Readers love lists, especially city rankings. There are lists that rank the world’s leading cities of opportunity, the most sustainable cities, the bike-friendliest cities, the top shopping cities, and even the most competitive cities in the future. What they all share is an attempt to measure cities. But what defines a city? The answer isn’t as clear-cut as it seems.

Every year, leading corporations fund the publication of an increasingly large number of benchmarking studies, which generate significant interest in the media. Even the UN has jumped on this bandwagon by adapting, for the first time, an urban goal within the Sustainable Development Goals framework. However, the basic question of what constitutes a city is often defined inconsistently across rankings. This could leave general-interest readers and policymakers, confused, or worse—misled.

The debate on how to define a city is not new. National census bureaus, international organizations, and independent researchers have all developed carrying definitions based on their unique perspectives. Readers must always be consciously aware of the specific unit of analysis underlying each ranking before deciding whether City A is more sustainable or competitive than City B.

The Economist Intelligence Unit’s Hot Spots 2025 report for instance defines a city as “the urban agglomeration or metropolitan area it holds together," which has neither administrative nor functional reference. One may argue that a city is “held together” by its public transport network, commuting patterns, the density of its population, or even support for its local sports teams.

In its Cities of Opportunity report on the other hand, PricewaterhouseCooper uses the strict administrative borders of an urban settlement to define a city. So the city of Paris is equivalent to its 20 districts and New York corresponds to the territory of its five boroughs. Such a definition is arguably clearer, but is inconsistent with the real patterns of urban settlements. In both Paris and New York, the metropolitan agglomerations go far beyond administrative boundaries.

Other reports simply avoid any discussion on this issue. Others simply extrapolate indicators from higher levels of aggregation. Arcadis’s Sustainable Cities Index for instance uses a national urban average to assess solid waste management and sanitation coverage at the city level, and relies on national energy efficiency and ease of doing business indicators. This is justified due to the non-existence of specific city-level data on such indicators, but it is obvious that average national indicators are seldom representative of specific cities.

In policy analysis, units of analyses directly impact rankings, and inconsistencies could lead to inaccurate interpretations potentially leading to bad policies. Looking at the quality of urban public services, GDP per capita, or transportation systems—all measures that form key parts of the Hotspots 2025 report—New York City and its five boroughs would rank differently than the larger metropolitan area. Given spatial segregation of high- and low-income households in most urban areas, city centers are arguably not representative of entire metropolitan regions.

While city rankings will continue to attract attention, both in the media and among policymakers, readers should step back and fully understand what is being measured in each case.The producers of urban indices and rankings on their part should also be more conservative in propagating policy lessons from their tools by clearly highlighting key definitional assumptions.