How Big Data Confers Market Power to Big Tech: Leveraging the Perspective of Data Science

• Published date: 01 September 2020
• Author: Cristian Santesteban,Shayne Longpre
• DOI: 10.1177/0003603X20934212
• Date: 01 September 2020
• Subject Matter: Articles

Article

How Big Data Confers Market

Power to Big Tech: Leveraging

the Perspective of Data Science

Cristian Santesteban* and Shayne Longpre**

Abstract

Data-hungry applications are central to the largest online platforms. Using a novel approach that

leverages data science to inform the economics, we demonstrate how data is a source of market

power. We highlight the importance of data heterogeneity, whereby small feature differences translate

into large value differences. We examine how concept drift, the existence of a nonstationary rela-

tionship between the predictive and target variables, implies that access to a continuous stream of data

is competitively advantageous. We analyze how an information bottleneck and high sample complexity in

existing applications lead to increasing returns to data. Finally, we show how user interaction control

enables personalization that raises switching costs. The combined effect is a potent data barrier to

entry that endows substantial market power to only the largest online platforms. Competition policy

should focus on enabling entrants unfettered access to vast continuous data streams similar to those

available to platform incumbents.

Keywords

digital economy, online platfor ms, Big Data, market power, barrie rs to entry, Big Tech, machine

learning, data science, artificial intelligence

I. Introduction

Online platforms such as Amazon, Apple, Facebook, and Google have undergone intense economic,

social, and political scrutiny in the last few years.

1

Formerly praised for their innovative technologies,

* RedPeak Economics Consulting and Department of Economics, University of Washington, Seattle, WA, USA

** Apple Inc., Cupertino, CA, USA

Corresponding Author:

Cristian Santesteban, RedPeak Economics Consulting, 3226 Fuhrman Ave E, Seattle, WA 98102, USA.

Email: cristian@redpeakecon.com

1. See, e.g., Derek Thompson, America’s Monopoly Problem: How Big Business Jamm ed the Wheels of Innovation,THE

ATLANTIC, Oct. 2016, https://www.theatlantic .com/magazine/archive/20 16/10/americas-monopoly- problem/497549/ (last

visited May 2, 2020); Cecilia Kang & David McCabe, FTC Broadens Review of Tech Giants, Homing in on Their Deals,

N.Y. TIMES, Feb. 11, 2020, https://www.nytimes.com/2020/02/11/technology/ftc-tech-giants-acquisitions.html (last visited

May 2, 2020); Tony Romm, Companies Burned by Big Tech Plead for Congress to Regulate Apple, Amazon, Facebook and

The Antitrust Bulletin

2020, Vol. 65(3) 459–485

ªThe Author(s) 2020

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/0003603X20934212

journals.sagepub.com/home/abx

these firms and their data-dependent services have taken center stage in debates over market concen-

tration, the efficacy of antitrust tools, and the boundary between consumer protection and competition

policy. Mounting evidence and the expanding dominance of these online platforms, referred to some-

times as “Big Tech,” have led many observers to claim that the largest platforms command significant

market power across many markets.

2

As such, the debate has shifted to examine the sources and

legitimacy of that power; whether antitrust policy has a role in curbing such power, and if so, whether

existing antitrust tools are up to the task. We focus mainly on the first aspect of this debate, and in

particular, whether and how the accumulation, analysis, and use of data confer online platforms with a

formidable competitive advantage that translates into substantial market power.

3

The focus of our analysis will be on the applications at the heart of the online platforms’ products

and services; namely, targeted advertising, and more generally, information filtering systems, as well

as web search and virtual assistants. All of these applications are rooted in data-hungry machine

learning (ML) and artificial intelligence (AI) algorithms. As others have noted, what some call “Big

Data” can provide market power to firms if it generates barriers to entry that prevent rivals from

competing effectively.

4

This point is particularly relevant in the presence of increasing returns to scale

that could result in “winner-take-all” markets.

5

In these environments, a differential advantage in

access to a key input for innovation, such as the high-quality consumer data collected by the online

platforms, could tip the market toward a single dominant firm. We propose that it could also lead to an

inefficient perpetuation of the d ominant firm’s market power due to a d istortion of the dynamic

competitive process.

6

Namely, a data-driven barrier to entry in these online marketplaces may bias

the playing field in favor of the incumbent platforms and inefficiently entrench their market power.

The reason is, curiously, because the very input required to generate the products and services to topple

Google,WASHINGTON POST, Jan. 17, 2020, https://www.washingtonpost.com/technology/2020/01/17/companies-burned-by-

big-tech-plead-congress-regulate-apple-amazon-facebo ok-google/ (last visited May 2, 2020). Microsoft, in contrast , has

gone largely unscathed. E.g. Steve Lohr, How Top Valued Microsoft Has Avoided the Big Tech Backlash,N.Y.T

IMES,

Sept. 8, 2019, https://www.nytimes.com/2019/09/08/technology/microsoft-brad-smith.html (last visited May 2, 2020).

2. See, generally, Fiona Scott Morton et al., Report of the Committee for the Study of Digital Platforms: Market Structure and

Antitrust Subcommittee,S

TIGLER CENTER FOR THE STUDY OF THE ECONOMY AND THE STATE, July 1, 2019 [hereinafter “Stigler

Center Report”]. Jason Furman et al., Unlocking Digital Competition: Report of the Digital Competition Panel,G

OVERNMENT

OFUNITED KINGDOM, Mar. 2019 [hereinafter “UK Report”]. Jacques Cr´emer et al., Competition Policy for the Digital Era:

Final Report,E

UROPEAN COMMISSION, 2019 [hereinafter “EC Report”]. Digital Platforms Inquiry: Final Report,AUSTRALIAN

COMPETITION &CONSUMER COMMISSION, June 2019 [hereinafter “ACCC Report”].

3. Of course, we are not the only ones to explore this topic. Many commentators have wondered whether these firms’ market

power may be due to the platforms’ reliance on services that rely on data. For example, Florian Zettelmeyer, The Economics

of Big Data and Personal Information, presentation at the Federal Trade Commission, FTC HEARING 6, SESSION 2, Nov. 6,

2018, https://www.ftc.gov/news-events/audio-video/video/ftc-hearing-6-nov-6-session-2-economics-big-data-personal-

information (last visited Feb. 24, 2020) (asking whether “the current rise of AI and machine learning” may be the key

ingredient, given that “prediction machines that have large effects on the quality of provision of services - those can’t work

without data.”) and Brett Gordon et al., A Comparison of Approaches to Advertising Measurement: Evidence from Big Field

Experiments at Facebook (KELLOGG NORTHWESTERN,Apr. 2018), https: //www.kellogg.northwestern.edu/faculty/gordon_ b/

files/fb_comparison.pdf. For a contrary view, see Catherine Tucker, Digital Data, Platforms, and the Usual [Antitrust]

Suspects: Network Effects, Switching Costs, Essential Facility,54REV.INDUS.ORG.683 (2019) (finding little evidence that

“digital data augments market power due to either network effects or switching costs”).

4. See Michael L. Katz, Multisided Platforms, Big Data, and a Little Antitrust Policy,54R

EV.INDUS.ORG.695 (2019) (defining

Big Data as “data sets that are large due to a combination of breadth, depth, and frequency of collection”).

5. See,e.g., ROBERT H. FRANK &PHILIP J. COOK,THE WINNER-TAKE-ALL SOCIETY:WHYTHE FEW AT THE TOP GET SOMUCH MORE

THAN THE REST OF US(1996).

6. This concern is separate from market power obtained on the merits of superior innovation. In these cases, the primary

competitive dimension is directly contingent upon the scale and quality of data. A rival firm could match or even exceed the

incumbent’s product on a number of competiti ve dimensions (user-interface de sign, marketing, business strate gy, and

engineering), but without access to the incumbent’s data or user base, their data-dependent applications will not be

competitive. We discuss this point more throughout this article.

460 The Antitrust Bulletin 65(3)