Speed Up Your Site: Web Site Optimization

Speed Up Your Site: Web Site Optimization

By Andrew B. King

New Riders Publishing, January 17, 2003

There's a time bomb on the web: user patience. It starts ticking each time someone opens one of your pages. You only have a few seconds to get compelling content onto the screen. Fail, and you can kiss your customers and profits goodbye .

You can't count on fast connections either. Most of your customers are still sucking content through a 56K straw. You have to serve up greased lightning or they'll bail. That's why you picked up this book. In it you'll learn how to cut file sizes in half. You'll trim (X)HTML, CSS, graphics, JavaScript, multimedia, and bandwidth costs. Real-world examples illustrate techniques with before and after code and percentage savings. After reading this book, you'll know how to make your pages literally "pop" onto the screen.



Users rarely agree about user interface design questions. Should the "add to shopping cart" button be red, orange, yellow, or blue? You will probably get about 25 percent in favor of each option if you ask your customers. For this, and many other reasons, you shouldn't design web sites by asking users for their preference. It is much better to learn by watchingsee what people do and not what they say. Then make the button the color that works best in user testing.

There are two exceptions to the general rule on which users disagree ; two questions where the answers are the same for virtually all users, year after year. Since my first Web usability studies in 1994, I have heard the following two comments in almost every test:

  • I don't like to read a lot of text on a computer screen.

  • I don't want to wait for slow pages to download.

This book doesn't talk much about the first problem, so let me give you a quick usability guideline: Cut your word count in half for online readers.

The second problem is where this book shines and should be mandatory reading for all web designers and Internet executives: Nobody likes slow pages, but we don't have to accept sluggish performance. Make it fast. Employ the optimization techniques you will learn in this book, and any web site will grow its base of loyal users and paying customers.

It is astounding how negatively users react to slow sites. Slow sites are difficult to navigate because users lose their sense of place and progresshuman short- term memory is fickle. Even worse , if a site is slow, it communicates contempt for customers and their time. Users will assume that additional pages will be slow as well, and that it will be painful to navigate the site. The back-button beckons.

Some web designers are in denial and think that their pages are so good that users will be willing to wait. Sometimes, this is even true: If users have specifically asked for something and know what they will be getting, they are sometimes willing to wait for higher-quality illustrations or other heavy design elements. The most common case is a user who has clicked on a thumbnail image to request a high-resolution scan. Mostly, though, faster is better.

This is the simplest equation in the entire field of Internet strategy. Many other issues can be debated hotly, and there is no easy answer. But to repeat, faster is better. This is a one-dimensional criterion where there is no doubt as to the preferred direction.

I have sometimes posed a challenge to those designers who believe that slow response times are acceptable: Artificially delay each of your page views by ten seconds for a day and see what that does to your online business. (Of course, it would be an even better experiment to speed up every page by ten seconds, but that requires you to actually read the book and redesign the pages. In contrast, it's easy to temporarily reduce performance.) So far, nobody has taken me up on this challenge. If you do, please send me the results.

There are an estimated 600 million Internet users in the world. Every year, these users waste about 6 billion hours waiting for web pages to download. This is the equivalent of about 3 million person years of full-time workor about the labor force of Denmark.

We can't totally eliminate download delays, but let's set a goal of reducing the World Wide Wait to half its current magnitude by 2005. That would gain the world economy almost $60 billion in recovered productivity. More important for you, the reader, it will increase customers' satisfaction with your web site.

Jakob Nielsen

Fremont, CA

November 2002