Upgrade to Pro — share decks privately, control downloads, hide ads and more …

antitrust uses and misuses (in the age of Big Data)

antitrust uses and misuses (in the age of Big Data)

Talk I gave at the conference "Designing antitrust for the digital era" (Brasília, Brazil). TL;DR: antitrust agencies shouldn't go around imposing fines and breaking up companies just because "network effects" or "data as an essential facility". Public policy must be evidence-based. And the evidence suggests that network effects and access to personal data are not preventing competition.

Thiago Marzagão

August 01, 2019
Tweet

More Decks by Thiago Marzagão

Other Decks in Research

Transcript

  1. antitrust uses and misuses
    (in the age of Big Data)
    Thiago Marzagão

    View full-size slide

  2. 76%
    54%
    Facebook
    Pinterest
    others
    16%
    33%
    48%
    60%
    35%
    27%

    View full-size slide

  3. • Facebook knows your likes, posts, comments,
    social graph, what you follow, what you mute/
    block, location, device, OS

    • plus your data from WhatsApp and Instagram

    • Facebook has 2.3 billion users

    • people are not porting their FB data elsewhere

    View full-size slide

  4. • Tinder knows what types of people you want to
    date

    • not the fluffy stuff you say you want (“kind”,
    “adventurous”), but the stuff you actually want

    • and how that varies across markets (US vs
    Europe, Millennials vs Generation Z, etc)

    View full-size slide

  5. • Uber knows where you live and work, your
    favorite restaurants and bars, your daily
    schedule, your hobbies, how impatient you are

    View full-size slide

  6. monetizing data is harder than
    most people realize

    View full-size slide

  7. • CADE: 2 years to approve Microsoft/Skype deal

    • Zoom: 4 months to go from 0 to 1M users

    View full-size slide

  8. about
    • data scientist at the Observatory of Public Spending

    • Ohio State University Ph.D.

    • researcher (thiagomarzagao.com/publications)

    • occasional professor of machine learning (thiagomarzagao.com/teaching)

    • amateur roboticist (github.com/thiagomarzagao/ev3py)

    • @tmarzagao, [email protected]

    View full-size slide