Performance Tracking > 2006 Valentine's Day |
WebSitePulse 2006 Valentine's Day Retail Performance Report Summary
The major metrics for any website performance - Uptime (defined as the total number of minutes a website or a web system is available during a given period of time), and Response Time, (including the following time components: DNS resolution, time to connect, download first byte, download last byte) are in the focus of the WebSitePulse Performance Reports.
Uptime and Response Time are critical to the End User overall online shopping experience, and play an important role for the success of any web-based operation.
WebSitePulse launched the Performance and Uptime Reports program to keep the general public and the Internet community informed about the progress that is made, as well as about the challenges webmasters, owners and operators are facing on a daily basis, especially during the online shopping-intensive periods of the year. Some of the most popular e-commerce destinations related to a particular event are selected for the Reports, and data from their performance measurement is made available on the WebSitePulse public website.
Most days during the monitoring period (February 1st – February 14th) at least seven out of the ten websites included in "2006 Valentine's Day Retail Performance Report" were available with no interruptions - at 100% Uptime.
Response Times for the online Customers across the country were consistent for most of the time. Considerable deviations from the average Response Times of FTD.com were calculated from data from the US West Coast.
WebSitePulse initiated monitoring of the ten websites on February 1st collecting enough data about their performance prior to the Valentine's Day. During the first week of monitoring, most of the websites followed stable patterns of responsiveness. In the evening hours of February 7th, Diamond.com Response Times were - for the first time since February 1st - longer than the average. On the 10th – with four days remaining for order processing and shipping – NeimanMarkus.com, Macys.com and VictoriasSecret.com seemed to have higher volumes of web traffic because their Response Times were unusually long in the different parts of the day. Godiva.com had its record long response times in the early afternoon that day.
NeimanMarkus.com and Macys.com continued this pattern on the 11th, and went back to normal on the 12th (when two days were probably not enough for deliveries).
Shorter Response Times characterized almost all the online retailers on the 13th when the pressure started to build up around Hallmark.com and FTD.com.
During the first half of the day – between 8:00 AM and 1:00 PM EST – Hallmark.com was showing response times way above the average and sometimes even timed out. For a short period, the "page not found" error notification was displayed on the home page, while the IT support people were probably working on the timeout issues. The fix seemed successful, as even with slightly increased Response Times, Hallmark.com was back in action in the afternoon. FTD.com was "under fire" next – between 4:00 PM and 9:00 PM EST on the 13th. Even though the longer Response Times (6 – 10 sec.) characterized this website for the entire monitoring period, 15, 18 or even 32 sec. would discourage most of the End Users.
During the last day of monitoring, more people chose Hallmark.com for their last minute post cards than the website was able to support. First, the Response Times exceed considerably the daily averages. Then, the website stopped accepting visitors and processing orders. It was obvious that the IT people at Hallmark.com were prepared for that situation: Online Customers were kindly informed that due to the extremely high volume of web-visitors the website was not accepting orders temporarily but they were invited to return later. WebSitePulse experts believe that the problem was not resulting from insufficient bandwidth capacity but rather from the limited ability of the back-end hardware and applications to process the higher volume of orders. One of the reasons for this conclusion was the extensively large in size floral background of the message.
An interesting situation involving Brookstone.com occurred on the 10th of February: while the monitored e-business transaction was functioning without interruptions, WebSitePulse system detected a persistent error on one of the brookstone's web pages. The unusual pattern required an explanation, and two of the WebSitePulse engineers analyzed the available data. The conclusion was that the server was functioning normally, and most probably, there was a little mistake with the coding of the page. Here is a snapshot of the resulting deformation of the webpage, easily missed by the human eye, but quite "visible" to an automated testing tool such as WebSitePulse monitoring agent. It is difficult to speculate about the potential impact that (human) error have had on the e-business transactions of brookstone.com. As long as the webpage was functioning, there should not be some serious consequence. However, Visitors susceptible to webpage defacing might have opted out from the transaction.
The overall performance of the online retailers was in line with WebSitePulse expectations: more pressure on the websites where Valentine's Days gifts could be ordered 3-5 days prior February 14th, and last minute crowds at the "doors" of the websites providing instant expressions of love and affection.
Even though a multitude of factors are playing their role for the overall shopping experience, WebSitePulse strongly believes that availability and response times are among the most critical ones and they are directly correlated to the End Users' purchasing decisions and behaviors online.
The summarized numbers and averages for the entire monitoring period are shown in the tables below:
Average Response times for the whole period
Average Uptime for the whole period
Best regards, and takeITeasy™ with WebSitePulse™
Iavor Marinoff, CEO
Performance and availability are measured for a typical shopping web transaction at 15 minute monitoring interval every day between the hours of 6AM and 3AM EST. The WebSitePulse application monitoring service is used to collect data and to create the reports.
The steps in each shopping transaction - opening the home page, browsing for a product, adding product to the shopping cart, and secure check out - are comparable for all ten websites, and reflect the real online buyers' experience. The simultaneous monitoring is performed from 3 US locations of the WebSitePulse monitoring network: Chicago, IL; San Francisco, CA, and Washington, DC. In case of a detected malfunction or unavailability, the system performs automated error verification from 3 independent resources and only after the above procedure is performed and has returned consistent results, the detected error is confirmed, recorded, and included in the daily performance reports.
WebSitePulse is a leading provider of global, independent, and objective availability and performance monitoring of web sites, servers, network components, web applications, e-business- and email round-trip transactions.
For more information about this report please contact:
George T., CTO
WebSitePulse launched the Performance and Uptime Reports program to keep the general public and the Internet community informed about the performance of specific websites and the challenges their webmasters, owners and operators are facing on a daily basis, especially when the web traffic intensifies due to some special events through the year. Some of the most popular web destinations, related to each particular event, are selected for the purpose of the Reports, and data from their performance measurements is made available on the WebSitePulse public website.
With the intention to broadly disseminate this information for the benefit of the Internet community, we encourage the use of this information without prior approval and under the following Terms and Conditions:
The Data from the performance measurements cannot be changed or manipulated and should be used "as is".