 |
|
|
||
 |
||||
 |
||||
 |
||||
 |
|
|
|
|
|
|
|
|
|
|
|
|
by Ron Richards, President, ResultsLab |
||||||||||
|
Link Architecture: Central issues in navigation design. |
||||||||||
|
The many roles of links, and their helpers < Below on this page |
||||||||||
|
Link strategies and examples < Best to read above first |
||||||||||
|
Roles of Links and Their Helpers |
||||||||||
|
Maybe all roses are beautiful, but as we’ll see, some links -- and the clicks they prompt -- can give your site results that are beautiful, or downright poisonous. I’d like to expose some of the variables, and issues, that a Web “landscape designer” should be careful about. Many site-architecture binds, and opportunities, arise from the basic facts -- that a link is: (1) one of the source pages exit doors, (2) the target page’s entry door, (3) a visually prominent elements on the Web page, and (4) a place where visitors make their navigating choices. A link can be A graphic that is used as a navigational link is intended to grab the eye. And if it’s an image map, it should be clear that certain areas in the map are clickable -- by virtue of its graphic prompts. Most navigational graphics fail to do either. More often than not they are designed to be cool, beautiful, or clever at the expense of strongly signaling clickability. Ironically, plain HTML links, by virtue of their color and underlining often do a better job of implying “click here”. Their navigational clarity and prompting power are based on the visitors clear knowledge that what they are seeing is a link They’ll grab even more attention if they are in a list with a line of white space between each link. These techniques helped us get 2.5 Fold Initial Gains for NetscapeWorld where I employed a double-spaced list of 9 grabber titles in ordinary HTML on the home page’s first two screens. To me, this outcome shows that even for a Web-savvy audience, grabbing power can be more important than beauty and more important than using advanced tools. NetscapeWorld readers are Web developers who “know” that top-of-home-page navigation “should” include a clickable graphic, yet even they clicked far more in response to the power of those ordinary HTML links carrying enhanced grabber language. A link can be Well designed link language offers some compelling value to the visitor -- such as learning, entertainment, surprising facts, or a response opportunity. See Grabber Development. Paradoxically, great grabbers let you know what you can expect at the other end of the link, and yet still manage to create curiosity. The key is to avoid using blind curiosity and cute wordplay language such as (in an online bookstore, promoting a book) “Mountain Valley Moments.” How much more would your curiosity be piqued if those words had read, “Mountain Valley -- the most peaceful and beautiful sanctuary in the world.” Most people will assume that clicking will get you a description of the valley, and an explanation of how it is a peaceful and beautiful sanctuary. You’ll notice that this grabber language posed an implied question and made it very clear what you’d get by clicking and yet probably made you curious about the answer. The above example demonstrates another powerful characteristic of great grabbers, one that is rarely employed in links: evocativeness. Evocative language causes the reader to have movies in their mind, and a total-body feeling of an experience. A link can be If a developer knows that certain content is important to communicate to a large fraction of visitors, it is a difficult decision whether to sprinkle that content with escape doors. Every added link offers a supporting resource to the visitor, at the possible expense of the resource they’re in. Every link on a page, especially if it’s designed (as it should be) as a grabber, is a competitor to every other stimuli on that page. In fact, the visitor is also holding in her mind, or her notes, a whole list of other places to go and things to do -- some of which are important areas in your site that she “temporarily” left. Every link a site architect adds competes with all those in-mind opportunities as well. The developer and the visitor share the goal of wanting vital, core content to not be missed. To prompt a visitor to leave vital, core content by presenting them with a digression can cost you, and the visitor, that communication opportunity forever, because what’s left is often never revisited. This double bind -- between giving easy access to supporting information, vs. keeping focus on core information -- can make Web architects more a victim to “writer’s block” than linear writers ever were. Maybe we should call it “linkers’ block.” One solution to “out of sight, out of mind” is to give visitors later links that prompt them back to the core area they may have left. That way you can find ways for the more detailed, ancillary content to repeatedly point them back to the core content. A link can be In a well-designed site the organizing themes and levels are constantly exposed -- giving the visitors clear cues to the content’s structure. Ideally, the non-link subheads and graphics explain the essence of each link-list, and distinguish the elements from each other. Great navigation architecture lets the visitor see how the elements relate to each other and to the elements that come above and below them in the hierarchy of concepts. But notice that you can do all the above in a wide variety of chunk sizes -- from very flat structures, to deeper multi-level ones. I believe such chunking decisions are among the most important in choosing site architecture. For example, in achieving 2.5 Fold Initial Gains for NetscapeWorld, one of the things we did was to remove an entire level. Formerly, the home page’s first screen showed three major table of contents pages, and listed a few key article titles in each area. The visitor clicked to one of the three table-of-contents pages, and then clicked links for each article in that area. We eliminated the intermediate level. Right on the home page’s first two screens, we encouraged readers to choose from a list of 9 grabber titles (cutting across all three areas) and to jump directly to those articles. Or the reader could scroll down right on the home page to the entire, consolidated, table of contents -- which contained headings for the three major areas, grabber article titles, and one-paragraph grabber-descriptions to sell the article. Since all the article descriptions were now on the many-screen-deep home page, visitors could print all the article descriptions in all three areas by one click on “Print.” And they could deliver the whole shebang to their colleagues and friends by sending the printout, or specifying one URL. We even put the entire long home page into e-mail letters to subscribers, to give them a complete overview of the month’s new issue all on one page. A link can be If the structure is well designed and exposed for the visitor -- with the right level of chunking -- it enables the visitor to scan and speed read. The faster they can find goodies they want, the more likely they are to stay in the site, and to explore deeply and broadly. But this is another two-edged sword since you don’t want them bypassing core content you know is vital, even prerequisite, for them. So, the trick is to develop a structure and grabbers that allow the eye to both fly, and be grabbed. One ideal outcome is for a visitor’s quick scan to convince them that the site is full of value, and to trigger a decision to spend real time and thought in the site, not just give it a quick browse. At that point, they may bookmark, or print, to help them explore extensively. A link can be Links are not just decision points. They are a place a visitor can remember and return to. They are a target in a visitor’s memory. Navigational graphics or buttons (such as at the right edge of this page) give the reader no help in remembering which of the links they haven’t followed yet. Fortunately, plain HTML links do this automatically by changing color. I believe strongly in providing several places in a content-oriented site -- places you make the visitor aware of -- where ordinary HTML links show them where they’ve been, and prompt them to visit some of the rest. One’s knowledge of color connotations should be employed in choosing the followed and unfollowed link colors for the site. A link can be We can’t currently track visitor behavior or reactions except when they click. We don’t know if they scrolled all the way down the page, whether they read thoughtfully, or whether they liked what they read. NetscapeWorld partly solved this by having a feedback area at the bottom of every article where readers could select their reaction to that article’s value, length, and technical level -- as well as give open ended responses in a form. Such data is wonderful feedback for the whole team. I found it especially valuable to study the open-ended feedback to find the reader’s most-used criteria words and concepts. One important incentive for designing in more links is that they let us track more fine-grained interest and behavior via the tracking reports. This goal should be in mind as we design and redesign architecture, especially deciding whether to chunk a big block of content down into several links. This is another tricky tradeoff. In our desire to track fine-grained interest (a truly great aid to making improvements), we could risk making the structure scattered, and default on both our own and the visitor’s goal of having appropriate guidance. |
||||||||||
|
The role of things that aren’t links Every picture or heading on a page that isn’t a link is either content that the page delivers, or an element whose purpose is to focus attention and interest on the links. Ideally, most non-link elements should do both of these things at once. The elements that give meaning to links, and direct attention to them, are as important as the links themselves. That’s because non-links can communicate claims, promises, or advice that many users will accept without drilling down for supporting evidence. For example, in a site to describe and promote an ISP’s services, a heading might say “Delivering consistent high speed and reliability even during busy hours in four ways” followed by links that say “Our big pipes,” “High power servers,” “Constant automated monitoring, “Hog prevention pricing tiers.” Each of those links (when clicked) would lead to a description of how the link language is true. For our clients, we often make link language longer than most links you see. Brevity is only good when it doesn’t cost you higher goals. And length rarely hurts if it adds grabbing power and evocativeness. With the more communicative, longer links notice that a somewhat pre-sold visitor might well not follow any of those links, having already been persuaded by the claims they communicate. Thus, a lot has been done by just using the graphic prominence of the heading and the links, but without prompting the drilling down that takes the visitor’s time and energy. This is the same principle on the basis of which all print subheads should be designed -- making it possible for many readers to bypass the body copy. Notice the profound difference between that kind of grabber-oriented structure vs. a topic-oriented structure. In a topic-oriented structure the headings might have said “Service Reliability” and offered links to “Bandwidth,” “Servers,” “Monitoring Systems” and “Pricing.” Most visitors would learn nothing from the headings, and wouldn’t click the links either. At best, they’d waste time drilling down for things the link itself could have told them. This error, and missed bet, is pervasive on the Web. |
||||||||||
|
Link Architecture: Central issue in navigation design. |
||||||||||
|
The many roles of links, and their helpers < Above on this page |
||||||||||
|
Link strategies and examples < Best to read above first |
||||||||||
|
|
|
Copyright © 1997 - 2007 ResultsLab. All rights reserved. |
|
||||||||||
|
||||||||||
|
||||||||||
