Module 01 · Network Design & Consolidation
With these hubs in these places, how much of the country can we actually reach in two days?
Coverage maps usually draw arbitrary mileage rings and call it service. This one derives the rings from what a driver is legally allowed to do: eleven hours of driving, a fourteen-hour on-duty window, ten hours off. Every destination is assigned to a hub, transit days are computed from that schedule, and the result is a cumulative curve — what share of demand is reachable within one day, two days, three. Hubs are variable from one to six, chosen from thirty-six real US distribution markets or dropped anywhere on the map.
Who uses it
Network design, service level owners, anyone defending or challenging a delivery promise
When
Annually, or whenever a site move is on the table
Time to first answer
Under a minute — pick a preset and read the curve
Quick start
From open page to a service level answer
Start from a preset
Tab 1 has four: the current three-hub network, a single central DC, an east/west pair, and a five-site national build. Click through them and watch the curve on tab 2 move. This alone tells you what each additional building buys.
Set your service target
Tab 3 → Target transit days and Target coverage. The default is 90% within two days. The header then shows whether you are hitting it, and by how much.
Read the cumulative curve
Tab 2. The line is the share of demand reachable within N days; the dashed red line is your target. Where the line crosses the target is the answer.
Find the exposed markets
If you miss the target, the flag underneath lists the highest-demand destinations that fall outside it. Those are the markets a customer will complain about.
Drag a hub and watch it move
On the map, hub markers are draggable. Pull one two hundred miles and the entire coverage calculation re-runs instantly. This is the fastest way to sanity-check a proposed site.
Screen map
What each tab is for
The two switches that most change the answer are on tab 2, not tab 3.
| Tab | What it controls | Leave it alone unless |
|---|---|---|
| 1. Hubs | Add, remove, move and disable hubs. Presets. Per-hub rate multiplier and capacity. Territory statistics. | Never — this is where you start. |
| 2. Service level | Population weighting, destination resolution, assignment rule. The coverage curve and transit-day distribution live here. | Never. Read it every time. |
| 3. Parameters | Driver mode, speed, dwell time, circuity, rates, and the SLA target. | You are testing team drivers, or you have contracted rates to load. |
| 4. Lookup | Query a single destination and see every hub ranked by cost, distance and transit days. | You are answering a specific customer question. |
Reading the output
How to read the curve
The single most common mistake is reading a coverage percentage that was computed with population weighting switched off.
| Metric | What it means | What to do about it |
|---|---|---|
| Cumulative coverage | Share of demand reachable within N days, weighted by market size. | If 1-day coverage is under 60% the network is probably too consolidated for next-day service, whatever the cost model says. |
| Transit day distribution | How demand splits across 1, 2, 3, 4 and 5+ day buckets. | A long tail in the 4–5 day bucket means one region is orphaned. Check which hub is serving it. |
| Territory table | Each hub’s node count, demand share, average distance, days and cost. | A hub carrying under 10% of demand is hard to justify unless it exists for a service reason. |
| Daily driving range | Miles per day implied by the current driver mode and speed. 605 mi solo, about 1,100 mi team. | This is the number that sets every ring on the map. If it looks wrong, your speed assumption is wrong. |
| SLA flag | Pass or shortfall against the target, with the most exposed markets named. | A shortfall is not automatically a problem — it is a decision to accept or spend against. |
Use cases
Three questions this actually settles
Use case 1
Do we need a fourth distribution centre to hit 2-day 95%?
| Set up | Tab 3 → set target to 2 days, 95%. Tab 1 → click through the presets from 1-DC to 5-DC, recording 2-day coverage each time. |
|---|---|
| Watch | Coverage climbs steeply from one to three hubs, then flattens. The fourth and fifth sites usually add far less than the second and third did. |
| Decide | If three hubs already clear 95%, a fourth is a next-day-coverage decision, not a two-day one. Take that finding into the TCO simulator and price it. |
Use case 2
Team drivers or another building?
| Set up | Fix the hub set. Tab 3 → switch between Solo and Team and record 1-day and 2-day coverage for each. |
|---|---|
| Watch | Team driving lifts daily range from roughly 605 to 1,100 miles. On a three-hub network that typically moves 1-day coverage by more than thirty points — comparable to adding two sites. |
| Decide | Team service carries a 15–25% rate premium. Compare that premium against the annual fixed cost of a building in the TCO simulator. Frequently the drivers win. |
Use case 3
Which markets are we quietly failing?
| Set up | Set the target to your actual customer promise. Read the shortfall flag on tab 2, then confirm individual cities on tab 4. |
|---|---|
| Watch | The flag names the highest-demand destinations outside the target, ranked by volume rather than by distance. |
| Decide | Either move the promise for those markets, or serve them from a different hub. Tab 4 shows the cost of each alternative hub for that destination. |
Pitfalls
Where people misread this tool
Every one of these has actually happened during testing.
Turning off population weighting makes Wyoming equal to California
Unweighted coverage counts nodes, not demand. On a 49-state basis that inflates the apparent difficulty of the west and deflates the northeast. A warning appears when it is switched off — heed it.
Transit days are driving days, not business days
The model counts driving time against DOT limits and rounds up, including dwell. It does not model pickup cut-offs, weekends or holidays. Real business-day lead time is often one day longer than the number shown.
Least-cost and least-distance are the same rule until you differentiate rates
Cost is a monotonic function of distance, so with every hub at rate multiplier 1.00 the two assignment rules produce identical answers. This was found during testing — the selector looked functional but did nothing. Enter per-hub rate multipliers on tab 1 to make it meaningful.
State centroids badly misrepresent large states
A single point cannot represent Texas or California. The 135-metro basis is the default for a reason; the state option exists only for comparison with coarser legacy analyses.
Data in and out
Files this tool reads and writes
| Direction | File | Purpose |
|---|---|---|
| In | hub_config.csv | Any hub set — including the optimal sites exported by the TCO simulator. |
| In | assumption_registry.csv | Rate and fuel surcharge from actuals. |
| Out | hub_config_for_optimizer.csv | The current hub set, formatted for the Consolidation Optimizer and the TCO simulator. |
| Out | network_coverage_result.csv | Per-destination assignment, distance, transit days and cost, plus the cumulative coverage summary. |
Where it sits in the loop
This is the service half of the design stage. The TCO simulator picks sites on cost; this tool tells you what that choice does to delivery speed. Run them together — the cheapest network and the fastest network are rarely the same one, and the gap between them is the actual decision.