web site optimization


Website performance news and views™

Home Sitemap Issues About Contact
Free newsletter
buy website optimization secrets book cover

New book on Website Optimization Secrets

Free Newsletter
Enter email:

home > reviews > shopping

Slow Shopping Sites Delay Santa: Scrooge Response Times

Flash: We've received sporadic reports of a frustrated Santa seen leaving the scene of some online shopping sites. We sent our action news team out to investigate. Upon encountering Santa near a popular store, we asked old St. Nick, why response times are poor? He promptly replied, "I'll never make Christmas with response times like these. Fat sites and high traffic brought the Web to its knees." As he jumped on his sleigh, and drove out of site, we thought we'd investigate the cause of his plight.

On narrowband connections, where most can relate, shopping online has become hurry up and wait. So in the spirit of the season, with not a touch of the Grinch, we bring you our first annual, Where to Shop in a Pinch.

'Twas the week before Christmas, and all through the Net,
Web executives dreamed of the perks they would get;
The graphics were hung by their logos with care,
In hopes that St. Nicholas soon would be there;

The children were nestled all snug in their beds,
While visions of Playstations danced in their heads;
And mom on her Wintel, and I on my Mac,
Had started to surf for one final knickknack.

When there on my screen 'lo what should appear?
But a broken down sleigh, and eight weary reindeer.
With a bent-over driver, so dull and so slow,
Weighed down by a burden more heavy than snow.

"What cruelty," he fumed "what heart black as coal,
Would keep me from meeting my gift-giving goal?"
I knew in a moment that it must be St. Nick,
Stuck in the quagmire of web sites not quick.

From oversized graphics to overstuffed code,
It became very clear what was slowing the load;
"Now, DASHER! now, DANCER! now, PRANCER and VIXEN!
These overslow web sites we must be a fixin'!"

As dry leaves that before the wild hurricane fly,
When they meet with an obstacle, mount to the sky,
So up to speed guidelines these shop-sites must go
To fill homes with presents and rake in the dough.

Poetry collaboration by Bob Peyser, Andy King, and either Clement Clarke Moore or Major Henry Livingston, Jr.

Speed Analysis of the Top Shopping Sites

We analyzed the home pages of Amazon.com, BarnesandNoble.com, BestBuy.com, Buy.com, Ebay.com, EddieBauer.com, JCPenny.com, LLBean.com, Overstock.com, Sears.com, Shopping.com, shopping.yahoo.com, Target.com, and Walmart.com to see how well they comply with the speed guidelines published in our book Speed Up Your Site. The home pages were analyzed using our free Web Page Analyzer plus a tool that takes into account dynamically generated content. The results of our analysis are listed in Table 1, sorted by total page size.

Table 1: Shopping Home Page Speed Analysis (file size in bytes)

Home pageTotal Page SizeHTTP RequestsHTML Size/ FilesJavaScript Size/ FilesCSS Size/ FilesImage Size/ FilesFlash SizeDownload Time* (seconds)
JCPenny.com87,76433 3,22511,526/114,654/168,359/30N/A13/36
BestBuy.com94,46061 12,284143,061/76,284/132,831/52N/A25/96
Shopping.com116,47026 6,151124,331/326,036/359,952/16N/A20/31
LLBean.com123,65029 27,05217,090/712,058/167,450/20N/A18/39
EddieBauer.com136,86719 19,35433,156/90/084,357/9N/A18/60
Amazon.com148,32856 68,3270/00/080,001/55N/A28/492
BarnesandNoble.com159,75949 65,8346,365/36,054/281,506/43N/A21/44
Ebay.com163,52168 40,35965,158/1663/145,056/4712,885/3325/70
Sears.com177,84862 66,2789,514/219,862/160,618/5721,576/143/70
Walmart.com181,89943 9,835117,762/212,253/1142,049/39N/A18/80
Overstock.com228,16247 46,08015,097/26,087/1160,898/43N/A25/60
Target.com288,87539 82,2460/00/0206,629/38N/A13/75
shopping.yahoo.com300,54543 41,560159,402/73,002/133,883/3162,698/3314/36
Buy.com316,71068 9,76110/08,654/1298,295/66N/A14/102
Average180,34745.9 35,59628,033/4.28,215/195,741/39 N/A21/60.6

*The download times were measured on a Macintosh PowerBook with a fast 56Kbps connection. The first number is the time it takes for useful content to display in seconds. The second figure is the time it takes for the entire page to load (i.e., useful content/entire page)

Note: The total page size includes dynamically generated content (ads delivered by JavaScript, etc.).

1JCPenny.com, Buy.com, Bestbuy.com, Shopping.com, and Walmart.com use HTTP compression to compress their HTML home page.

2Note that Amazon's top navigation bars load in half the initial time shown, but with no alt attribute values do not qualify as useful content.

3Ebay.com and shopping.yahoo.com use a small Flash movie to load other Flash movies, similar to GM's technique to delay loading.

Results Summary

The shopping homepages averaged 180,347 bytes in total size, requiring 45.9 HTTP requests per page. HTML file size averaged 35,596 bytes, due in part to five sites that used HTTP compression (Buy.com, Bestbuy.com, JCPenny.com, Shopping.com, and Walmart.com). JavaScript averaged 28,033 bytes in 4.2 files, while CSS contributed 8,215 bytes in 1 file. Images made up the bulk of the total payload with 39 images contributing 95,741 bytes on average. Overall, these pages averaged 21 seconds to display useful content and 60.6 seconds to load the entire page on a 56Kbps modem (see Table 1).


JCPenny.com had the smallest total page size at 85.7K (see Figure 1). It also featured optimized HTML that was compressed with HTTP compression for a lilliputian 3,225 byte HTML page. BestBuy.com came in second at 92.2K (See Figure 2). Some of the larger sites (Target, shopping.yahoo.com, Buy.com, and Walmart) had relatively fast initial load times, due in part to their minimal use of external files in the head of their HTML documents.

JCPenny.com home page - Dec. 17, 2003

Figure 1: JCPenny.com, the fastest and smallest site tested

BestBuy.com home page - Dec. 16, 2003

Figure 2: BestBuy.com, only the second site under 100K

There was a large variation in the number of external CSS and JavaScript files among the shopping sites tested. Target and Amazon used no external CSS or JavaScript files, while Ebay used a total of 17. By embedding style and behavior within their XHTML files, Target and Amazon save costly HTTP requests that delay the initial display of useful content. Target and Buy.com in particular displayed useful content relatively quickly while having slower than average total load times (75 seconds and 102 seconds respectively). This shows how an efficient layered design can overcome bulky pages to create faster initial display times.

Buy.com scored a near-perfect high-low hand, with the slowest total load time (102 seconds) yet tied for second for initial content display time (14 seconds), see Figure 3.

Buy.com home page - Dec. 16, 2003

Figure 3: Buy.com

Overall, the three biggest contributors to average page size were image size (53.1%), HTML page size (19.7%), and external JavaScript files (15.5%). The pages with slower initial load times had larger HTML files, more HTTP requests, or larger total payloads.

Delayed Flash Gratification

Two sites used GM's technique of using small Flash movies that load other movies after the page loads. Ebay.com and shopping.yahoo.com use a small nearly-empty Flash "stub" movie that loads quickly and loads rich media ads after the bulk of the page has displayed. This gives better response times while providing a rich media experience for the user. Yahoo Shopping's faster than average response times (14 and 36 seconds) show how effective this technique can be.

HTTP Compression

Five of the top fourteen shopping sites tested used HTTP compression to compress their HTML home pages. Bestbuy.com, Buy.com, JCPenny.com, Shopping.com, and Walmart.com saved 84.5% by compressing their home page HTML from 265,544 bytes to 41,256 bytes. They did not utilize HTTP compression to compress external CSS or JavaScript files, although any code included within their XHTML would be compressed. Typically gzip compression saves 80% to 85% off of HTML file size, due in part to its redundancy. See our Use HTTP Compression Speed Tweak of the Week for more details.

Graphic Intensive Experience

As you would expect, all of these sites used a number of graphics and in some cases Flash movies to attract the eye and open the wallet. However, each unique graphic requires a round-trip to the server, introducing indeterminate delays. EddieBauer.com had the fewest graphics (9) with 19 total HTTP requests, while Buy.com had the most graphics (66) but like Target and Amazon used no external JavaScript, speeding up the initial display of useful content. Ebay tied Buy.com for the most HTTP requests (68) required, due in part to their 16 external JavaScript files, 5 of which were in the head of their XHTML document.

In summary, the most common performance problems that we found were:

Suggested Solutions

To speed up initial response times we recommend combining and refining external JavaScript and CSS files referenced in the head of these HTML documents. Optimize HTML, graphics, JavaScript and CSS to minimize file size. Replace graphic rollovers with CSS2 rollovers to save space and minimize HTTP requests. Applying HTTP compression to all the remaining textual content that can be compressed (JavaScript and CSS files in the head for example) would also speed page display for most modern browsers.


In our opinion, all of these shopping sites need an express checkout lane. Four sites loaded useful content at least 33% faster than average (JCPenny.com, Target.com, shopping.yahoo.com, and Buy.com). However, all of the shopping sites we tested did not meet our response time guidelines at 56Kbps. JCPenny.com was the smallest and fastest site tested, using HTTP compression and optimized HTML. On average, the fourteen shopping sites tested averaged over a minute to fully display on a 56Kbps modem (60.6 seconds).

The art and science of web design is to balance visual appeal with fast response times. Online merchants who best optimize their user's experience, especially for narrowband connections, will reap the greatest rewards.

And I heard him exclaim, as he drove out of site; speedy shopping to all, and to all a good night.

Further Reading

The Art and Science of Web Design
An excellent guide to "intelligent content that can figure out how to display itself correctly" created from dynamic publishing systems (databases and scripted templates). By Jeffrey Veen.
Gomez' E-Tail Index
The Gomez Holiday E-Tail Performance Index shows how fast and reliable the top 37 shopping destinations are, and reveals which sites have been naughty and nice this holiday season.
Internet Merchants Struggle With Traffic
Some online retailers tested by Keynote in December failed to fully load, or complete common transactions. The New York Times, Dec. 15, 2003.
Minimize HTTP requests
By combining external files you can save costly HTTP requests and speed up your site.
Online Holiday Shopping Sites Still Straining Under the Load, Reports Keynote
The success rate for 13 online retailers dipped to 93.7% in December according to Keynote Systems, Dec. 10, 2003.
Optimizing JavaScript for Download Speed
Shows how to speed up your JavaScript downloads.
Port80 Software's Fortune 1000 Compression Survey
Port80 Software found that only 3% of the Fortune 1000 uses HTTP compression.
Reports: E-Tailers Lagged in Black Friday Stampede
"Gomez said Walmart.com ranked dead last among the Web sites it tracks in terms of Black Friday response time for purchases, with the site taking as long as 54 seconds to complete transactions." ECommerce Times, Dec. 4, 2003
Response Time: Eight Seconds, Plus or Minus Two
Summarizes current research into the psychology of delay on the Web, and offers web design guidelines. See also Flow in Web Design. By Andy King.
These Sites are a Shopper's Dream
Broadband shoppers are five times more likely to buy than their dial-up counterparts. Business Week, Nov. 25, 2003.
Use HTTP Compression
A brief introduction to content encoding. From Speed Tweak of the Week by Andy King.

About the Author

Andy King is the founder of five developer-related sites including this one, and the author of Speed Up Your Site: Web Site Optimization from New Riders Publishing. He publishes the monthly Bandwidth Report, the weekly Speed Tweak of the Week, and the semi-weekly WebReference Update.


Any trademark or tradenames used in this article are owned exclusively by their owners and they do not endorse or sponsor this site. Optimization Week, Santa, and the author are not affilliated with any of the shopping sites tested. All opinions expressed herein are based on information that we believe to be reasonably accurate at the time of publication. If you find any errors or misstatements of fact please contact us.