Parcel delivery: separate packages, stops and route miles
A parcel operator can lose packages without losing many delivery stops. Another can gain packages by adding remote addresses that require extra driving and a new route. Package volume alone cannot distinguish these operating paths. The useful research object is a route’s geography and stop workload, not a generic cost-per-package trend.
Build a route-hours worksheet containing packages, unique serviced stops, driven miles, driving speed and stop service time. Keep depot handling, loading, breaks and exceptions separately visible. The purpose is to identify which workload actually changed and whether that change could alter deployed driver and vehicle capacity. A favorable average yield does not answer that physical question.
Anchor the claim in an actual operating disclosure
In UPS’s 2025 Form 10-K, locate Management’s Discussion and Analysis, U.S. Domestic Package, and the expense discussion mentioning additional stops associated with Ground Saver insourcing and the offsetting impact of lower volume. This is first-party evidence that stops and package volume can affect costs through distinct channels. It does not disclose the route-level inputs used in this article.
For a real company, extract precisely what management claims: fewer pieces, fewer stops, fewer miles, fewer routes or lower cost. Preserve the period, service product and change in operating responsibility. A parcel transferred to a postal partner can leave the company’s final-mile stops while remaining in its reported commercial volume under the relevant disclosure basis.
Do not reconstruct secret route data from group revenue. Use disclosed observations where available and label the remaining route inputs as scenarios. A scenario can test whether an explanation is mechanically possible without pretending to measure the company’s actual savings.
Reproduce the base route
Assume a fictional route delivers 180 packages to 120 stops over 60 driven miles. Driving speed averages 20 miles per hour and each stop requires two minutes, including the assumed parking and delivery activity. Loading, depot work, breaks and other activities are excluded from this first calculation and must be added later if the question concerns paid hours.
| Component | Calculation | Hours |
|---|---|---|
| Driving | 60 miles / 20 miles per hour | 3.00 |
| Stop service | 120 × 2 minutes / 60 | 4.00 |
| Route work | Driving + stop service | 7.00 |
Packages per stop are 180 / 120 = 1.5. Stop density is 120 / 60 = 2 stops per mile. Packages per route hour are 180 / 7 = 25.71. These ratios describe different operating dimensions and should not be treated as interchangeable productivity scores.
Now remove thirty packages while retaining all 120 stops and the same geography. Route work remains seven hours under the stated assumptions. Packages per stop fall to 1.25 and packages per route hour fall to 21.43. A 16.67% volume decline has not reduced modeled route hours. More packages had previously shared each parking and address visit.
For a separate scenario, reduce stops from 120 to 100 while miles stay at sixty. Driving remains three hours; service becomes 100 × 2 / 60 = 3.33 hours, giving 6.33 total. The saving is forty minutes, or 9.52% of base route work. It is not automatically a 9.52% reduction in payroll, because the driver shift and vehicle assignment may remain in place.
Assume variable route labor is valued at 30 currency units per hour only for an analytical comparison. The first scenario’s route-work cost stays at 210, increasing from 1.17 to 1.40 per package. The second scenario’s modeled route-work cost is 190. If 150 packages remain, cost is 1.27 per package. Preserve the excluded loading and fixed shift costs beside these figures.
Test whether savings change deployed capacity
Suppose the paid operating arrangement is an eight-hour shift and the remaining hour covers loading and other activities. Forty minutes less route work may create spare shift time without removing a shift. To claim cash savings, obtain evidence of changed overtime, driver assignments, route consolidation or variable contractor payments.
Route consolidation is discrete. Two routes of 6.33 hours cannot simply be combined into one eight-hour route. Their driving paths, delivery windows and vehicle capacities must support the combination. An arithmetic sum of saved minutes across a large network does not prove that equivalent full routes were actually removed.
For a network worksheet, build one row per comparable route group. Record route count and the distribution of work hours, not only their average. A group averaging seven hours can include both five-hour routes and nine-hour routes. Reducing average work may leave overtime on the constrained routes while freeing unusable time elsewhere.
Counterexample: more density with more service complexity
Consider a fictional apartment building replacing ten separate residential stops. It receives thirty packages in one visit. The route now has fewer stops and more packages per stop, but access control, elevators and individual apartment delivery make that visit last twenty-five minutes instead of the assumed two.
If the original ten stops each took two minutes, stop work rises from twenty to twenty-five minutes despite the lower stop count. Driving might fall, but the net saving must be recalculated. An operational stop definition is not a universal unit of labor. Commercial docks, signature requirements, pickups, failed attempts and large items can create the same problem.
A different boundary is a remote stop. Adding one address might add ten miles rather than the average half-mile implied by the base density ratio. The marginal route is shaped by geography. Do not use network-average miles per stop to value every new customer.
These counterexamples do not make density useless. They identify the necessary conditioning variables: service product, building type, package size, access requirements and geographic cluster. Compare within those groups before interpreting a change in average density as pure execution improvement.
Gather and reconcile the inputs
| Claim | Evidence required | Worksheet action |
|---|---|---|
| More pieces share stops | Compatible package and stop counts | Recompute packages per stop |
| Less driving | Miles and route coverage | Check route exits and geography |
| Faster service | Comparable service-time observations | Retain package and building complexity |
| Lower deployed labor | Paid hours, overtime or assignments | Separate freed time from removed shifts |
| Outsourced final mile | Responsibility and contract scope | Rebuild the stop and cost boundary |
When counts conflict, check whether a pickup and delivery at one address are one stop or two, whether unsuccessful attempts are included, and whether commercial handoffs use another counting rule. Keep the original definitions. Reclassifying all records into a guessed common definition can destroy the evidence you need to test.
Missing route miles should remain unknown. Do not obtain them by dividing costs by an assumed fuel rate, since that reverses the intended test and bundles wages, vehicle expenses and network handling into invented distance. If only group totals exist, report a range of plausible route mechanisms and specify the missing observation that would distinguish them.
Update the model after a major customer exit, final-mile insourcing or outsourcing, a service-area redesign, changing package-size restrictions or new evidence of route consolidation. State whether the evidence changed pieces, address visits, driving geography, service complexity or paid deployment. That distinction allows a concrete operating judgment without inventing savings from volume alone.
Reconcile a network change without inventing route detail
For a separate fictional district, assume ten routes each have seven route-work hours plus one paid support hour. Paid deployment totals eighty hours. Redesign reduces route work by five hours across the district, but no route can be eliminated and support time is unchanged. There are now seventy-five required hours inside the same eighty paid hours. Report the five-hour workload reduction and the unchanged paid deployment separately.
If later records show two hours less overtime, that is new evidence of a different cash channel, not permission to relabel all five hours as realized savings. Fuel consumption can also change independently through vehicle type or driving conditions. Keep route workload, fuel use and purchased labor as linked but distinct observations. The model should explain which specific evidence closes each commercial bridge.
A service address can have both a pickup and a delivery during the same visit. Check whether the disclosed stop count combines them and whether both parcel directions enter package volume. Otherwise an apparent rise in pieces per stop can be only a change in the counted activities, not improved geographical consolidation.
FAQ
Is a higher number of packages per stop always better?
No. Access, handling requirements and package size can increase the work at the same stop.
Does a lower average route time establish payroll savings?
No. Actual shift, overtime or contractor-payment changes need separate evidence.