Google Fiber in Kansas City: there is no such thing as free infrastructure

#internet #infrastructure #google

5 November 2019


by Tooran Alizadeh¹, Edward Helderop² & Tony Grubesic²

¹University of Sydney; ²Arizona State University

Imagine having access to ultra-high-speed internet, up to hundred times faster than any pre-existing service, with a competitive price, more reasonable than anything that you have ever seen before. Imagine if a major corporation decided to build such a dream-come-true telecommunication infrastructure for free. Imagine if cities were to organize campaigns, create Facebook pages, upload YouTube videos, and collect signatures to put forward their case for priority access to the next generation of telecommunication infrastructure. And finally, imagine that when a city was selected to receive the ultra-high-speed telecommunication infrastructure, you, as the citizen, had the power to determine exactly where it would be deployed! This is one way that cities could provide their citizens with the next generation of telecommunication infrastructure—just the way Google Fiber was introduced.

In February 2010, Google challenged US cities to compete for being the site of its first attempt at building an ultra-high-speed fiber-to-the-premises network (FTTP), with speeds up to 1 gigabit per second. More than 1,100 cities staged elaborate stunts as part of their applications. Kansas City, however, was announced as the winner of the competition. In 2013, Google expanded their FTTP network to Austin, Texas, and Provo, Utah; and later in 2014 announced expansion plans to 34 US cities. Google Fiber was widely acclaimed in its first few years, with Kansas City being anointed as a “broadband mecca.” The Fiber project was also roundly welcomed as a triumph for free markets, and as a model for telecommunications deployment for other cities to follow. [1]

The honeymoon phase, however, did not last long for Google Fiber. In August 2015, Google announced its intention to restructure the company, moving it into a new umbrella corporation, Alphabet Inc.

This was then followed with numerous media reports [2] suggesting that Google Fiber was under pressure by Alphabet to limit the scope of the project and the number of cities involved. In part, this pressure stemmed from corporate losses related to Google Fiber, nearly $3.6 billion in 2016. [3]

Interestingly, the details and facts on Fiber are relatively limited, with Alphabet never releasing details on the size of its investment in any of the Google Fiber cities. What is known, however, is where the construction took place, slowed, or stalled completely (e.g., San Jose and Portland). Below, we briefly examine the ups and downs of the Google Fiber—with a focus on Kansas City—highlighting what cities can learn from the experience.

PNG - 2.5 MiB
Location map of Johnson County, Kansas and Jackson County, Missouri.
T. Alizadeh et al. / Telematics and Informatics 34 (2017) 973–986, with permission from Elsevier.

Early attempts at fiberhood selection were disastrous from an equality perspective

After announcing Kansas City as the inaugural Google Fiber site, the first major step for the rollout was to determine exactly where the Fiber would be deployed.

To make this aspect of the project more transparent, Google carved the city into ‘fiberhoods’—the neighbourhoods that would receive the ultra-high-speed broadband infrastructure. For network eligibility, it was required that at least 5% of the residents of any fiberhood had to register for the service—paying a small fee of $10 during a six-week registration period. The progress of fiberhood pre-registrations was then displayed on the Fiber website using a basic map. Green areas indicated that the pre-registration goal had been met; and yellow signified that the pre-registration goal was yet to be achieved.

The pre-registration system adopted by Google proved to be a recipe for disaster; and drew strong criticism. The process manifested a sharp visualization of socio-demographic disparities and the digital divide in Kansas City: affluent white neighbourhoods easily met pre-registration targets, but lower-income, predominantly black and Hispanic neighbourhoods did not. In other words, the pre-registration process displayed how naive Google was about the implications of the deep socio-economic inequalities and pre-existing digital divide in Kansas City. For example, Google had completely overlooked that many households (up to 70% on the impoverished Missouri side) had always been without at-home internet access, making the online pre-registration process almost impossible for them. If anything, early attempts at fiberhood selection did not present any hope of erasing the digital divide in the city. Instead, it made the divide more visible.

Grassroots campaign with community groups saved the day

Under significant pressure from both the public and media in Kansas City, Google made the decision to revise their early fiberhood strategy. This was done through a grassroots campaign involving numerous community groups and significant work in the field, for Google.

In July 2012, a team of 60 Google employees got involved with grassroots community work, talking to neighbourhood associations, going to town hall meetings and church meetups to proselytize the promise of Google Fiber. The learning curve for Google was steep. They realized that many residents living in minority-dominated areas did not speak English (e.g. Hispanic communities), were not in the traditional banking system, and did not have a credit or debit card number to pre-register. Google’s effort on the ground was supported by community groups which ended up offering pre-paid debit cards—made available partly through a crowd-funding platform—for the $10 pre-registration fee. In time, the extensive grassroots community-based worked, with nearly 90% of all fiberhoods turning green in Kansas City, including many of the minority-dominated communities.

PNG - 7.3 MiB
Spatial distribution of Google Fiber in Johnson County and Jackson County.
T. Alizadeh et al. / Telematics and Informatics 34 (2017) 973–986, with permission from Elsevier.

Equitable outcome despite the lack of data transparency

To examine the Google Fiber roll-out from an equity perspective, we reached out to Alphabet to request service data on multiple occasions. All requests were denied—met with either silence, or a refusal on the grounds of security concerns. This was not unexpected. Telecommunications companies, as a rule, typically refuse to share data with academics and/or policy-makers.

Not easily deterred, we instead developed a novel data mining approach which was combined with exploratory spatial data techniques to highlight the provision footprints of Fiber for Kansas City, Provo, and Austin [4]. Specifically, our efforts to determine the spatial provision of Google Fiber for each city included several steps. First, city boundary, address points and cadastral data were obtained for each city. Next, a spatial grid was produced for each location, including its neighbouring communities. Then, a randomly selected set of addresses from each grid cell was used as an input into the ‘Check Your Address’ function on the Google Fiber website. This process was automated. Each individual query generated a response from the Fiber website that provided details on the availability of Fiber for the queried address. Lastly, we explored the demographic and socio-economic differences between areas with access to Fiber (‘haves’) and those without access (‘have-nots’). The resulting patterns were quite telling. Given the persistence of the digital divide in the US, especially within urban areas, Google Fiber had managed to manifest an equitable roll-out; and to maintain a socio-spatial distribution that favoured neighbourhoods with younger, lower income, minority populations.

It is important to remember, however, that there were significant costs to the communities embedded in this process—costs not borne by Google. A closer examination of urban governance and the Fiber projects [5] across the US has highlighted massive regulatory concessions and incentives provided to Fiber during the construction phase in Kansas City, Provo, San Antonio, Huntsville and many other cities. For example, Kansas City provided Google access to all city-owned conduit, fiber, poles, rack space, nodes, buildings, facilities, central office locations and available land. Moreover, Kansas City did not impose any charges for access to these facilities, nor did the city require any permit or inspection fees. Additional concessions were made by Kansas City, including an agreement to use third-party inspection firms, selected by Google, to ensure compliance during the installation process. In addition, Kansas City provided Google with access to all municipal GIS data and technical information databases, cooperation in publicity and marketing efforts effort and assistance to Google in obtaining settlement-free interconnections with anchor institutions in the city that had existing network connections (see the Development Agreement for more information).

In short, although Google designed and installed the network, much of the administrative cost was absorbed by Kansas City and its surrounding communities with taxpayers forced to cover most of the indirect costs for the network. It has been argued that Kansas City’s support for Google’s network went well beyond deregulation and in some instances local efforts were described as ‘corporate welfare.’ For example, the media [6] questioned fee waivers for the use of rights-of-way; citing these corridors under Kansas City’s streets and on its utility poles as a scarce, taxpayer-owned resource. When a city offers a private company access to those resources for free, it’s forgoing an opportunity to collect revenues. Questions were also raised as taxpayers, rather than Google, paid to hire extra city staff to supervise the project.

There is, however, a second side to this debate. Specifically, there are questions pertaining to the economic logic of a private firm building new fiber networks without taxpayer subsidies. If one accepts this argument, then there is an obligation to ensure that the resulting infrastructure and associated services are equitably distributed for all residents, regardless of location or socio-economic status. In the case of the Kansas City metropolitan area, where taxpayers did subsidize the roll-out of Fiber, the results of our analysis suggest that both Kansas City and Google Fiber met this obligation in terms of achieving an equitable distribution of telecommunication services.

What can cities learn?

Google Fiber is a complex project with important lessons for cities seeking equitable telecommunication infrastructure: First, Google Fiber’s success in deploying an equitable socio-spatial roll-out was only possible after Google realized the need to run a highly localized, grassroots campaign—in collaboration with community advocate groups—to reach out to the socio-economically impoverished sections of the city. This multilateral partnership between a telecommunication provider (Google), local government, and a community (i.e., institutions and residents)—is more complex than the broadband debate and industry generally allows for. Second, Google Fiber’s refusal to release detail information on their network to the public, slowed, but did not stop our research into the equity implications of the Fiber network. Indeed, advancements in data mining approaches and spatial analysis techniques make it difficult for both governments and telecommunication providers to fully avoid public and academic scrutiny. Last, but not least, we praise Google Fiber’s effort in working with community groups in Kansas City, as well as ensuring the equitable provision of Fiber in Provo and Austin. However, we acknowledge the massive cost borne by taxpayers in each city—via the numerous concessions and incentives received by Google Fiber. In other words, there is no such a thing as free infrastructure. Cities need be aware of the hidden cost whenever partnering with corporations.

↬ Tooran Alizadeh, Edward Helderop & Tony Grubesic

Tooran Alizadeh is a senior lecturer, director of urban design program, and also a recipient of the prestigious Research Accelerator Fellowship (SOAR) at the University of Sydney. Tooran is an interdisciplinary academic researching policy and planning implications of telecommunication infrastructure with a focus on the broadband deployment, and smart cities.

Edward Helderop is currently working on his PhD in Geographical Sciences. His research interests include GIScience, big data, and network analytics (particularly as applied to urban infrastructure systems). His previous research explored turnover and resiliency in plant-pollinator networks.

Tony Grubesic serves as Director for the Center for Spatial Reasoning & Policy Analytics at Arizona State University. His research interests include spatial analysis, geocomputation, urban informatics and policy evaluation.