Submitted by: Submitted by michelb
Views: 436
Words: 474
Pages: 2
Category: Business and Industry
Date Submitted: 05/24/2012 04:55 PM
Online Shopping Site Performance
Topic Selection – What
Team 7 will be researching the topic of measuring and improving Internet performance, specifically online shopping sites during high-traffic time periods. The data that will be used for the analysis is provided from Cedexis who collects the largest and fastest growing repository of such data at ~1 billion data points daily across ~200 countries and 32,000 networks (including the world’s largest ISPs and mobile carriers).
Data analysis performed by Team 7 will focus on the following dimensions:
1. Major online shopping sites – Amazon, eBay
2. Major shopping days –Cyber Monday
3. Country – USA
4. Candidate areas to explore with our hypothesis testing include:
a. Usage/performance experienced by visitors of online shopping sites during high-traffic time periods
b. Usage/performance experienced by visitors over mobile vs. non-mobile during high-traffic time periods
Development of the Hypothesis Testing Plan
The hypothesis testing will compare within 95% level of confidence if the connect time is the same between Cyber Monday and other non high-traffic days. The team’s current assumption is that connect time will be slower on Cyber Monday and will use the data collected from Cedexis to confirm or to disprove the assumption. Attached to this document is an example spread sheet (Titled “Game 7 ALL DATA”) with data collected from Cedexis for Game 7 of the World Series. This spread sheet includes similar data to what will be analyzed for Cyber Monday. The Hypothesis test is presented below:
H0: Performance (connect time) is the same on high-traffic days vs. normal-traffic days
H1: Performance (connect time) is not the same on high-traffic days vs. normal-traffic days
Supporting Literature
Several sources cited below illustrate the mission-critical nature of network performance to delivering customer experiences online. Major citations are listed below:
1. Users abandon...