Ten Mistakes to Avoid In Smart Cities

Introduction / "Smart" cities are old, only the technology is new

Smart city is one of those sprawling terms with a variety of meanings. To most people, the definition amounts to this: a city that connects disparate data sources to make more of existing resources than was possible, supplanting boilerplate solutions such as more roads, bloated administrations, and bigger budgets. 

The deeper meaning depends on your point of view. To Internet of Things (IoT) and IT vendors, the smart city is a market. To city leaders, it is a more efficient use of transportation and utilities that reduces pressure on budgets. To the general public, it means a more livable city, while to cynics it is a scam. A few public interest developers view the smart city as a new platform for delighting the public. 

Gordon Feller, cofounder of the smart-city advocacy group Meeting of the Minds, adds several key benefits, maintaining that smart cities embrace innovation, ensure privacy and safety, encourage engagement for all voices, and “foresee self- organizing systems to make the city more livable, sustainable, equitable, and prosperous.”1 

Using data to improve city life is actually nothing new. In his book The Ghost Map (Riverhead, 2006), Steven Johnson documents a 19th century London neighborhood beset with cholera outbreaks. Experts at the time insisted the cause was “vapors,” but a local doctor, John Snow, disagreed. He had pored over several kinds of information to validate an emerging theory—that the disease was carried not in the air but in well water. He used maps to demonstrate the correlation of deaths and the neighborhood’s water pumps. His map indicated that outbreaks seemed to be connected to the Broad Street pump, and the handle was removed. The series of outbreaks soon ended. 

Today, Google Maps functions in a somewhat similar way to steer individuals away from bad restaurants. On a city scale, though, the effect is multiplied, especially if disparate data can be woven together for a complete mirror of conditions. Even better, public- service-oriented data analysts, in the spirit of Dr. Snow, might develop insights or services no one has ever dreamed of. 

The smart city buzz is loud, but smart cities themselves are still young. With youth comes learning. “I don’t think people realize how in the beginning we really are in this industry,” says Peter Torrellas, a manager at Siemens, a global technology vendor to cities. “If [at a trade show] you saw all the booths and CIOs, you would be left with the impression that the whole problem is solved and that our job is just to do more with it. That couldn’t be further from the truth.” 

Vendors, leaders, and administrators are making countless mistakes as they create and manage smart cities. These are just 10 of them—the ones raised most often or most forcefully in six months of reading and talking to vendors, city leaders, and smart city advocates. 

Mistake One: Failing To Take An Analytical Approach To The Data 

Not-so-smart cities use data to generate reports that answer known questions but give little thought to eventual reuse of data. Truly smart cities store data to answer unanticipated questions with data discovery. 

“Connecting data is a mindset,” says Teradata-affiliated smart cities consultant Bob McQueen. It is also hugely important to understanding networks, and a connected city is a network. “You can’t understand cities without an analytical approach,” McQueen explains. “If you design for reports, all you get are reports, but 

if you design for analytics, you get the reports for free and then much more.” 

McQueen recalls when an airline executive was asked if he had seen benefits from the analytical system he built. Yes, but the benefits were a “thimbleful” compared to the “barrel of benefits he got from questions he didn’t know he was going to have.” 

What questions? It could be about a competitor or a new technology. It could also be a response to issues of operational integrity. What’s going on? Why is this system failing? 

McQueen maintains that virtually every client has had some black swan event—something that has never occurred before and will never happen again but may be worth hundreds of millions of dollars. You can’t get to the wow moments until you rule out the data errors and the systematic patterns. What’s left is where the barrelful of benefits comes from. 

Mistake Two: Ambiguous Ownership Of Data 

Cities often sign away too many rights to the data their citizens and visitors produce. “An unsophisticated customer will fail to ask [about data ownership] until a year down the road,” says Todd Walter, a chief technologist at Teradata. The city realizes its problem when it wants to analyze the data or integrate it with other data. “It’s absolutely necessary for a smart city,” Walter explains, “but they find it’s a separate negotiation and a separate price to get their own data.” 

How important is ownership of the data? “It’s the key to the whole shootin’ match,” says Bob Bennett, chief innovation officer for Kansas City, Missouri, who demands ownership from vendors. The technology “is all very, very cool, but that doesn’t make you smart,” Bennett says. The transition from cool to smart comes when the combined data from all those things helps you make decisions better or faster on behalf of citizens and visitors. To do that, a city has to own its data. 

Some prospective vendors balk at Bennett’s requirement for ownership. If they do, he says, “It’s a pretty short conversation.” His lesson: a city must articulate and enforce its ownership of data “at a very painfully descriptive level.” 

A manager (who cannot speak publicly) at a global technology vendor to cities explains that “the reason the private sector

is asking for the data is because they need to pay for [the technology]. The reason the city is giving it away is they don’t have money but need broadband or whatever they want to get out of the situation. There should be mutual benefit and some aspect of responsible governance. Right now, everyone’s trying to figure it out on their own. There’s no cheat sheet.” He believes cities should coordinate to solve these common problems. 

Mistake Three: Assuming That Sensor Formats Are Fully Baked 

You might think that in a world flush with product standards, sensors, too, would have standards. One car manufacturer—an entity that shares issues with cities—replaced engine sensors in a new model. According to Teradata’s Todd Walter, the new supplier assured them their sensor was entirely interchangeable with the previous supplier’s. Everything went fine until panicked reports came in. The new cars wouldn’t start. The car computers expected temperature readings in Celsius, but the new sensor provided temperatures in Fahrenheit. 

Sensor data has very limited standards so far, Walter explained to me in an email. “There’s no common representation of anything, from the identity of the sensor itself to the units it reports.” Analysis requires a detailed understanding of the metadata of every version of every sensor. Only then can translation and normalization be done—and only then can the data be analyzed meaningfully. 

Streaming makes analysis even more complex. The translation and normalization has to be done in-line, and none of it is static. Every new sensor, even a supplier’s updated sensor, requires updates wherever translation and normalization take place. 

“We in the analytics industry,” Walter says, “have long bemoaned the quality of data from upstream applications and the time we need to spend on ETL.” Because IoT requires far more preprocessing, translation, normalization, and manipulation of data, “the IoT world is much worse.” There’s not even a time or rate sync in the stream, Walter explains. Some sensors send every second, some every hour, and some at GHz rates; thus, combining data from multiple sensors is difficult. Gunshot detection systems, for example, sample very frequently while others sample every hour, yet each type has to be used together to spot the gunshot at the instant it happens. 

Sensors also lie, according to Walter. They send zeros, random values, and unreasonably high values. An engine sensor reports the unlikely 2,000 degrees Fahrenheit, or it simply fails to report anything at all. Creating predictive models is difficult. 

Cities have to press for standards, built-in sync services, self- diagnosis and self-checking, and hard failures instead of slow fades. Without these, Walter predicts a bumpy road ahead. 

Mistake Four: Failing To Secure Data And Devices 

Among the daily outpouring of articles, blog posts, and social media on smart cities, relatively little attention is paid to cybersecurity. 

“Do we have a standard that is being deployed uniformly in all major cities today, especially considering the dynamic IoT space?” asks a manager (who cannot speak publicly) at a global technology vendor to cities. “We do not. I have not seen a comprehensive threat analysis and programmatic response for a complete city, from back office and public services to built infrastructure.” 

The apparent lack of preparation comes in spite of recent reminders. In May 2017, the WannaCry virus froze computers in 74 countries, and in October 2016, a massive denial-of-service attack deployed thousands of ordinary (and until then asleep) devices to choke the Internet. 

Such attacks may not even require much technical know-how. An article at Ars Technica warned of ransomware for hire. 

 A new trend is multilevel marketing, which lets victims recover their encrypted data only by infecting someone else’s data. 

Dan Graham, Teradata director of marketing for IoT, maintains that cities are especially ill-equipped to manage cybersecurity and are more vulnerable to cyberattacks than almost any other entity. Unlike the federal government, cities don’t have the FBI, NSA, and CIA 

to help guard secrets. Many barely fund their police and, unlike corporations, their vulnerability goes beyond data centers and out to traffic lights, wastewater systems, and police channels. 

“We will see some high-profile attacks on cities in the next years, exposing their failure,” Graham says, and cities will remain “complacent until burned after years playing whack-a-mole with security.” Graham advocates for cities to band together and pressure technology start-ups for a solution. As a group buying software, he says, they have a good chance of creating enough demand to attract suppliers. 

2 Dan Goodin, “Booming Crypto Ransomware Industry Employs New Tricks to Befuddle Victims,” Ars Technica, November 5, 2015. https://bit.ly/2tfNsbk 

Mistake Five: Failing To Account For Hidden Costs 

One group’s “smart” is another’s usurpation, but smart is hard to oppose, points out Wired contributor and science fiction author Bruce Sterling. “You can’t say, ‘I don’t like the smart thing and I prefer to be dumb.’ You’re stealing somebody’s lunch. You’re not confronting what you’re actually taking and where the cost and benefits go.” 3 

Sterling illustrates his point with a 2014 protest by taxi drivers in Milan, Italy. They unfurled their banners and raised their chants

at the heart of what they took to be a beast: the Wired Next Fest

in Milan. Phones and databases take jobs from humans, the taxi drivers shouted. “We are already overtaxed, underpaid—the state is not taking care of us!” A spokesperson from Uber tried to explain the company’s business in the city but was shouted down. 

The drivers’ critics say they just want their old monopoly reinstated. Perhaps they weren’t universal monopolists the way Uber may become, but taxi-company complacency neglected to adopt even simple low-tech innovations. 

Some emerging technologies will actually make monopolies harder—or easier—argues Meeting of the Minds’ Gordon Feller. For example, blockchain, a software platform for digital assets, could reduce the friction of financial transactions to make way for smaller financial players. “There are all sorts of scenarios,” Feller says. 

Feller compares such disruption to the power and arrogance wielded by railroads in the 19th century, to which society responded effectively. “The pushback by cities is a clear indicator that the immune system at the local level is healthy, even as 

the corrupting influence at the national level is not clear. . . . There’s always going to be tension between local monopolies and universal monopolies.” 

3 Bruce Sterling, “Smart City States,” YouTube, July 15, 2014. https://bit.ly/2uaZ16F 

Mistake Six: Assuming People Like To Deal With Data 

Creators of smart cities often seem to forget that data has to meet the public where the public wants to meet it. That requirement calls for much less data than those who like data assume. 

Professional data analyst Max Galka found out firsthand about the public’s appetite for data after he established Revaluate, a website for Manhattan apartment hunters. It teemed with data on all aspects of each dwelling and was the kind of site he loved. 

His users perused data displayed in Excel-like tables. Each apartment had about 100 different factors, including rat problems, noise level, landlord quality, air quality, neighborhood crime, complaints about restaurants on the ground floor,

neighbor complaints about the building, and statistics about the neighborhood overall. Subjective observations supported the data. 

Galka found that “people felt overwhelmed.” A competing site, in fact, won more traffic—and it offered just one score for each apartment building. “I had a difficult time imagining how other people wanted to see [the data].” 

He learned to simplify, which seemed wrong to his data-analyst eye. “If I were to look at a building’s overall score but saw that there wasn’t any detail behind it,” he says, “I wouldn’t put much credence in it.” He found, however, that consumers preferred his new, simpler approach. 

“I think there is a big disconnect between the way that the data analyzers see the world,” he says, “and the way that decision makers do.” The data guys spend too much time focusing on the technical details without really understanding what questions they should

be trying to answer. Furthermore, when it comes to communicating the results, they present it in an overly analytical way that only other data geeks understand. 

One group did prefer Galka’s site. They worked at a large tech company known for its big data. 

Mistake Seven: Failing To Measure The “Soft Stuff” 

Something’s missing in the usual metrics, says Meeting of the Minds’ Gordon Feller. Having spent a seven-year stint at networking vendor Cisco, he knows metrics. “City leaders get the need for hard metrics on transit, water, and all that,” he says. “What most overlook, though, is the ‘soft stuff,’ the stuff that makes a city livable, fun—a place you want to spend time.” 

Although many analysts complain that such factors are hard to track, Steve Pepple, a Bay Area designer, showed how it could be accomplished. He wanted to use data to explore neighborhoods for “pockets of activity, vibrance, and new things” in order to better understand mobility, housing, and neighborhood change.4 

Pepple was inspired by the research of Eric Fischer, who had

pored over San Francisco newspaper ads dating back to 1948

and surprised urbanists with his conclusion: the usual levers for regulating demand wouldn’t mitigate the city’s lack of affordable housing.5 Pepple opted for shorter experiments than Fischer’s. Like Fischer, he used publicly available data, including social media data. With Instagram photos, for example, he tracked the flow of crowds before, during, and after the main event. 

Another experiment used what he called “color quantization” to create a “spatial database of color” to render population density and relative activity level. Eventually, in partnership with a design firm, he created tangible, street-level screens to show passersby how they could interact with data about their neighborhood and arrive at their own discoveries. 

According to Dell EMC Services’ Bill Schmarzo, the “soft stuff” also includes tribal knowledge among city staff. “There are always going to be people [on city staff] smarter than our models. We want to learn from that.” When they see their knowledge incorporated, their engagement improves. 

Operating a backhoe, says Schmarzo, is one of those functions that appear simple but are actually subtle. People would assume that digging a hole with given dimensions could be reduced to push-button simplicity. Schmarzo, who once operated a backhoe, describes how the six control knobs are played like an instrument. 

Subtle operators sense through the machine when they have hit a new level of soil or when they have just come close to a pipe. 

The hard metrics help people do simple things, but it’s the soft metrics that pay off with the subtler skills and better engagement. 

4 Steve Pepple, “Data Experiments and Design Exploration About Place,” Medium, May 20, 2016. https://bit.ly/2uJriCI 

5 Kate Abbey-Lambertz, “There’s a Profoundly Simple Explanation for San Francisco’s Housing Crisis,” Huffington Post, June 2, 2016. https://bit.ly/2vD4DEv 

Mistake Eight: Failing To Allow For Ripple Effects 

Change one thing in a city and unforeseen effects ripple throughout neighborhoods. Everything being connected to everything else is true in nature and in cities, too. 

Consider the case of a new light-rail extension. One city planned to build a new metro line to an area that had insufficient public transit. Residents, mostly lower income, commuted by car to jobs in the nearby city. Planners said the new metro station would increase the quality of life in the area. 

Already into the planning stage, planners wanted to check their projections. What would change 15 years after the station’s opening? Michel Movran, CEO of CoSMo, described to me how his team built a model, ran simulations, and presented the results. Although the new extension would attract ample ridership, an unexpected result showed up: traffic congestion would increase near the station. The city manager was aghast; everyone had thought the new station would reduce traffic. 

Gentrification was at the root of it. New, higher-income families would move in. Although the current population averaged one car per family, the newcomers would average two. They would tend to work in places closer to transit than current residents, who would not be well-served by the station. The station was attractive to the newcomers, though. 

Some ripples are avoidable. Siloed organizations prevent planners from foreseeing effects, says Movran. At a workshop he organized in Singapore, for example, city workers met others from outside their department for the first time. “Very often, people from transportation never talk to people from water,” he says. “They don’t even know each other!” 

A light-rail extension will likely cause new demand for commercial real estate nearby, but it may also displace current residents.

To mitigate such disruption, Kansas City’s Bob Bennett and his planning team have recruited community leaders: neighborhood- watch leaders, Moms Against Violence, street business organizations, school leaders, and others. Priorities shift when that group is in the same room with the engineers and city planners. 

Where do the ripples end? Perhaps the real mistake is thinking they ever will. The true smart city, much as the truly intelligent human, never stops learning and growing. 

With any luck, a city’s maturing connectedness will become

just another public feature along with transit, water, and parks. Connectedness will come to be assumed, and the “smart” in “smart cities” will drop off as connectedness is taken for granted. As the city grows smarter, its connectedness spurs imagination. Rich new features bubble up—features that spur even greater citywide vision and ambition. 

The last ripple may even be the first ripple of a new ideal that is just out of reach, which it always has been. 

Mistake Nine: Failing To Cultivate Long-Term Vendor/City Relationships 

City administrators and technology vendors come from different worlds according to those who have observed them together. That often leads to friction, disagreement, and poor results. 

Peeter Kivestu, director of travel industry solutions and marketing at Teradata, has sat across from city administrators who, he says, feel they are gazing upon “a bunch of sophisticated hustlers trying to rip them off.” Some vendors believe the city people can’t be trusted to achieve results because of disruptions in politics and policy. “We all have to make a big effort to find common ground,” Kivestu says. “Good results depend on good partnerships.” 

Bill Schmarzo, CTO of Dell EMC Services, finds fault with vendors that overcomplicate the conversation with cities. “They come out there and talk about ‘smart this’ and how you’ve got to have this technology and that gateway,” Schmarzo says. ”Until you know what questions you have, it’s really hard to know what technology and what data you need.” 

According to Kansas City’s Bob Bennett, many vendors “are very much ‘pay up front and see how it works.’” Though Bennett has found cultivating long-term vendor-city relationships difficult, he believes the effort pays off, as it did recently when a high-profile project slipped behind. 

Kansas City had planned to open a new light-rail extension and

a string of new Wi-Fi kiosks in time for the March 2016 Big 12 basketball tournament championship. Testing of the new rail line kept live electricity exposed, which delayed wiring of the Wi-Fi kiosks. Still, the city wanted to show good faith to its citizens, who had put up with several years of disruption. 

City/vendor negotiations resulted in the opening of a small number of kiosks. It might have been done with a lessor relationship, Bennett explains, but probably not. “When I asked for that favor, it was because of the relationship we’d built up that got us through. It’s good to have a partner you know and trust, who does what they say they’ll do, makes the appropriate investment, and continues on.” 

Mistake Ten: Not Anticipating Short-Term Politics In Long-Term Projects 

Incoming city leaders have a habit of making their mark by turning away from their predecessor’s projects, according to Meetings of the Minds’ Gordon Feller. 

City decision makers may not be in their jobs for the entire life span of most projects. A short project that results in meaningful change could run five years. Midsize projects may run 10–15 years, and long ones run 25 years or more. Those who initiate such projects are often “termed out” after four years; city council members may be gone in two years. 

Caltrain, the commuter rail line from San Francisco to Silicon Valley, fell victim to such a shift in February 2017. A $647 million federal grant, which the transit agency had been counting on to electrify its 50-mile corridor, was frozen. The project has stalled and political pressure to release the funding as of June 2017 has not worked. The ostensible reason: the electrification project is remotely related to California’s controversial, costly bullet train project. The state’s Republican delegates in Congress, all critics of the bullet train, have found a friend in President Trump. 

Another shift occurs when commitment for one project leaks over to a different one. In the early 1990s, the portion of a half-cent sales tax slated for refurbishing an old railroad bridge for ACE (Altamont Corridor Express) was somehow redirected toward building an extension to BART (Bay Area Rapid Transit), the commuter rail/subway system serving San Francisco and its environs. 

Shifts in commitments happen all the time, says retired transportation consultant Kenneth Ryan. “It’s a major mistake to assume that existing law will be static.” 

According to Siemens’s Peter Torrellas, some cities (e.g., Charlotte, North Carolina, and Dallas) are trying a new strategy that entails institutionalizing change with a third-party body of powerful local stakeholders. They come from academic institutions, foundations, and local industry. In Nashville, they might be in the music industry; in New York, finance. When a new mayor takes office, he or she faces an entity sitting slightly outside the city that’s driving a specific change. “The agenda doesn’t change just because a different stakeholder walks into the room,” Torrellas says. 

References 

I interviewed the following people between January and May 2017, unless otherwise indicated, for this report. 

Bob Bennett, chief innovation officer for Kansas City, Missouri 

Gordon Feller, cofounder of the smart-city advocacy group Meeting of the Minds and former executive at Cisco Systems 

Max Galka, founder of real estate site Revaluate and professional data analyst (interviewed May 2015) 

Dan Graham, Teradata director of marketing, IoT 

Peeter Kivestu, Teradata director of travel industry solutions and marketing 

Bob McQueen, Bob McQueen and Associates Michel Movran, CEO, CoSMo

Kenneth Ryan, retired transportation consultant Bill Schmarzo, CTO of Dell EMC Services 

Peter Torrellas, national business manager for state and local government for Siemens’s Building Technology Division 

Todd Walter, Teradata chief technologist, Americas 

? Ted Cuzzillo 2018