The 15-Minute City Beyond Downtown: Metrics That Actually Matter

Measure network minutes, not circular buffers

The 15-minute city is often drawn as a circle around a home, but people move along streets, crossings and paths. A straight-line distance of seven hundred meters can take twenty minutes if a motorway, railway or steep slope sits in between. Indicator reviews identify mobility, time, distance, speed and amenity functions as central variables, which means the spatial unit is less important than the network used to calculate reach.

Start with a pedestrian network graph rather than a radius. In a Copenhagen-based GIS study, even a city known for sustainable mobility shows spatial clusters of insufficient access in suburban locations. Density largely predicts both walking accessibility to local destinations and access to other areas by public transport. If your metric cannot see these clusters, it gives little guidance for interventions.

• Use street-network routing, not Euclidean buffers.
• Include barriers, crossings, gradients and lighting as network constraints or separate quality checks.
• Run sensitivity checks with slower walking speeds for children and older adults.
• Report median network minutes, not average as-the-crow-flies distance.

Split everyday and citywide services

A pharmacy and a regional hospital do not belong in one basket. Everyday functions such as food, primary care, childcare, transit stops and green space should be within a short walk. Rare and citywide destinations such as universities, theatres and major hospitals are served by transit and subcentres, not by replicating one in every block. Mixing them into one score rewards downtowns with many amenities and hides functional gaps in residential edges.

A study of Krasnodar, Saratov and Naberezhnye Chelny using open data and 2GIS illustrates the difference. For daily services, 86 to 92 percent of residents live within a 15-minute walk across three morphologically different Russian cities. Once rare citywide facilities enter the index, differences widen sharply. Low-rise areas show average walking times about three times longer than historic centres and twice as long as Soviet microdistricts.

Practical rule: compute a basic index only from everyday services, and handle rare destinations in a separate transport-access layer.

• Basic set: food, pharmacy, primary education and healthcare, transit stop, public green space.
• Citywide set: assessed through stop access, frequency and transfer reliability, not walking radius.
• Percentages can look similar for daily services even when morphology differs.
• New high-rise districts often lag because modern regulations are less prescriptive about social provision.

Six metric families and their blind spots

A systematic review grouped performance metrics into six families: amenity-based, population-based, distance-based, gravity-based, behaviour-based and weighted scores. Amenity-based counts how many service types are reachable. Population-based turns that into the share of residents with access. Distance-based measures the actual route. Gravity-based adds attractiveness and capacity. Behaviour-based observes how people actually travel, and weighted scores combine indicators into a single number.

Each family fails in a predictable way. An amenity count does not know whether a facility is open, accessible or economically viable. A population share can hide the fact that the missing eight to ten percent is concentrated in the least connected neighbourhoods. Gravity and weighted models depend on weights that are rarely grounded in local preference data. Behavioural metrics are closest to lived experience but require mobile or survey data and raise privacy concerns.

Resist the urge to publish one magic number. A defensible panel shows two or three of these families side by side.

• Amenity-based: fast but blind to quality and operating status.
• Population-based: policy-friendly but averages out spatial inequality.
• Distance-based: honest about routing but silent on choice.
• Gravity-based and weighted scores: richer, yet sensitive to unproven weights.
• Behaviour-based: most realistic, needs careful data governance.

One opportunity is not enough

A distinction proposed in an assessment framework for 35 cities is useful for strategy work. The basic 15-minute indicator requires at least one opportunity for each essential service type. The optimal indicator requires at least three independent opportunities, which captures diversity and choice rather than mere presence. A single store can close, change owner or fail to meet demand.

Stress-test your index by removing the largest provider in each category and recalculating. If the share of residents with access collapses, the territory is formally close to the threshold but operationally fragile. Classification thresholds such as achieved, achievable and unlikely using fifty percent and twenty-five percent population benchmarks help prioritise cities and districts, but they should never replace a block-level look at why access is missing.

• Baseline: at least one opportunity per type.
• Optimal: at least three independent alternatives.
• Stress-test each type by removing its largest provider.
• Use city-level categories for prioritisation, not for final site decisions.

A minimum panel you can maintain

For a territory strategy, a sustainable panel answers three questions: is there access, is there choice, and what is the route quality. Suggested core indicators are median network minutes to the three nearest providers of each everyday type; share of residents with a full basic set; share with fewer than three alternatives per type; median time to a transit stop; share of routes with continuous sidewalk, lighting and safe crossings; and the local share of trips made on foot or by bicycle.

Source data from open facility catalogs, street networks, transit feeds and, where lawful and aggregated, mobility data. Publish the assumptions along with the number: walking speed, facility set, data vintage and how barriers were treated. Otherwise an index becomes marketing rather than management.

Compare morphotypes such as historic core, microdistrict, new high-rise and low-rise within the same territory. This quickly separates deficits solved by a single provider, by rerouting, or by capital investment in a subcentre.

• Cap the published panel at about seven indicators.
• Show medians, not only averages, for network time.
• Disaggregate by urban form rather than comparing whole districts.
• Refresh facility and network data annually or after major openings and closures.