BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.com//pydata-amsterdam2026//speaker//H8X39U
BEGIN:VTIMEZONE
TZID:CET
BEGIN:STANDARD
DTSTART:20001029T040000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20000326T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-pydata-amsterdam2026-B9QGMW@pretalx.com
DTSTART;TZID=CET:20260910T150500
DTEND;TZID=CET:20260910T153500
DESCRIPTION:Standard A/B tests assume that treating user A has no effect on
  user B (SUTVA). In highly-networked products - where users message\, call
 \, share content with each other or buy/sell from each other - this assump
 tion breaks down. The result: experiments that systematically underestimat
 e\, miss entirely\, or even produce directionally wrong estimates of the t
 rue impact of product changes.\n\nThis talk surveys the landscape of appro
 aches for dealing with network effects in experimentation: from simply ign
 oring them (and when you might get away with it)\, through bi-directional 
 metrics\, translation coefficients\, double-sided gating\, and cluster-bas
 ed randomization\, to bipartite experimental design. For each approach\, w
 e'll discuss the intuition\, when it's appropriate\, and its practical tra
 de-offs.\n\nTarget audience: Data scientists\, analysts\, or any product p
 erson who runs A/B tests and wants to go beyond naive user-level randomiza
 tion. Familiarity with basic experimentation concepts (treatment/control\,
  statistical significance) is assumed\; no background in causal inference 
 or graph theory is required. Concepts are built up visually and intuitivel
 y.\n\n**Key takeaways:**\n- Why ignoring network effects can lead to biase
 d\, incomplete\, or even directionally wrong experiment results\n- A pract
 ical decision framework for choosing the right measurement approach based 
 on your product's interaction model and constraints\n- How bipartite estim
 ators can measure spillover effects while respecting user privacy constrai
 nts
DTSTAMP:20260710T150510Z
LOCATION:Main stage
SUMMARY:Your A/B Test Is Leaking: Practical Lessons in Measuring Network Ef
 fects - Dror A. Guldin
URL:https://pretalx.com/pydata-amsterdam2026/talk/B9QGMW/
END:VEVENT
END:VCALENDAR