Key points
- Indicators decree what a city must pursue; informality, care and displacement go unmeasured and stop existing.
- What exists in downloadable form gets measured: Global South cities fall to the bottom through statistical privilege.
- Söderström: the smart city is corporate storytelling in which metric and solution come from the same vendor ecosystem.
Smart for whom is the question almost nobody asks when a city hall celebrates its position in a smart city ranking: fourteenth, seventh, third in the world. The big indices, from the IMD Smart City Index to the IESE Cities in Motion, present themselves as neutral thermometers of a measurable reality, and they are not: they are devices of government dressed as statistics, and opening the methodology shows the machinery at work. The first problem is that indicators do not describe the city; they define it. When a ranking decides that urban intelligence is measured by sensor density, bandwidth, digitised procedures or connected parking spaces, it decrees that this is what a city must pursue, and everything left out of the spreadsheet stops existing for public policy.
Indicators do not describe the city: they define it
Informality is not measured, nor the care work that sustains the daily metabolism of any metropolis, nor the silent displacement of the neighbours who can no longer afford the district they made habitable, nor power. Rob Kitchin warned it clearly: urban data are never raw, they are always cooked, and someone chose the recipe. The second problem is textbook: most indices are built on the availability of data rather than its relevance, measuring what exists in downloadable form, and the result is a tautology, the most digitised cities score best on digitisation and the cities of the Global South appear at the bottom, not because they work worse but because they do not produce their data in the format the ranking can read. It is a geography of statistical privilege dressed as technological merit.
Availability, weighting and Goodhart's law: an opinion with decimals
The third problem is weighting, that hidden moment in a methodological annex where someone decides that connectivity weighs three times as much as housing; change those weights and the whole podium reorders. A ranking is not a finding but an opinion with decimals, and when that opinion becomes a target, Goodhart's law applies: the measure stops being a good measure. City halls optimise the indicator instead of the city, installing smart lampposts on the main avenue while the periphery still lacks night buses, because the lamppost scores and the bus does not.
Söderström, Hollands, Zuboff and Kitchin: the ranking as corporate catalogue
The indices do not come from nowhere. Ola Söderström, Till Paasche and Francisco Klauser showed in 2014 how the smart city works as corporate storytelling, a narrative built by technology vendors who first manufacture the metric and then sell the solution that metric demands; the ranking is the catalogue, and the consultancy that publishes it and the firm that installs the sensors share ecosystem, sponsorship and language. Shoshana Zuboff would call it data extraction turned into infrastructure of accumulation; Robert Hollands asked in 2008 whether the real smart city would please stand up; and Paolo Cardullo and Kitchin showed that on the participation ladder of these models the citizen is almost never a citizen: a user, a consumer, a data generator, never the one who decides.
What should be measured and why it does not fit on a podium
What should be measured is the question that follows. Perhaps the real time a person without a car takes to reach a public hospital; how many households were evicted last year; what share of the municipal technology budget is locked in proprietary licences the council can neither audit nor abandon; who owns the data the city produces every second. None of those questions fits on a podium, which is why they are not there. Alternatives exist, such as indicators of access to services by neighbourhood, gender budgets or public algorithm registers, which measure the city by what it does with its residents rather than by what it buys.
The lesson is that the next time a mayor celebrates climbing four places in a global smart city ranking, it is worth not applauding yet and asking who built the rule, who sells it and who is left out of the frame. The city that is not measured does not disappear: it simply stops counting, and that, and nothing else, is the most important political decision taken in a spreadsheet.
Frequently asked questions
Why are smart city rankings misleading?
Because they define urban intelligence by what they can measure in downloadable form, sensors, bandwidth, digital procedures, omit informality, care, displacement and power, depend on arbitrary weightings that reorder the podium and work as a catalogue for the vendors who sell the solutions the metric demands.
What should a city measure instead of a ranking?
The real time a person without a car takes to reach a public hospital, how many households were evicted last year, what share of the technology budget is locked in licences that cannot be audited and who owns the data the city produces, indicators that measure what the city does with its residents rather than what it buys.
