UPDATE: Our new site is up, and we have begun posting articles once again. Check out our blog on Montpanas.com, and read the first post in our new weekly installment of usability news.

We are happy to announce that our company, Paradyme, LLC has been renamed to Montparnas. This has been a long process which we are still bringing to fruition, with our new site, Montparnas.com set to launch on October 1st.We will continue to host the current articles on this site, but stay tuned for more articles (more regularly) on our new site, where all archives will also be available.

We look forward to continuing the dialogue in our new virtual home and will post again when it is live.

Iskold’s claim that there is an infinite demand for Google’s goods and services is an erroneous claim. Google primarily makes its money by selling advertising, and there is certainly a limit to the size of online and other advertising markets. In 2006, for example, eMarketer estimated that the size of the online advertising market in the US was roughly $15.6 billion, and Google captured about a quarter of that pie. Undoubtedly, the world wide advertising market is bigger. And as companies move toward internet marketing from traditional advertising and as Google moves into traditional advertising, the pie will get bigger. However, this is a far cry from ‘unlimited’. Google’s market is neither unlimited nor is Google the only player. In fact, as stated above, Google only capitalized on a quarter of the online advertising market–its forte.

Google is not the darling, magical money machine that Iskold and others believe it to be. It is a very well situated company in a large market that is becoming increasingly competitive and will continue getting more competitive in the future. We do live in a free market global economy (roughly), and why would anyone pass up on the opportunity to eat from the succulent advertising money pie?

Does Your Copy Hold Up To A Quick Glance?

by Jessica Neuman Beck (via GUUUI)

Jessica Neuman Beck covers key principles on ensuring that copy holds a user’s attention and at least conveys as much information in as little as a glance. The article covers visual support such as increased margins, brief paragraphs, using pictures, headlines and captions: white space is our friend. Another great reminder was to use what we know about the way people scan a web page (F Shape pattern and beyond - see our previous write-up).

Effective User Assistance Design: Ten Best Practices

by Meghashri Dalvi

This article covers a great number of excellent points about designing effective user assistance materials. This is a must-read for all as it not only provides insight on how to support users with effective materials, but also on how to create effective products by keeping the users’ needs in mind. Points include: step into your users shoes, adopt good designs, test on real users and use many, many examples in your user assistance.

Advertising versus Research: different ways to try and change your customers’ behaviour

by Paul Adams of Re-Frame.info

This article is a great study of missing the mark in advertising and product development. Knowing your audience and knowing how to communicate and cater to them effectively can be very powerful tools to achieving success.

Transitioning from User Experience to Product Management

by Jeff Lash and Chris Baum

Boxes and Arrows authors offer great points connecting Product Management and User Experience Design practices. As products and interactions continue to evolve, UX practitioners are becoming more involved in the product definition tasks and the transition to this strategic role is increasingly becoming common. Also see the second part in this series.

Another interesting finding in the study was the contrast between behaviors of participants that were role playing for the study and those that were actually inputting sensitive information. Definitely worth a read and the final paper is set to appear at the IEEE Symposium on Security and Privacy from May 20-27, 2007 in Oakland, California.

Results show that users follow a fairly uniform scan path when browsing through pictures, and a more random path while specifically searching through them.

In fact, not only does the study suggest that users’ viewing patterns depend on the nature of the web page (text-rich versus image-rich) but also by the users’ tasks (browsing versus searching).

Summary of Study and Findings

The study involved twenty participants that were asked to perform three tasks—searching for a product category, browsing product categories, and searching for a non-existent item—on the two pages on the Coleman web site.

The pages used for this study.

Gaze Patterns while Browsing an Image-Rich Page

Unlike the ‘F’ pattern, the findings from this research suggests that users tended to spread their visual attentions uniformly over the image-rich test page. Test subjects tended to be drawn to the product images equally across the visible page. One thing that did seem to hold true was that users focused significantly less on content that was below the fold and required scrolling.

A sample heat map from the study showing browing patterns of the image-rich test page.

Gaze Patterns while Searching on an Image-Rich Page

A more surprising outcome centered around users’ viewing patterns when searching for an image corresponding to a target product category. Users seemed to instantaneously identify the correct category image. Consequently, the gaze heat-map showed a very defined hotspot on the category image.

A sample heat map from the study showing browing patterns of the image-rich test page.

When test participants were asked to find a non-existent category image, their viewing patterns tended toward those associated with browsing. In fact, the gaze pattern heat map suggests that their visual attention was even more broadly and uniformly spread out over the page.

A sample heat map from the study showing viewing patterns while searching for non-existent image.

Designers Beware

Eye tracking has been all the rage over the past couple of years and while this practice can help us gain some insight, the latest study from SURL highlights the dangerous pitfalls from taking as natural law the findings from these limited and sometimes poorly designed studies. Eye tracking studies such as performed by Nielson and by the Poynter Institute produce results that are ONLY true for the test subjects and test conditions of those studies. Their findings DO NOT unequivocally translate to any web page beyond those specific ones in the studies.

The conclusions drawn from these studies by their authors or readers are often too absolute and sweeping. While they can help usability engineers and interaction designers gain some insight into users interactions, they hold little or no predictive power across any other web pages that were not included in the study. For example, it would be foolish to think that the ‘F’ viewing pattern holds for the Google home page.

Nielson himself stresses that the ‘F’ pattern is a rough viewing pattern that does not apply to all kinds of web pages and tasks. The fact that the study included 232 users only makes it statistically significant for the particular tasks and web pages in the study and for nothing else. To draw an analogy, if we conducted an eye tracking study with 500 users of the Google home page and found that users view it in a ‘G’ pattern, we could not apply those results to Amazon or Netflix and expect that users of those web sites will follow similar viewing patterns.

How users view a web page depends on a number of factors that are unique for every web page, and how users will browse or search entities on those pages is also unique. If we do not have the luxury of conducting an eye-tracking study on our own pages, we can also turn to more accessible solutions like analytics software such as Google Analytics, click pattern software such as Clickdensity, and mousing pattern software such as ClickTale. Beyond those, there are numerous other ways that we can gain insight into users’ interaction patterns with the web pages that we create, and if all else fails there is always our hard-earned intuition and common sense.

Our office ran out of blank DVDs the other day. I chose to order more DVDs online to forego the hassle of driving to the store. I went to the Office Depot web site and quickly scanned the main navigation and the middle of the screen. The main navigation had two choices that seemed to match—‘Office Supplies’ and ‘Technology’. In my mind it was really a toss-up.

Since I was late for a meeting and am rather adverse to playing guessing games, I decided to try a competitor. I typed ‘officemax.com’ into my address bar and landed on the OfficeMax home page. There I quickly scanned the main navigation and the middle of the screen. Although Office Max has similar categories—‘supplies’ and ‘technology’, many key sub-categories are exposed in the middle of the home page. Therefore, I quickly found the path that I needed—‘CD-RW, DVD Media & Storage’.

Navigation mechanisms often offer a myopic view of available paths, forcing users to take paths that may or may not bring them closer to their goals. This is not a problem if the categories and sub-categories that they contain are very intuitive for all. For example, if I wanted to buy a digital camera from the Best Buy web site, I would almost certainly know to click the ‘Cameras & Camcorders’ link in the main navigation. However, there are many cases where users do not enjoy such intuitive choices, and poorly implemented navigation forces users to choose paths which lead in uncertain directions.

Case Studies

There are a number of ways to give users farsighted visibility into the navigation structure. It is impossible to cover all of the methods, but by comparing some popular ones we can get a better understanding of their particular benefits and drawbacks.

Home Depot vs. Lowe’s

Home improvement ecommerce web sites are excellent case studies of navigation systems since they offer such a wide selection of products within numerous categories. Allowing users to quickly find items that they seek goes far beyond information architecture. In fact, it is impossible to create an information architecture that will be intuitive to all users since there are many gray areas in the logical groupings. Far more important is creating a user interface that offers users greater insight into the navigation pathways.

The Home Depot web site has a daunting set of 22 product categories in its main navigation. Some of the categories are rather intuitive while others much less so. For example, what can I expect to find in ‘Building Supplies’? I would expect a huge assortment of products to be located in that category; it has little meaning or value to me as a consequence. Not only that, many of the categories clearly overlap or are redundant. Looking at all 22 main product categories causes a great deal of uncertainty and confusion.

Further, without having a great deal of foreknowledge, I am mystified as to how the products contained in ‘Building Supplies’, ‘Electrical’, ‘Electronics’, and ‘Lighting & Fans’ differ. I am also confused by what I can expect to find in ‘Outdoor Power Equipment’ versus ‘Power Tools’.

Because the web site does not offer any visibility into those categories, the only way for me to resolve these questions is to navigate to each section and survey the products contained within them. This approach inherently introduces great potential for navigation errors as well as a great many needless clicks and wasted effort. Except for the most motivated, this is likely to be hugely aggravating for most users. Unlike Microsoft’s iron hold on MS Office users, web-based services like the Home Depot ecommerce site do not generally enjoy relatively high levels of customer loyalty since users can easily navigate to a comparable competitor such as Lowe’s.

The Lowe’s web site does not have as many top-level product navigation choices, but it does offer a mechanism that exposes sub-categories within those top-level ones. For example, when a user mouses over ‘Building Products’, corresponding sub-categories are displayed and even third-level categories.

Users can quickly see that the ‘Building Products’ category includes ‘Electrical’ items such as ‘Breakers’ and ‘Wiring Devices’. One immediately understands the nature of products found in ‘Building Products’ and can even go directly to some popular ones such as ‘Treated Lumber’. This practice not only reduces navigation error, but also reduces the number of clicks needed to get to popular product categories as far down as three levels deep.

There are some drawbacks in this case, however. In order to expose the sub-categories, users must mouse over the corresponding link in the main navigation. It would be preferable to expose the sub-categories upfront, without the need for any user action, but it is sometimes impractical to do so given space and aesthetic considerations.

OfficeMax vs. Office Depot vs. Staples

The most user-friendly way to expose deeper levels of navigation is simply to list out the categories. This gives users a better sense of the contents of top level categories and offers users shortcuts to their target destinations. The Staples web site does this very effectively by listing out the main navigation categories as well as main subcategories in a prominent place on the home page. Users do not have to point their mouse anywhere to see the sub-categories, and they have to do little scanning to locate these links.

The OfficeMax web site also does this effectively, although it places the category listings further down on the home page, requiring users to do more scanning.

Office Depot also allows users to see sub-categories within the main navigation categories, but does so less effectively by implementing a mouse-over interaction similar to Lowe’s web site. This is certainly better than not providing visibility into the main categories, but requires user interaction to do so. At the same time, this implementation frees up a lot of real estate that can be used for product promotions. An additional thing to note about this implementation is that expanding navigation menus sometimes break in users’ browsers due to version incompatibility or because users disable JavaScript.

Takeaways

Effective navigation allows users to quickly and easily find their way toward a destination without having to guess, get lost, or click numerous links. Whenever space and aesthetic limitations permit, navigation systems should both give users an insight into each pathway as well as reduce the number of steps required to get to key areas. Ensuring that these goals are met will help to deepen users’ loyalty by minimizing frustration and will make their interaction with the web site more valuable.

This painfully frustrating scenario seems outlandish, yet many websites put their users in this precise situation. Navigation systems often do not tell users where they have been or where they can go, and sometimes they don’t even let the users know where they are. Instead, users are forced to blindly stumble to their destination and through their tasks; they are forced to rely more on good fortune than on effective wayfinding.

Navigation, by definition, is finding a target point based on one’s current location and at least one more reference point – usually a visible destination or previous bearing. Anything that does not satisfy these criteria is not truly ‘navigation’ but rather un-systematic drifting. In addition, even perfect navigation is useless if the reference points used for triangulating one’s trajectory unexpectedly shift and change. This is unlikely in a physical setting since landmarks usually do not move much, but it is a real problem in a virtual space like a web site or application where the rules are dictated by human-developed code.

Therefore, effective navigation systems must do four things:

1. Tell users where they are – Users must know where they are so they have a reference point for adjoining areas or states
2. Tell users where they can go – Users cannot go somewhere if they do not know how to get there
3. Tell users where they have been – Users do not navigate in one direction; they often backtrack
4. Navigation must be consistent – The rules for navigation and reference points cannot change or they become meaningless

Meeting these criteria ensures that a user is able find their way using consistent reference points. This makes navigation through a web site or application quicker, more intuitive, and generally easier. An effective navigation mechanism facilitates the flow of traffic to key areas and functionality and, thereby, increases the value of the web site or application.

How does one create navigation systems that embody these four elements? Before discussing ways of implementing an effective navigation system, let’s first study navigation that fails to meet these criteria in order to learn how to identify navigation problems and to understand how missing these elements causes it to fail.

Examples of Bad Navigation

One of the worst perpetrators of bad navigation is none other than Google. The darling of usability engineers and interaction designers has very disparate and unintuitive global navigation that makes it difficult for users to navigate from one service or section to another. (I hope that Google takes the following evaluation as constructive criticism—yikes.) Let me take you through just one common session, where the user (me) goes to Google to find out about Picasa checks his personalized home page and Gmail.

Where Can I Go?

Probably the most annoying pitfall of navigating through the Google maze is that I have no idea where I can go to find the things that I need. Maybe I exaggerate a little but I still often feel completely lost despite being–what I would consider myself–a power user; I subscribe to at least 8 Google services and products, have been using Google since about 2001 and have even worked at Google.

For example, I recently wanted to learn about using Google’s Picasa for organizing and sharing my pictures. The task was to find out about Picasa. There are many different paths that one can take to completing this task, and I took what I think is the most common. I started on the Google home page. I scanned the minimalist page looking for a trailhead that would lead me to my destination. None of the links at the top of the page came close, I wasn’t looking for ‘Advertising Programs’ nor ‘Business Solutions’, and I didn’t care to learn ‘About Google’.

Where can I go to find out about Picasa?

From my days at Google, I luckily remembered that Google’s Services are listed under ‘About Google’. However, how would a user that is not intimately familiar with the Google Maze think to look there? Sure, the user could have searched for it by name or some related keywords, but this is meant to illustrate a very common navigation pitfall—not letting users know where they can go to complete their tasks.

Where Am I?

I navigated to the About Google section of the web site. In the upper left of the page, the logo read ‘About Google’. I located the ‘Google Services & Tools’ link in the middle of the page and navigated to the corresponding page. As I scanned over the page to get familiar with its contents, I noticed that the logo at the top left of the page now read ‘Google’ and there was a search field in the header.

What?! Where am I? How do I get back? How do I navigate around this section of the site?

Although disoriented, I assumed that I must have been in the Google Services & Tools section of the web site, so I pressed on. I found and clicked on the ‘Picasa’ link and was taken to the Picasa landing page. I scanned over the content on the Picasa landing page and found out that it was only available for Windows, so I decided to leave.

Where Have I Been?

Even though Picasa didn’t work out, I wanted to check out some other cool Google services.

But wait. Where am I now? I think that I should be in the Google Services & Tools section. How do I see the other services and tools? Do I have to navigate back to the previous page? Can I get to some other services from this page? How do I get back to the Google Services & Tools page? Do I have to hit the back button on my browser? Forget it. I just want to check the headlines on my personalized Google home page and check my Gmail. How do I get back to my personalized homepage?

I scanned the Picasa landing page and there was not a single link to the Google home page. Frustrated, I typed ‘www.google.com’ into my address bar.

What Now?

Google uses a myriad of navigation paradigms. Sometimes the link to Google’s home page is at the top of the page, sometimes it is through a linked logo, sometimes it is at the very bottom of the page, and sometime it simply does not exist. For a long time, you could not get back to Google’s home page from Gmail. Thankfully, they finally fixed it, but there are still a number of confusing and conflicting navigation systems. For example, there is still no apparent way to get back to Google’s home page from AdWords.

In addition, Gmail now lists some of the user’s services in the upper left of the page and there is a link to ‘all my services’. This is also the case on Google Docs & Spreadsheets and some other services. However, on many other services such as AdWords, AdSense, Google Analytics, etc., this is not present. In fact, on Google’s personalized home page these links also do not appear, and a totally new paradigm is introduced where the user has to follow the ‘My Account’ link.

Even more egregious is that there is an identical ‘My Account’ link in the upper right in Google Analytics that leads to a totally different page—the account page for Google Analytics. Creating similar navigation paradigms that lead to completely disparate states is certainly a cardinal sin in user experience design.

My Account link goes to main My Account Page

Above is the main My Account page

This ‘My Account’ link does not go to main Google My Account Page but rather to Analytics Account

Implementing Effective Navigation

There are no magical, universal rules for implementing the four axioms that comprise effective web navigation, but there are a few common starting points that can be modified or used for inspiration.

Telling Users Where They Are

Telling users where they are is hugely important and relatively easy to implement. On web sites that employ static navigation systems such as primary and secondary navigation, letting users know where they are is a matter of highlighting the primary section and subsection. For web sites that may have a less linear navigation pattern, the use of breadcrumbs can be a very good way of letting users know where they are as well as where they were.

Telling Users where They Can Go

Often, navigation mechanisms fail to let users know where they can go by using labels that are misleading or vague. Therefore, special attention should be paid to naming of primary navigation links. Another way to help users understand where they can go is to expose sub-sections within top-level navigation categories. This is particularly true in cases where the user’s intuition can easily fail. If the user cannot rely on her intuition, it is important to reveal deeper content. For example, if a user is looking for real estate on Craigslist.org, the user may expect to find it in either the ‘for sale’ category or under ‘housing’. However, this possible source of error and confusion is alleviated by the fact that ‘real estate for sale’ is exposed under ‘housing’.

Telling Users Where They Have Been

This point is much more difficult to implement and probably less critical than the other three. Despite this, one must be cognizant of the fact that users often backtrack, and letting users know their previous bearings greatly aids their navigation. This is particularly critical in cases where the use of dynamic methods like Ajax break the users ‘back’ button functionality.

Beyond highly interactive pages, dynamic navigation mechanisms such as filtering systems or complex, web-like information architectures require that users know from where they are coming. Navigation aids such a bread crumbs can do a fairly good job in meeting this criterion.

Systematic Navigation

There are few things that are more damaging to users than confusion and misinformation. It is critical, therefore, that navigation mechanisms are consistent and persistent. A navigation mechanism should never have an evil twin that looks the same but does something different. Also, if navigation paradigms must be different from one section to another, it is important to communicate this to the user using visual cues and explanatory copy. For the most part, creating systemic navigation is just a matter of clearly defining navigation mechanisms and common sense. If the link to the web site’s home page is in the upper left, keep it there throughout the entire site.

Takeaways

Navigation should not be confused with an infrastructure through which users move. Navigation is a systematic and controlled movement to one’s destination. Links, buttons, and other pathways that take users from page to page and state to state are just the infrastructure similar to roadways. In order for users to effectively move through the infrastructure they need navigation mechanisms that let them know where they are, where they have been, and where they are going as well as a landscape that does not shift chaotically. Trying to meet these criteria in the design and implementation of navigation systems will improve usability for users and will help web site owners meet their goals by ensuring that traffic flows to strategically critical areas.

The formula factors in navigability and speed, among other things, to give a result from 0 to 100%. Although it may seem nice and appealing to say “Oh yeah, well, my site is 100% perfect,” what exactly does this statement mean? What is the definition of perfection? The study never defines it, and I’m sure you will not find anyone who can truly define this for everyone and every website. Beyond an imprecise target, this study suffers from many other pitfalls which debunk the final formula; it is a prime example of drawing erroneous conclusions from bad experiment design, implementation, and analysis. The end result is a formula that claims to predict perfection, but falls far from this illusive goal.

The Formula

Pwebsite = { ((14.14*EaseNav) + (13.56*Speed) + (13.11*CleanDes) + (10.89*Func) + (10.89*Up)) – ((12.63*Pops) + (10.32*Ads) +(5.21*MultiM)) } / 6.26

OR (for sites dependent on security)

Pwebsite = { ((14.14*EaseNav) + (13.56*Speed) + (13.11*CleanDes) + (10.89*Func) + (10.89*Up) + (9.77*Secure)) – ((12.63*Pops) + (10.32*Ads) +(5.21*MultiM)) } / 7.24

Where:
Pwebsite = the degree of perfection of the website
EaseNav = ease of navigation
Speed = the speed at which pages load
CleanDes = clean and simple design
Func = functionality
Up = the site is always live
Secure = the overall security protections
Pops = the site tries to give you pop-ups
Ads = excessive advertising
MultiM = Flash and other multimedia

As evident above, the formula rewards websites that people perceive to be easily navigable, fast, functional, secure, reliable (responsive), and possess a clean, simple design. The other factors negatively taken into account are the use of pop-ups, excessive advertising, flash and other multimedia.

All-in-all, one would say that the formula correctly identifies many perceived positives and negatives of websites. The study was quite extensive, including 2,500+ residents of the UK and thus the formula accurately indicated trends of advantageous and disadvantageous aspects of a website. Unfortunately, the formula goes further to claim to deliver an extremely accurate prediction of a site’s ‘perfection.’ By using coefficients rounded to the nearest hundredth, the formula pretends to be highly accurate and comprehensive with small margins of variability. This arbitrary precision is problematic based on the nature of the study and the study’s analysis suffers from quite a few flaws, including:

• It is founded on qualitative and subjective questions which cannot efficiently inform such precise calculations of ‘perfection’
• There is no variance provided for the coefficients. Why are these rounded to the hundredth place?
• The formula does not account for non-linear relationships. What evidence is there that these relationships are not quadratic or logarithmic?
• The formula does not show correlations among the variables
• The formula clearly is not robust and does not account for unknown factors

Applying the Formula

By applying the formula, we can unveil another big issue with the calculations, namely that it assumes that the idea of perfection is the same for all websites. Let us examine one type of website under the microscope of this formula: Video Websites (YouTube, Revver, Blip.tv, etc.).

Not having data for the positive attributes of these sites, let us examine the “negative” aspects. According to the formula, sites are penalized for having pop-ups, excessive advertising and flash. In the case of these sites, they certainly employ pop-ups (full-screen viewing windows) and flash for viewing the content, and would thus be penalized for these features. This is to say then, that even if these sites were to be perfectly accessible, secure, functional and usable, they would never reach perfection because they employ useful features that enhance the user experience and actually make the site work. Does this seem a little off? There are actually times when multimedia integration is necessary and beneficial for a site. See discussion on appropriate use of flash within websites and accessibility tips for Flash (PDF).

When these so-called negative features are suitably integrated in a website, they should increase the level of ’site perfection’ not diminish it. These features cannot be completely written off as bad and decreasing the site’s positive reception as they can sometimes improve the user experience and overall site value. Moreover, we cannot simply add the good and subtract the bad, and claim that we have reached a perfect balance.

Other types of sites that may misrepresented by this formula include:

• Portal sites that effectively use pop-ups
• Marketing sites (due to the ambiguity of excessive advertising)
• E-commerce sites that effectively use advertising and pop-ups

The Buck Does Not Stop Here

Although this study provides great insight into some important factors for websites they do not cover them all, and certainly not for all websites. Are there not other attributes which could play a factor, such as the accessibility of the site, the customizability, the quality of the content, credibility etc.? Why is there no offset variable which can account for these factors? The exclusion of credibility and accessibility is particular astounding, considering many will not even use a website that does not encourage trust or that they cannot actually use.

Rate Your Site Quiz

Beyond the study, Rackspace offers an automated version to calculate your site’s score. The quiz asks eight questions rated on a scale of 0 to 10 to determine your score. No mention is made here of the viability of score or the subjectivity involved. To be meaningfully used, the information filled in should be based on a large enough sample size representative of the site’s user base. My major concern with the quiz would have to be question 7 which asks “Does your site contain advertising” with the extreme being “a lot.” Unfortunately, this further misconstrues the variable of “excessive advertising” as it assumes that having a lot of advertising is actually excessive (note negative connotation) when it is the way the advertisements are integrated that can cause negative ramifications.

What Does Make a Site Perfect?

Aiming for perfection is more than just meeting a few guidelines set out in general terms, and following the script of what is good and bad. There are and will be sites that score a perfect 100% on the above mentioned formula that may not be meritorious, such as one that forgoes integrating advertisements and thus must be discontinued due to lost revenue or one with great functionality (does what it says), but that doesn’t include tipping point features which give the edge to a competitor. Beyond going through a checklist of items and scoring a high score on an arbitrary scale, perhaps we should measure our sites based upon its practical value. A few questions to examine include:

• Do users have a positive experience interacting with the site? (admittedly a loaded question)
• Is the site accessible to all?
• Can users easily accomplish desired tasks?
• Are desired click-through rates achieved?
• Is the marketing effective?
• Is the content valuable?
• Is the website readable?

Asking such pointed questions and finding answers for them through investigative processes will yield much more useful information about where a website is successful and where lacking than plugging in numbers into an arbitrary formula. In the end, in-depth investigation will give us more than an abstract number to attach to our websites; it will give us direction toward improving what exists.

The screenshot (right) shows the expanded interface with category choices to the left of the player and users that liked the video to the right. This incorporates the major features into one simple, unified display of category navigation, viewing and rating and socializing through interests. Hiding the peripheral panels on start also reduces the overload of choices for the user once they first arrive on the site.

I think this is a great start to a product that should do very well and can serve as a nice model for other sites that want to let users explore and discover without having to wade through excessive screens, disorganized layouts and cluttered screens. One thing that most will notice almost immediately is the lack of advertisements. Although, this may or may not be a temporary state, the focus on the actual product and offerings in this case seem to heighten the user experience, and a non-dependence on advertisement is something for which StumbleUpon is known.

One small point which we noted is that the contrast of the text and the background is not very great. Although the contrast does meet W3C’s accessibility standards (Passed at Level 2 with contrast ratio of 5.31) according to Juicy’s Colour Contrast Analyser), the categories are a bit difficult to read particularly at varying screen brightness and angles.

Do Page Views Matter?

Page views are a useful metric for judging the volume of interaction with a web site, but comparing page views can also lead to erroneous conclusions. Let’s explore some common pitfalls.

MySpace and Inflated Page Views

In the user interaction and usability world, much has been made of the lackluster user experience of MySpace. Indeed, many technophiles and designers cannot understand why so many people actually use MySpace despite the fact that it is aesthetically unappealing, slow, and riddled with vestigial clicks. This is a fascinating enigma with complex answers, but the important thing to consider from a marketing standpoint is the vestigial clicks. It has been suggested that MySpace is in no hurry to streamline navigation because the current information architecture inflates page views. And it seems many analysts and industry watchers keep falling for this trick.

Although a brilliant marketing move on the part of MySpace, it should be a big red flag that each user is taken to a marketing splash page and is forced to navigate to their own profile page. Other social networking web sites take users directly to their profile page upon logging in.

Another important point that has to be made is that MySpace is very static while many of Yahoo’s web services are dynamic. By using AJAX based techniques, Yahoo has a number of dynamic web pages and applications in which all of the interactivity happens on one or a handful of pages. Yahoo Mail is a perfect example of this. Technically speaking, the user never leaves a single web page while reading and composing messages. On the other hand, the MySpace experience requires users to go to no fewer than three discrete pages to send a message and get back to the profile.

Declining Page Views Are Great

By making services such as Yahoo Mail more dynamic and self-contained, Yahoo made a very conscious effort to reduce page views to better the user experience and increase user lifetime. What matters is not how many page views Yahoo’s users generate but how long they stay on their web properties, how long they will be loyal to those services, and how many ads are displayed per average session as well as their CPM (cost per one thousand ad impressions).

Show Me the Money

From a business or investor perspective, the ultimate litmus test for any investment is its propensity to generate revenue. Therefore, even if MySpace generates 5% more page views than Yahoo but the CPM of corresponding page views is 50% less, Yahoo is still by far the stronger of the two. We don’t have numbers for the average CPM for current Yahoo properties, but its valuation of Facebook–a popular social networking web site similar to MySpace–estimates eCPM to be $0.42 for 2006. The total ad revenue is roughly the total number of ad impressions multiplied by their CPM, so the other piece of the puzzle is the total number of ad impressions.

Page Views vs. Ad Impressions

Page views are not commensurate to ad impressions. As stated above, multiple ads can be displayed on one page during one session, as is true for Yahoo Mail and other highly dynamic, AJAX-based web services. Even static pages often display multiple impressions for the same ad units to maximize their click-through rates. In other words, ads often rotate while a user is on one page.

Given these two facts, what really matters is the average number of ad impressions per user rather then the average page views. As the web evolves page views are becoming more and more meaningless. A much more telling metric is the average amount of time that a user spends on a web site or web service. Although, this can also be inaccurate because users can have Gmail open in the background while surfing elsewhere or doing other things. I believe the ultimate measure of interactivity is a combination of time spent on a web site, the number of clicks generated by the unique user, as well as the number of ad impressions generated in that period. (Of course, this applies to web sites that generate revenue by advertising.)

Nielsen NetRatings produced a very interesting report in May surveying the total time that users spent on popular web sites. For example, the average times spent on MySpace, Yahoo, Google, MSN, and AOL are as follows:

As is evident from the table, Yahoo had quite an edge on MySpace, MSN and especially Google.

Enough with the Hype Already

Undoubtedly, MySpace continues to grow and is here to stay as a hugely important player in the internet game. It is exciting to see who is growing and who is struggling, but let’s forget all the empty hype of page views and pull ourselves away from obsessively observing who will have the highest page views this week. The real question that we should be asking is who will find the magic formula of attracting, engaging, and retaining users that will ultimately lead to the highest levels of success.