Observing how well our test users did with Web-wide tasks tells us how people approach Web sites when they have no predisposition to use a specific site. This is common when users are trying something new, such as researching a purchase of something they haven't bought before, and it is exactly the situation in which a Web site needs to be its most competitive and attractive.
On average, our test users spent 1 minute and 49 seconds visiting a Web site before they decided to abandon it and move on. On the final site they visited while working on a task, they spent an average of 3 minutes and 49 seconds.
Interior pages accounted for 60 percent of the initial page views. Recognize this and support it. Don't try to force users to enter on the homepage.
When doing a task, users visited an average of 3.2 sites in addition to any search engines they might have used to find these sites. More interesting, they revisited sites an average of 0.4 times when performing a task, meaning that basically they didn't. A site has only a 12 percent probability for being revisited, so once you have lost a user, you have almost always lost them for good.
The users in our testing first went to the homepage of a site 40 percent of the time. Considering that most Web sites have thousands of pages, this means that the home-page gets disproportionately more visits. Furthermore, users often turn to the homepage when they want to get a general idea of what a site does, even if they entered on an interior page. So it's certainly a good idea to pay extra attention to homepage usability.
Let's say that you were looking for the lyrics to the song Singin' in the Rain and found them on the All Musicals site, maybe from a search engine or through a link from another site. This screen shot shows what you would see above the fold. This site doesn't encourage users to do anything else once they have read the lyricswhich in this case may have been just to check whether it was a "glorious" feeling or a "wonderful" one. The name of the site is at the top of the page, but it is presented in a way that looks like a headline, not a logo. There's no tagline to indicate the purpose of the site. There are a few links, including a money-making link to buying the CD, but these don't look clickable because they are not colored or underlined. Also, taken together, the links don't look like a navigation bar and users will tend to ignore that part of the screen because of its association with the obnoxious banner. Even worse, the title of the musical is not a link, so users have no easy way of clicking through to a list of other songs in the same musical. This site doesn't understand how to exploit deep-link visitors, so it's throwing away a lot of traffic.
Despite the importance of the homepage, however, interior pages accounted for 60 percent of the initial page views. A Web site is like a house with a thousand doors, and visitors can enter anywhere. We strongly encourage you to recognize this and support it. Some Web sites are designed to force users to enter on the homepage, but by doing so they go against a very ingrained element of the Web: the deep link.
Deep links enhance usability because they are more likely to satisfy users' needs. Generic links, such as those to a company's homepage, are less useful than links that take users to a specific article or product. So you want to encourage third-party sites and search engines to link directly to those pages on your site that address specific issues.
News.com offers several follow-ups to users who might arrive at this page via a deep linkfor example, from a blog posting about Super Bowl commercials. Below the story are three category links to lists of stories about similar topics. Let's say that you are interested in the concept of showing commercials on cell phones. You can click on "mobile/wireless" to stories about recent developments in mobile technology. Stories that are more specifically about the Super Bowl are listed to the right. Finally, there's a box with unrelated but high-traffic current headlines. (Our main suggestion would be to swap the two sets of headlines so that the list that is the most tightly connected to the current story appears closer to the body text instead of being in the right hand column, where it's more likely to be ignored.) Finally, of course, it's easy to access other features of the site by going to the homepage or using its internal search engine, both of which are represented in the expected location.
The Homepage: So Much to Say, So Little Time
As this table shows, experienced users who went first to the homepage on a Web site spent an average of ten seconds less time there than did users with less experience. This difference tells us that users get more ruthless in evaluating sites as they gain experience; they become faster at scanning pages and quicker at dismissing things they don't like.
Average Time Spent on the Homepage
Initial homepage viewings when the homepage is the first page visited on a Web site. Make your point quickly. You have very little time to make a good first impression. For time spent on subsequent visits, see the table on page 32.
Of course, it's always difficult to predict the future, but it's a good bet that this trend will continue. The more years they have been online and the more comfortable they get with judging Web sites, the less time they will spend on the homepage.
With half a minute at your disposal, all messages have to be ultra-lean and to the point. No long-winded paragraphs that users won't read anyway. Most adults can read about 200 to 300 words per minute, depending on their level of education. You might think that this allows you to present a 100-word welcome message on your homepage. Not so. Ten to twenty words are more realistic. Users will spend most of their 25 to 35 seconds figuring out where to go next, not reading word-for-word about what makes you special.
What does this company do? Even with all the text on this homepage, it doesn't say. A simple tagline or description at the top of the page would help. The strong emphasis on headlines and news overpowers the page, leaving little room to showcase the company's products and services. Also, notice the list of links at the bottom of the page. Cross-linking users to different Web sites without proper contextor warningis problematic because people aren't expecting it. If you do want to link to other sites, do so by showing their names or a clear description, not their URLs.
Dial Before You Dig seems to be a simple enough site, with little complexity on the homepage. But what does the company do? Apparently whatever it does has never been easier, but what exactly is that? With only a few seconds to communicate with new users, don't waste it telling people that your new Web service is now available. And the place to tell people how to get their password is on the login page. At any other time they won't be interested.
You might be tempted to think that users will read more later in their visit. Unfortunately, this is not so. If people ever return to the homepage, they'll spend even less time admiring your carefully honed slogans. As this table shows, users spend less and less time on the homepage with each subsequent visit. After all, the main goal of a homepage is to guide users somewhere else, and the more people understand the page, the less they will look around on it. They go straight for the navigation and click where they want to go.
Page Views by the Screenful
Gone in 30 seconds: Users spend very little time on the homepage and scroll minimally, especially on subsequent visits. For simplicity, the numbers in this table are averaged across both high-experience and low-experience users.
Only 23 percent of users scrolled the homepage during their initial visit and even fewer scrolled on subsequent visits. This is because users knowor think they knowwhere the important areas on the homepage are after one visit. Even those few users who scrolled didn't scroll very much: less than one additional screenful on average.
Interior Page Behavior
Low-experience users who entered a site through a deep link and visited an interior page first spent an average of 60 seconds there. High-experience users spent about 45 seconds on an initial interior page visit.
One of the major reasons to support deep links is because users read substantially more content on interior pages of a site than they do on the homepage.
We see the same phenomenon for interior pages as we do for homepages: With experience, people get faster at scanning their first page view and deciding what they want to do on the site. Users spent about 70 to 80 percent more time reviewing their entry point when they entered on an interior page than they did when they entered on the homepage. This is because the interior pages they visited were more directly related to their tasks.
With 45 to 60 seconds on an interior page, users could theoretically read about 200 words, but they usually spend some of this time assessing the site's navigation system and deciding where to go next. They may read as many as 100 words of initial information, however, which is substantially more than the 10 to 20 words they read when entering on the homepage. One of the major reasons to support deep links is because users read more content on the interior pages of a site.
Homepage vs. Interior Pages
People spend more time on interior pages than homepages. Again, people with more Web experience fly through screens at a faster pace while novice users tend to scrub the screen more carefully.
Say a user is researching music players and has arrived at the product page for iPods on Apple's Web site. On average, people read about the amount of text we have highlighted with a red box in this figure. In practice, users are not going to read these two paragraphs word-for-word. Instead, they will scan several of the top paragraphs, reading less than half of each of them. Users will spend the remainder of their 45 to 60 seconds looking at the photos and scanning the bulleted feature list and other page elements. In total, this page contains 523 words (including a disclaimer not shown here), which would take the average user two minutes to readmore than twice the time they are likely to spend on the page. The text is written at an eighth-grade reading level, which is our recommendation for adult users, but it will be too difficult for many teenagers, who are an important target audience for this product.
When users visited interior pages during their browsing of a site, they spent an average of only 27 seconds on each page. This extremely short time emphasizes the importance of being crystal clear on each page about what users will get out of it. People don't have time to read everything, so they will judge pages in a few seconds.
This pie chart shows where users clicked on the page. It may seem surprising that users spent more time in the content area than in areas that are usually used for navigation, such as the top of the page or the left or right columns. However, we know from eye-tracking studies that users spend the vast majority of their time looking at the content area and only rarely scan navigation areas.
Where users clicked on Web pages to navigate elsewhere on the same site, averaged across 4,719 clicks.