor becomes a lookup instead of a data-entry task.
When there is no manifest, the fallback is capturing the shipping label and reading the address from it. That works, but it costs a few seconds per parcel and those seconds are the difference between clearing at noon and clearing at four.
One manifest upload saves roughly 6–10 seconds per parcel. On 800 parcels that is over an hour of labor.
The two-beat scan
Good sorting is rhythmic. Scan the parcel, the screen shouts a bin, you drop it in, you scan the bin. Two beats, no thinking, no lookup sheet taped to the wall.
The bin assignment should be computed, not chosen. Clustering by geography, weight cap and stop count produces routes that are 15–30% shorter than routes built by a human sorting into 'north / south / downtown' buckets, because a human cannot see the whole day's geography while holding a box.
- Beat one: scan the parcel. The system resolves the address and picks the bin.
- Beat two: scan the bin QR. Placement is now provable and the parcel is locked to that route.
- Mismatch: the screen blocks it. Wrong bins are caught at the bin, not at the doorstep.
Rebalance late, not early
The clustering that made sense at parcel 40 will not be optimal at parcel 600. Rather than re-planning constantly, let bins fill and run a single rebalance near the end of intake. That moves the outliers, evens out the stop counts, and leaves you with routes that a driver would actually recognize as sensible.
Rebalancing before dispatch is cheap. Rebalancing after a driver has loaded their vehicle is not — so make the cutoff explicit and hold it.
Dispatch the floor in one action
Publishing bins one at a time invites partial dispatch: three routes offered, four bins sitting on the floor because someone got pulled into a call. Dispatching everything at once means the floor is either out or it isn't, and any exception is visible immediately.
Each bin becomes its own multi-stop load with an optimized stop order, mileage, and estimated drive time attached. Carriers see the payout and the stop count and accept the ones that fit their day.
Handle the exceptions on purpose
Two exceptions will happen every single day: a parcel with no usable address, and a parcel that was scanned in but never bin-confirmed. Both need a visible queue and an owner. Left alone, they turn into the 4pm scramble that makes people believe hub sorting doesn't work.
- Unmapped address → hold, correct, re-scan into a bin.
- Scanned but unconfirmed → find it on the floor before dispatch, not after.
- Damaged or mis-shipped → pull it out entirely and notify the partner the same day.
Sort, route and dispatch your hub in one place.
FASTLANE gives drivers one-tap routes, shippers automatic parcel sorting, and retailers a last-mile network they can actually audit.