Home / Blogs

A Look at How Retailers Handled the Online Rush as Black Friday Sales Hit $1 Billion Mark

This post was co-produced with Alan Dyke, Senior Software Engineer at Neustar, Web Performance Services.

Not everyone rushes to the mall on Black Friday anymore. Instead of visiting brick and mortar stores, more and more shoppers are perfectly content to go online. In fact, this year Black Friday online sales reached $1 billion for the first time. Of course, Cyber Monday is the busiest e-commerce day of them all. Sales for Cyber Monday reached $1.5 billion this year, a 30 percent increase from 2011. E-commerce continues to be the bright spot in retail as sales at brick and mortar stores slow.

Amazon has found that a 100 milliseconds delay can cause a one percent drop in revenues. With Cyber Monday generating 2.2 percent of domestic sales, Amazon like the other e-tailer behemoths, has huge incentives to squeeze the last millisecond of performance from its site.

Website providers have been moving to cloud computing so they can instantly scale to meet increased traffic. But what happens when everyone needs to scale at the same time?

Top 100 Retailers: Neustar’s Analysis

For this Thanksgiving holiday season, Neustar monitored the websites of 15 of the top 100 retailers to see how they behaved under the influx of bargain hunters. In monitoring the 15 sites, Neustar throttled our monitors to simulate the connection speeds of the typical web user, with maximum download speeds of one megabit per second and 50 milliseconds of latency. These are typical of a business DSL connection or a home connection, and much better than what most mobile users can expect. Each website was sampled every five minutes from multiple locations, starting on the Monday before Thanksgiving and through Cyber Monday. We set timeouts of 30 seconds, which are likely to exceed the patience of most web users.

The timings shown are the total page load time for all page elements. In some cases, this may include items that have little or no impact on the user experience. As a result, the raw page load times need to be read with a careful eye. For instance, the average Amazon page load time of 13 seconds may appear to indicate a very poor performance, but a closer look at the page behavior shows that the main page loads in 1.5 seconds, followed by images, most of which will be below the ‘fold’ for users, and lastly a single flash object adverted to the right margin delivered by Doubleclick.

For most of Wednesday, the object was not arriving before our 30 second cutoff. This resulted in Amazon showing a lower success rate in our table below. However, despite the imperfections, a relative change in a webpage’s load time generally indicates a poorer web experience.

Load Time: WeekLoad Time: Cyber MondayDeltaSuccess %
overstock.com4.263.9492.49%99.00%
Macy’s4.494.8106.90%99.00%
JC Penny5.425.2396.49%99.00%
Staples6.766.8100.59%99.00%
Barnes & Noble7.857.4294.52%99.00%
Target7.78.09105.06%99.00%
Dell10.8710.495.68%99.00%
Home Depot10.9111.62106.51%98.00%
Sears11.9711.6497.24%97.00%
Amazon13.0712.1492.88%85.00%
Kohls13.3612.3692.51%98.00%
Gap13.2514.9112.45%98.00%
Toys R Us13.5416.13119.13%98.00%
Walmart13.1816.56125.64%99.00%


A Closer Look at Walmart

The good news is that most sites we looked at survived Cyber Monday without incident. The single most notable feature of the analysis was the near doubling of Walmart’s load time at about 11:10pm PST on Thanksgiving Day. Analyzing the sample times shows a dramatic shift from a consistent page load time of about nine seconds to about 16 seconds.

A quick side by side comparison of the Walmart HAR file immediately preceding the change and the one after the change shows that the addition of a single JPG image could explain most of the increase in the page load time.

This image was below the ‘fold,’ so it may have had little or no effect on user satisfaction. The image, a not particularly large 800 pixels by 300 pixels, was 405 kilobytes in size. It was compressed on Tuesday morning to less than 95k, and its load time fell from eight to less than two seconds.

Walmart has a great reputation for analyzing the relationship between website performance and business value, so they would have insights into how much, if anything, this oversight cost them.

A Closer Look at Toys “R” Us

Another e-tailer worth mentioning is Toys “R” Us. Their website didn’t show any significant changes, but started to exhibit significant slowdown from about 9am EST, and suffered degraded performances throughout the day, recovering around 9pm EST. We infer that a large volume of web-traffic caused this slow down. How much revenue did Toy “R” Us lose from customers abandoning the site?

Toys “R” Us website performance on Cyber Monday.

See all the results for yourself at our interactive graphs for Cyber Monday and for the Thanksgiving Week.

By Diem Shin, Product Marketing Manager at Neustar

Filed Under

Comments

Comment Title:

  Notify me of follow-up comments

We encourage you to post comments and engage in discussions that advance this post through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can report it using the link at the end of each comment. Views expressed in the comments do not represent those of CircleID. For more information on our comment policy, see Codes of Conduct.

CircleID Newsletter The Weekly Wrap

More and more professionals are choosing to publish critical posts on CircleID from all corners of the Internet industry. If you find it hard to keep up daily, consider subscribing to our weekly digest. We will provide you a convenient summary report once a week sent directly to your inbox. It's a quick and easy read.

I make a point of reading CircleID. There is no getting around the utility of knowing what thoughtful people are thinking and saying about our industry.

VINTON CERF
Co-designer of the TCP/IP Protocols & the Architecture of the Internet

Related

Topics

Brand Protection

Sponsored byCSC

Cybersecurity

Sponsored byVerisign

DNS

Sponsored byDNIB.com

Domain Names

Sponsored byVerisign

Threat Intelligence

Sponsored byWhoisXML API

IPv4 Markets

Sponsored byIPv4.Global

New TLDs

Sponsored byRadix