公开(公告)号：US20220158908A1

公开(公告)日：2022-05-19

申请号：US17307187

申请日：2021-05-04

Applicant: Coupang Corp.

Inventor： JUN YE ,  Chul SEO

IPC: H04L12/24 ,  G06F40/205

Abstract: A system for live analysis of testing logs. The system including a memory and a processor configured to execute the instructions to perform operations including receiving a plurality of log entries; processing the plurality of log entries; storing the processed plurality of log entries in a database having an inverted index; receiving a query from a user device; returning test data, metadata, and statistics related to the one or more log entries; displaying the test data, the metadata, and the statistics on a GUI; comparing an exposure ratio to a range, the exposure ratio being based on a first amount of log entries associated with a first test version and a second amount of log entries associated with a second test version; and upon determining the exposure ratio is outside of the range, directing all requests from subsequent test users to a default test version.

公开(公告)号：US20220067754A1

公开(公告)日：2022-03-03

申请号：US17005232

申请日：2020-08-27

Applicant: COUPANG CORP.

Inventor： Xiaowei GONG ,  Beibei YE ,  Jun YE ,  Chul SEO

IPC: G06Q30/02 ,  G06F16/955 ,  G06F16/958

Abstract: Embodiments of the present disclosure include computer-implemented systems and methods for predicting a minimum detectable effect. The system may include at least one processor configured to execute instructions to perform steps. The steps may include sending a first webpage to a first user device and a second webpage to a second user device. The second webpage may include at least one characteristic different than the first website. The steps may include collecting user interaction data from the first and second user devices and determining a current minimum detectable effect of a user experience. The steps may include retrieving a set of historic minimum detectable effect values associated with an earlier period of time and determining a percentile rank of the current minimum detectable effect based on the retrieved set of historic minimum detectable effect values. The steps may include predicting a first and second future value of the minimum detectable effect of the user experience and aggregating the first and second future values. The aggregated first and second future value may be compared with a threshold to determine whether or not to stop an experiment and implement a change on a website.