Drive-time vs radius trade areas: why distance is not access

A radius trade area is a circle. You pick a center and a distance, say three miles, and everything inside counts as your market. Drawing one takes seconds, which is the whole appeal, but the circle assumes the ground is flat and travel runs at a constant speed in every direction. A drive-time trade area starts from the same site and measures how far a person can actually get within a time budget, say ten minutes, along the real road network. Its shape stretches down highways and pinches wherever a river or a rail corridor blocks the way. The circle reports distance from the site. The drive-time shape reports what the site can reach.

• Barriers. Rivers, interstates, rail corridors, and large parcels with no through-roads all fall inside a circle while staying out of reach on any reasonable trip. The ring counts the households on the far bank. The road network does not, because no road crosses to them.
• One-way and limited-access networks. On divided highways and one-way grids, the trip out and the trip back follow different paths and take different amounts of time. Network routing captures that asymmetry. A ring has no notion of a road, so it treats the journey as the same in both directions.
• Time of day. Congestion at rush hour can cut a ten-minute reach to a fraction of its midday size. A circle looks identical at 8 a.m. and 8 p.m. A drive-time shape can be built for the conditions that match when your customers are on the road.

• Radius (ring). A fixed circle around the site. It draws instantly and works as a first pass, though it knows nothing about roads, barriers, or traffic.
• Drive-time or walk-time (isochrone). A shape built from the road or pedestrian network that marks everywhere reachable inside a travel-time budget. It tracks real access, which is why most retail and service analysis leans on it.
• Customer-derived. A trade area drawn from where real customers come from, using loyalty records, card transactions, or anonymized mobility data. It rests on observed behavior rather than assumption, but you only have the data for sites that already operate.
• Probabilistic (Huff and gravity models). Instead of a hard edge, every location carries a probability of visiting your site, set by how attractive the site is and how costly it is to reach, with nearby competitors pulling that probability down. It handles overlapping, fuzzy markets well, and it demands the most data to calibrate.

The choice has direct consequences for the brief. Whatever boundary you draw is the boundary your data gets summed inside, so a change of method changes the totals. Replace a three-mile ring with a ten-minute drive-time and the population, the median income, and the daytime demand inside the boundary all shift, sometimes sharply, because you have pulled in the neighborhoods a highway feeds and dropped the ones a river walls off.

Competitors move with the boundary too. A rival inside your ring but on the far side of an uncrossable interstate competes for almost none of the same trips, and a drive-time shape tends to leave it out. A different rival, just past the edge of the circle but one freeway exit away, may belong squarely inside your real catchment. Score competition against the access-based boundary and both the rivalry and any overlap with your own nearby stores read differently than the ring suggested.

Rings still earn their keep in the right spot. When you are screening a long list of candidates down to a shortlist, a radius is quick to compute and easy to defend in a room, and the rough cut seldom turns on network detail. Dense, uniform street grids with no major barriers are another fair case, since there the straight-line circle and the real network nearly agree. A practical default: triage with a radius, then switch to drive-time before any number reaches a real estate committee or a capital decision.

Cars are the wrong unit for some formats. An urban cafe, a convenience store, or any business that runs on foot traffic draws its customers off the sidewalk network, so a walk-time isochrone is the trade area that matches reality. Five minutes on foot covers the blocks someone will actually cross for a coffee, and it honors the same barriers a car faces plus a few of its own, such as a freeway with no pedestrian crossing. Where transit carries the demand, the catchment hangs off stations and lines rather than rings. Match the travel mode to the format and the trade area stays tied to how people really arrive.

If the method moves the numbers, the method has to ride along with them. A trade area recorded as nothing more than a population figure cannot be rebuilt by anyone else. Record it as a ten-minute peak-hour drive-time over a network of a stated vintage, with the demographic year attached, and a colleague or a skeptic can reproduce it or argue with it. Geod stores the trade-area definition next to the demand and competition counts, so the brief shows both what fell inside the boundary and how the boundary was drawn. A reviewer can then check the method instead of taking the result on faith.