Flog: The Official Fluid Blog

The Get Elastic blog had an interesting post last week on American Eagle’s (http://www.ae.com) incorporation of product merchandising elements within the drop down layer of its product catalog navigation (http://www.getelastic.com/merchandising-in-navigation). As the Elastic Path folks point out, it’s an effective technique for surfacing more product content without a click or a separate page view. It also obscures a lot of content and can impair overall site usability.

The technique inspired a lot of discussion at Fluid. Our typically crotchety CTO had this to say:

“I find it mildly interesting, but more of a distraction than a benefit. When I use navigation, I typically know where I want to go or quickly want to scan the options. It’s like going into a Macy’s and seeing a big ad for something random on the floor index when I’m looking for shoes. Really? Do you really need to distract me there as well? Maybe I’ll forget about buying the shoes altogether…”

While our Director of Information Design countered with his own take on balancing the needs of the retailer with those of the audience:

“I think that American Eagle’s implementation could be improved, and I agree that the size of the banner distracts, but I also think there is great value in merchandising like this. It’s definitely possible to do something less obtrusive and more integrated - a more natural extension of the navigation rather than an additional banner tacked on. There is a definite business potential here for popular online retailers who can get manufacturers to pay for placement like this, with the benefit of getting their product and brand recommendations in front of more eyes. Taking it one step further the retailers, in turn, could use this incremental revenue to keep shipping costs low or free.”

And finally, from a member of the design team:

“I agree that this is really interesting from a merchandising point of view but the UI is super busy and visually just taxing. It would be interesting to know if people are actually clicking through on any of these and how it’s effecting their overall site conversion rate.”

Personally, I agree - the UI is busy and ultimately distracting. But I am also not the target consumer. That would be teenage boys and girls, for whom I suspect visual clutter and informational density is a plus. Not to mention the “Hey, I wasn’t expecting that! Cool!” factor.

As web sites and web applications become increasingly complex, they are also becoming decreasingly static. Web sites no longer present themselves as passive, static blocks of text and images. They move on their own and in response to user actions in order to:

• Capture attention (e.g., animated banner ads)
• Provide feedback after user actions (e.g., mouseover and on-click states)
• Provide deeper levels of information and facilitate understanding (e.g., infographics that illustrate complex data, such as planetary orbits)

Unfortunately, too much of this animation and motion is gratuitous and does not serve to enhance the user experience, and in many cases it actually distracts the site visitor. Usability professionals often encounter test participants who use their hands (or even sticky notes) to cover parts of the monitor where animated banner ads appear, because the ads attract their attention and distract from the content they are trying to read.

The animated GIF for banner ads that annoyed us for years before declining are being replaced by slicker Flash ads with embedded audio and video, elaborate animation, and even interactivity. Additionally, web sites and applications are using more motion and multimedia to add value to content and attract attention. In some cases the result is a sensory cacophony that overwhelms visitors and reduces the quality of the user experience.

Is animation bad? Should motion and multimedia be avoided? No, but we do need to consider when, where, and why we choose to use it. Animating a logo or image or infographic simply because we can is gratuitous. There should be value and improvement to the user experience, and the animation should support and enhance the content and goals of the site. Here are some basic principles for effective motion graphics and animation:

1. Remember: motion attracts attention. Using too much animation and motion on a single page results in competition for attention and often frustrates visitors. If you choose to use motion to draw attention, give the visitor control and the ability to stop the motion if it distracts them from their goal.
2. Animated graphics are only better than static graphics when they make it easier to understand complex information by being more visually explicit.
3. The content and format of a graphic should closely correspond to the content and format of the concepts and information to be conveyed (also known as the Congruence Principle.) For example, it is more difficult to understand the variations in the stock market looking at tabular numeric data than by looking at diagrams of value over time. Animation adds the ability to include changes in time and space in a more visually explicit way.
4. To be effective, animations need to be correctly understood by the viewer (also known as the Apprehension Principle.) Animations are often too complex or move too fast to be accurately processed and understood by visitors. Make the animation interactive, and give visitors the ability to pause, rewind, restart, and even control the speed or flow of the animation so that they can better focus their attention and thinking on the important and more complex portions.
5. Avoid clutter and unnecessary complexity. Provide enough information and visual cues to help the visitor understand, but do not include extraneous information or design elements that may confuse or distract.
6. Be organized and focused before starting to create the animation or motion graphics. Write a script, create storyboards, and have a plan to convey the information in a concise and focused way, otherwise you may wander away from the goal and include unnecessary information and/or steps.
7. Use the animation to tell a story. A coherent narrative helps visitors better understand the information in a meaningful context.
8. Support animations and motion graphics with corresponding text. Do not assume that the animations are sufficiently explicit to understand without supporting information. If you are using audio to support the animation, give the visitor the ability to control playback and volume.
9. Consider using visual metaphors to help visitors better understand complex information and concepts and to reduce ambiguity.
10. Avoid design myopia. You already understand the information and concepts, therefore your animation and motion graphics design will make sense to you. Show the animation to other people, test it with your target audience, and evaluate whether or not they understand it correctly. What seems obvious to you may be less so to others.

“That which we will call a…” Hey, just what are we going to call it?

Nomenclature, terminology, labels. What we call things and the names we use are very important, and deciding upon the “perfect label” can often be more difficult than we anticipate. As information designers and architects we are tasked with not only organizing information, but also often with naming it.

Sometimes the exercise is quite straightforward: developing a product catalog for an online toy store may be more an exercise in deciding whether building blocks are categorized as “puzzles” or ” learning toys” than in naming them “building blocks.” In either case, the label for the toy seems obvious. After all, we would not change the name from “building blocks” to “plastic construction units.”

But sometimes the nomenclature exercise is not as easy or straightforward. Sometimes we struggle with what to call something, especially when we are trying to decide upon a site’s navigation and interaction terminology, headers, subheaders, text links, and buttons. For example, should we call a section “Online Learning Center”, “Training Services”, or “Guides and Documentation”? Although each of these options conveys a similar type of content, each creates slightly different expectations about what information will be presented, the format of the information, and the extent and form of the interactions possible.

The labels and terms we use are very important, because they assist, guide, and inform visitors to our site (or users of our applications) and help them work efficiently and accurately. Poor nomenclature:

• is ambiguous, incomplete, or even erroneous.
• causes confusion.
• causes frustration.
• increases abandonment.
• decreases credibility and trust.

Good nomenclature should go unnoticed, because site visitors do not have to stop and think, “What does this mean?”, “Where will this take me?”, or “What is going to happen?”, because it is immediately obvious and meaningful to them. There are a few practices we can put in place to help us develop better labels and more useful (and usable!) nomenclature:

1. Use active language (e.g., “See the Ford Model T Specifications”) and avoid passive language (e.g., “More about the Ford Model T may be found here.”)
2. Use familiar terms that are meaningful to the target audience, and try to avoid internal, (company or product specific) terminology (e.g., “Mental Health” rather than “Psychopathology” on a community health web site.)
3. Use directive language that sets accurate expectations for the reader about what will happen and what will be presented when they follow the link or press the button (e.g., “Download the Update” communicates that the download will begin right away, whereas “Get the Update Here” is somewhat vague and may either start the download or take the visitor to another page.)
4. Be concise. Long labels are more likely to introduce ambiguity, and there is almost always limited screen real estate to work with. Tooltips that show up on mouseover are a nice way to include additional, explanatory information if necessary. Also, eliminate any unnecessary words, for example, use “Read the full article” instead of “Click here to read the full article.” There is no need to write “Click here to…” as long as the link looks like a link to the visitor. If it does not look like a link, then make it look like one - do not fix it by adding “Click here.”
5. Be consistent. The terminology, tone and voice, and grammatical structure should be parallel across the navigation system. If you are using active verbs in the navigation, then all of the navigation options should include them (e.g., “Search for Music”, “Listen to Music”, “Share Music”, and “Buy Music.”)

As information architects and information designers we spend a good deal of our time studying the organization of information. We dissect it, study it, and analyze it to discover the underlying structure. We create taxonomies, classifications, and categorizations to map out the relationships. We look for connections among disparate pieces and distinctions between similar pieces. We diagram it, visualize it, and try to give every piece a meaningful and unique name. We create navigation systems and site maps and architectures to capture and represent everything we have learned. We design page layouts, schematics, and wireframes to give everything a place to be seen. We try to put it all back together so that others can find the information more easily.

And often we fail.

Everything is findable, understandable, and logically organized. All of the pieces are there, the labels are meaningful, ambiguity is absent, and the prototype succeeded in testing, but site visitors and customers do not come. We have included everything they asked for, we have made it accessible, and we even made the connections among pieces for them, but they leave before they even get to benefit from our efforts. Our carefully crafted architecture and specially selected nomenclature languish, even though all of our test participants commented about how easy and efficient it was to complete the representative tasks we gave them.

We deconstructed the information to better understand it, but during the reconstruction process we left something out: the user experience. No matter how findable, accessible, or well-named the information is, if it is not wrapped in a desirable user experience, it will languish.

The whole is greater than the sum of the parts.

User experience is more than information design, visual design, and interaction design. We often use terms such as compelling, immersive, persuasive, and entertaining to describe a user experience, but where in our taxonomies, architectures, and wireframes do we represent these qualities? What color, shape, sound, or logo embodies them? What interface component facilitates them? Capturing and defining user experience is outside of any single design effort.

User experience is an emergent property. It arises from a wonderful convergence of good design (information, visual, and interaction), useful and usable information, and the intrinsic desires, needs, or motivations of our visitors and customers. To encourage a successful user experience (can we really ever say we create it?), we have to make the visitor or customer want what we offer, to make them feel they need it, and to help them enjoy themselves when they seek and use it.

How do we make information desirable? Not just by making it available, but by making it real. Show people how it connects with their lives, how it is useful and valuable, and how it can help them. When we reconstruct the information and make connections among the pieces, we also need to make connections between the value of the information and their lives and daily experience. We cannot stop at findability and usability. We must strive for desirability, for in the end, is it not our desires that are often the most motivating?

A common mistake when creating personas is equating one persona with one style of online behavior. Although this simplifies the creation and presentation of personas, it is not an accurate reflection of actual site visitors, their motivations, and their behaviors. Do you have the same goals and behave the same way all the time?

People behave differently under different circumstances and have different needs and goals at different times. Although it may be convenient to think of John Q. Persona as “the guy who abandons his shopping cart ten times before committing to the purchase,” this is not necessarily the way he behaves all the time.

Nearly all web site visitors will change their online behavior based on their current goals. Sometimes we have extra time to surf, and sometimes we are in rush and need to get in and get out efficiently. Sometimes we need to search extensively (e.g., I want a good suspense novel), and sometimes we already know what we are looking for and where (e.g., I want the latest Anne Rice book.) But rarely do we do the same thing every time.

Personas should be representative of actual site visitors, and that means that they will have a range of goals and behaviors. Some behaviors may be more common than others, but they will exhibit a variety of behaviors over time. Yes, John Q. Persona may be more likely to abandon shopping carts than other personas, but perhaps he only does that when comparison shopping for himself. His behavior may be different when he shops for gifts for others, or maybe his behavior changes when he browses from home versus browsing from his office during lunch.

To get the most out of personas, we should separate the personalities (i.e., the personas) from the actual behaviors. Create personas to learn about who site visitors are demographically, what motivates them, how they react to options or designs, and what content or experiences are important to them. Create experience flows to describe the most common behaviors that are observed or desired for the site (e.g., gift shopping vs. product research.) Then play mix-and-match with the personas and the experience flows: which experience flows are most likely for each persona, then rank order them. For example, John Q. Persona may exhibit more cart abandonment behavior than any other persona, but he may also be interested in comparison shopping, product research, and interactive merchandising displays.

Personas are meant to be realistic representations of actual site visitors, so it is important to remember that real visitors behave differently at different times and under different circumstances, therefore personas should also exhibit tendencies, preferences, and ranges of behavior.

A recent blog post on MSDN discussing the usability process behind the new Microsoft Office 12 “ribbon” used to present function icons in the interface described how they learned from observation that users could scan the icons more quickly when the icons are not all the same size. But not everyone thinks (based on their intuition) that different sized icons are a good idea.

I (Jensen Harris, the blog author) was reading a blog entry of someone who was kind of critical and dismissive about what we’re doing and our designs. One of his criticisms was “how bad the usability of the Ribbon would be because it’s got icons scattered all over of various sizes.” What we’ve learned is actually the opposite. People can scan disparate patterns more easily than homogenous patterns. When we use more toolbar-like layouts–a bunch of equally-spaced, equally-sized buttons, people scan them less quickly than when each chunk has a memorable layout. So we actually try explicitly to vary the layouts between chunks–it helps people find the thing they’re looking for more quickly.

This is a great example of how we are re-discovering some basic perceptual principles that have been studied psychologists for nearly 100 years: humans (and all insects and animals, for that matter) are designed to perceive change. We notice more quickly when things are different or when they change, and we get perceptually “bored” when things are all the same and never (or rarely) change. We know from software design experience that rows and rows or columns and columns of the same icon in file management UIs offer no scanning advantage, but if just one of those icons differs it seems to literally jump off the page at us. We notice the difference.

It is possible for historical reasons that function icons in toolbars are the same size because is it easier to design and build a toolbar where everything fits together like uniform bricks. It’s more difficult to create a toolbar where all of the pieces are different sizes and must be properly arranged in order to “look right.” Perhaps we have a perceived need for a grid-based system because it is simpler, not because it is better.

Yes, icons do vary in color and content based on the function, so there are differences among the icons already, but a standard icon size introduces a regularity that makes it more difficult to see those differences. Reading researchers have known for a long time that we read LONGER TEXT PASSAGES WRITTEN IN ALL CAPS MORE SLOWLY THAN WE READ PASSAGES WRITTEN WITH BOTH UPPER AND LOWER CASE LETTERS, because the ascenders and descenders provide us with word patterns based on the differences among the letters that we recognize more quickly. All caps letters are uniform rectangles, but a mixture of upper and lower case letters are a sequence of different sizes and shapes. We notice the differences.

Change and difference can occur in more than size, shape, and color; it can also occur over time. Our attention is drawn by things that move. Yes, we can see things when they are stationary, but when something moves it attracts us, it makes us want to look at it. Predatory animals are particularly sensitive to movement, but when the prey remain motionless it is much harder for the hunter to perceive it. Camouflage only works well when you are sitting still. Many animals are equipped with the coloring and instinct to blend in: fawns lay motionless in the dry grasses and stick insects cling motionless to branches. If they move, they become dinner.

We can use these perceptual principles to our advantage: facilitate scanning by using color, shape, and size cues to create memorable patterns, and draw attention with the use of differences, change, and motion. Although it may be easier to build UIs based on a grid pattern, the reality of our perceptual systems (and of the organic world) is that patterns are based on differences not similarities.

Overall, drag-n-drop is a popular interaction method, most people get it, and it is often requested. However, there is a distinction for many people between desktop environments and web sites or web applications. The “click for action” paradigm on web sites is very strong, and drag-n-drop is not yet as common on the web. We often find that we have to explicitly tell people that they may drag-n-drop items, but once they know it is available, they get it and use it properly.

The one problem with drag-n-drop (on the desktop and on the web) is that there is no uniform way to visually indicate drag-ability and inform people that the functionality is available. Buttons look clickable and underlining links is a well-known convention, but nothing really looks drag-able. People often must try to see if something is drag-able, hence the need to inform people explicitly.

Drag-n-drop in file management dialogs in a desktop environment is a good example of a learned behavior: the filenames do not really look draggable, but we have learned to drag them around and drop them where we want them. Similarly, we have learned that we can drag-n-drop appointments in our calendars as well as re-arrange our email.

The whole idea of dragging an electronic document (or image or other virutal object) composed of electromagnetic bits and putting it in a “place” relies on the metaphor of space: physical things can be found in places and can be re-arranged. We can re-arrange files in a cabinet, books on shelf, pictures in a scrapbook, or bricks in a new patio. While tangible objects have the affordance of being grasp-able, electronic objects may not.

On the web we have to create designs and provide guidance to help people draw the analogy between physical and virtual dragging and use drag-n-drop in situations where the concept of space applies to the virtual object. We need to help people transfer the skills they have learned in one situation (e.g., desktop applications or making a scrapbook) to another, new situation (e.g., placing merchandise in an online shopping cart by dragging rather than clicking.) Luckily, most of the virtual objects we drag-n-drop around do have the property of location (or space), so the drag-n-drop paradigm typically works well.

Are there instances where a virtual object lacks the analogous property of space and would therefore be confusing to drag-n-drop? Are there situations when dragging and dropping a virtual object would make little sense as an interaction method or would be more complicated than other interaction methods? For example, although it may be possible to select dates for a flight or hotel reservation using drag-n-drop calendars, it may not be as efficient or usable to abandon the drop-down lists or clickable calendars.

Drag-n-drop is an effective, easily learned, and relatively common interaction method, and as web technologies improve and grow we will have many more opporunities to incorporate drag-n-drop in our interaction designs for web sites and web applications.