UI logo
The Webzine for Interaction Designers
uidesign.net
 
     
 
  Site Search

Advanced Search
 
  Subscribe
Receive site update email alerts.
Enter your email address.
 
  Resources
Recommended Books
Links
 
  Site Info
Update Notification
Send Feedback
FAQ
Copyright/Link Policy
Review Scoring
Site Goals
About us
 
 
January 27th, 2000
  << Part 1: Requirements  
 

Site Redesign
Part 2: Content Analysis

 
     
  Introduction  
 

Now that we have some Requirements, we can go on to examine those requirements and deduce the content and required navigation elements or content relationships (aka links) which are needed at the site.

It ought to be easy to do Analysis given that the most popular content at this site is a paper describing just that - User Interface Analysis. Well, yes that may be true. However, that paper is already more than 1 year old. Times have changed and so has the state-of-the-art in Interaction Design. Not wishing to detract from my own paper, it's important to realize that it is not the only answer. My thinking has changed over the last 12 months and has been affected by newer publications. We will be using the material from my User Interface Analysis paper. However, first let's consider some more goal and task related material.

During the Requirements phase, we captured a number of Persona definitions and for each Persona we captured at least one Goal related to a visit to uidesign.net.

In a medium or larger scale web project, it would be important to expand this material. We would take each Persona and write a Usage Scenario describing how to achieve each goal. We would imagine the Interaction with the website in the broadest terms. Each of these Usage Scenarios [Fleming 99] could then be abstracted into Essential Use Cases [Constantine 99]. These techniques marry the methods of Alan Cooper [Cooper 99] and Larry Constantine and are I believe, best of breed methods at this point in time.

However, we must consider at least one of the Constraints from our Requirements session. The Constraint that the project be delivered on a low budget and in a small scale. Therefore, interesting as it may be to develop Usage Scenarios and Essential Use Cases (and I do this for a living), we will skip these procedures and take a "short cut".

[ I will return to the rigour of writing Usage Scenarios and Essential Use Cases in a future White Paper ]

The legitimate "short cut" is a simple one, justified by the strong problem domain knowledge available, i.e. I know my own site and subject matter. We are going to determine the Content needed to achieve the Goal without elaborating the procedure involved in finding that Content. Once we have determined the Content required, we will simply produce a design which makes that Content as readily and freely available as possible.

So let us firstly revisit the requirements and analyse how that may affect the content. At the same time, we are going to use the User Interface Analysis strategies from number 5 through 8 to help determine the content. [ Note the lower numbers may be useful when you are completely stuck but I believe they have become obsolete - replaced by Persona and Goal Analysis ]

 

 
  Requirements Revisited  
 

Let us examine the Requirements Document Section by Section

Owner's required functionality

The owner's requirements are that the site should be interesting to a diverse group of visitors who share some interest in the field of Interaction and User Interface Design. Such that the content will be read and improved software design will ultimately result in improved user experience with software technology. The site should facilitate the publishing of periodical material, formal "how to" papers, editorial comment - opinion or advice, feedback from readers and other useful related material such as book reviews and resources for Interaction Designers or User Interface Designers.

From Strategy 6, we clearly have several collections of data - White Papers, Editorial, Letters. In addition, Letters (feedback) are normally related to a Paper or Editorial and it makes sense to "group" these for the readers' convenience.

Strategy 6(d) challenges us to determine whether collections are named. White Papers - merely the plural of a singular White Paper. There appears to be no collective noun for White Papers and little benefit to the User in grouping them by say "1999" or "2000". [ To draw this conclusion involved skipping ahead to examine Persona Goals and then iterating back. Static Documents aren't a good medium for showing iteration in the Analysis process ]

From Strategy 7, we also realize that we have a collection over time, a history of news and "editions" of the Webzine. To maintain the integrity of the published material, it makes sense to allow the reader to view these older editions providing suitable hints are given to orient the reader that they are reading older material. The previous site design did this very poorly.

Owner's preferences

The site should teach by example. It should demonstrate many "best of breed" examples of good Interaction Design including good navigation, good usability, ease of use, good visual communication. It should generate interest and it should promote itself and the Owner (this is the hidden agenda - there is always at least one in the requirements and it's important to find it, to know it and to understand it). It should also encourage more frequent return visits to the site.

This gives us a simple but challenging design requirement. The site must create a "sense of immediacy" and must clearly communicate most recent updates and when they happened. The Content requirement is an abstract one but implies time stamps and a group or collection of Most Recent Updates.

It should also be fast to load and make effective use of lightweight graphics.

This is not a Content requirement but a design and implementation preference.

 

 
 
Related Articles
Most Recent
Most Popular
 
  Site Search

Advanced Search
 
Content to deliver Goals
 

We can validate our Persona definitions by taking a look at the site logs, the referral information and the pages being hit most often.

Reviewing the Logs

In order to do this review, we must consider Page Requests, rather than hits. Hits tell us very little as one is generated for each graphic. However, hits do give us information about the "bale out rate" i.e. how many pages ceased requesting before all graphics were downloaded. However, interesting as that is, it's not relevant to our Content Analysis.

The first thing we discover is that the number of pages viewed against unique URLs is quite high. The best guess is that each visitor views on average 5 pages per visit. So at least 80% of referrals should be internal. As it turns out this is the case.

Next we can look at those external referrals from other websites and search engines. This is the remarkable statistic. Only 5% are from external sources. The leader, by far, is google, followed by AltaVista. This tells us that other search engines aren't much good for this technically specialist field. It also tells us that the Users are most likely to know about technically superior products such as google.

Other leading referrals come from foruse.com and the Swing book site of Manning Publications.

The key conclusion to be drawn is that 75% of the readership are relatively regular readers who have bookmarked the site or have memorized the URL.

Of the remaining 25%, 10% are coming in through Search engines and perhaps discovering the site for the first time. The other 15% are coming in from referrals, linked from other sites or through usenet referrals.

The other interesting data to come from the log is the page people first view. The homepage is naturally the most popular but the White Papers Index, the Book Review Index and the Links page all act as portals. These four pages account for more than 80% of "entrance" pages. In the parlance of David Siegel, uidesign.net has 4 "core" pages. The homepage accounting for 50%, the links page 25% and the other two splitting the rest.

So now compare the site log analysis against our Personas to validate them. We know that 10% of first visits are from Search engines and that 10% of visitors hit a Content page and not a Core page. This "type" of visitor fits the profile of Diane (from below).

We know that around 20% of people hit the links page on first entry. This fits the Gary profile.

We know that many many readers are regular or periodical returners who come straight in from a bookmark or typed URL, hitting the home page. This fits the Graham profile.

We also know that around 5% of visitors come in through usenet or from other Usability, Interaction or HCI websites. This fits the Walter profile.

So, our Web Logs have told us that our Persona definitions are within a reasonable band. They can't tell us for sure that the Personas are accurate but they could have told us whether the Personas were wrong. Simply, the usage profiles from the logs would not have matched the implied usage profiles based on the Goals of the Personas.

Turning Goals in to Content

Walter, an HF/HCI post-grad student

Walter has only worked in industry for one summer internship. He didn't get any real work to do or it was only hacking C++ code for some back end interface. He is currently doing his literature review and has turned to usenet and deja.com for suggestions. He found uidesign.net through a link from my tagline.

Walter needs to find material that can be classified as "State of the Art" and he needs to know if he's missing anything. He finds the book reviews (or at least some of them) useful. He browses the White Papers and cherry picks those with novel Interaction material.

Walter arrives on a Content page such as Book Review from a usenet link. His goal is to determine whether there is any "State of the Art" material on the site. Firstly, he will view the Content on that page - is it relevant? If the answer is yes, he might stay. Walter needs to be oriented quickly. He needs to know where he is and what might be available which is "State of the Art". He needs access to a list of Papers and a list of Book Reviews and it would be helpful if those came with a brief summary His background in HCI should be sufficient to help him. He will be looking for references to certain authors or use of certain terms which he knows to be "current".

If he can't find it from a quick browse, he may prefer to Search for it. He wants to be thorough and knows that he must check every article for possible material. To be able to Search the Site using keywords from his current research will be very useful.

Diane, an Undergrad CS student with a term paper on HCI to submit by Friday

Diane is having a lousy year. Her boyfriend is cheating on her, her allowance is almost done for this month, her roommates are real messy and her pet cat got sick. She works evenings to supplement her allowance and parties when she is not working. She doesn't have time to do homework. Term papers have to be thrown together over lunch, just before their due date.

Diane hits uidesign.net from AltaVista search engine. She arrives on a book review but really needs a White Paper to plagiarise. She needs to see a list of available papers and get a quick grasp of the content. She selects a paper, saves it to disk and starts editing.

Diane arrives on a Content page such as White Paper - maybe even a middle page of a multi-part paper. She came in through a Search engine. Her goal is to find something which looks slightly obscure, but still cool. The paper she's on may not be it. She quickly examines the Content. Is it relevant? Is it vaguely suitable? What else might be available? Anything related? What other Papers are here?

Like Walter, Diane needs to be oriented quickly. She needs access to a list of Papers with a brief summary.

Gary, Windows GUI Developer in a big Insurance Firm's IT Division

Gary stands out from his immediate colleagues as a visionary. Someone who cares a little bit more than simply when his paycheck will arrive. Gary wants to do a better job. He knows his boss is related to Dilbert's boss and the only way the department's output will improve is if he makes it happen in his own time. He reads a few web sites periodically and tries to improve his knowledge. He doesn't have time to read books but he does own a few weighty tombs on Windows APIs which he references when required. Currently, Gary doesn't know much about UML but he is keen to learn that too. It might just be his passport to a better position elsewhere. Gary found uidesign.net from usenet. Gary bookmarked the links page. He passes through occasionally looking for other sites.

Gary needs to read stuff which makes his job easier. He wants advice on menus, grouping fields, code construction, notification mechanisms, advice on when to use a widget and what alternatives there might be. He needs to know stuff about APIs and stuff about UML and software development processes. He comes in by the links page, occasionally he might stay but more often he just wants to choose a link and surf off somewhere else.

Gary comes by regularly. He needs to see "What's new!" in the links page. He needs the "diff" from last time. This can be facilitated by a "New" link tag and by the "Last updated" date on the page. Hopefully, he remembers when he last came by.

However, Gary's Goal may not lie with an external link, there may be something cool and helpful at uidesign.net since he last visited. The old design didn't help Gary with this, so the new design should help to orient him and should give him some idea of what's new or recent at the site.

Graham, Program Manager at esiteforyou.com

Graham graduated in Art and Communication. His hobby interest in Graphic Design and his Apple Mac experience, got him a job as a Graphic Artist at esiteforyou.com while it was still a small startup called Design Partners. The sudden explosion in eCommerce and the change in the business name saw huge growth and VC investors. Graham is now head of a whole eCommerce web team. He still knows a lot about Graphic Design but is out of his depth in many other areas of design and software development. He found uidesign.net from a Search Engine and bookmarked it. He visits occasionally as he does several other sites. His interest is in improved usability and better design.

Graham likes to see what's new. If something interests him then he may choose to read it, otherwise he doesn't want his time wasted and will move on to something else somewhere else.

Although listed last, Graham is our most important Persona. Our logs have shown that Graham represents the majority of visitors. The regular (perhaps loyal) front page visitor.

Graham comes regularly. We owe it to Graham to help him achieve his goals. He needs to know "What's New" perhaps with some summary or introduction to each of the new articles. We don't want to waste his time. If the current material isn't relevant we want to tell him upfront. So he goes away but he'll come back again.

We don't want Graham to have to click links just to find out what's beyond. The White Papers are long and heavy to download. We only want Graham to click that link when he's sure it's right for him.

That way he'll be happy to wait while all the graphics load. He knows that what is coming is of interest and is worth a 30 or even 60 second wait.

 

 
 
Summary
 
 

From a process of two techniques, Interface Analysis Strategies and Goal Delivery Analysis, we have determined considerably more detail for our Content and Navigation. We have a clear idea of what needs to be made available in order to Contain the Content at the site and what Navigation we need to make available in order to best facilitate finding the material to deliver the goal.

In addition to delivering the User's Goals, we can also facilitate the Site Owner's Goals.

The first one is easy. Deliver relevant content to those who need it. This is totally compatible with the User's Goals.

The second to turn surfers and occasional readers into regular readers can be achieved through better orientation on arrival. A better sense of "where they are". This has to be coupled to a good initial impression and an instant degree of trust.

 

In Part 3 of Site Redesign, we will look at Design - both graphic design and interaction design. We will look specifically at how to create a good initial impression and how to create and build that all important trust.

We will also look at how we can deliver the Content Analysis Model as an Interaction Design. Taking a look at the Content and Navigation required to deliver a solution for each Persona and accommodate the Data Groupings, Collections and Histories deduced from our Interface Analysis.

 

Do you have any comments about this paper? Do you have a different Analysis approach? Are there conclusions with which you don't agree? Please send feedback to mailto:david@uidesign.net

 

References:

[Constantine 99] Software for Use, Addison Wesley, Constantine & Lockwood, 1999

[Fleming 99] Web Navigation, Jennifer Fleming, O'Reilly, 1999

[Cooper 99] The Inmates are Running the Asylum, Alan Cooper, Sams, 1999

 

 
  << Part 1: Requirements    
  Comment on this article...  
   
Most Recent
Related Articles
uidesign.net
hosted by likk.net
           
 
Copyright uidesign.net, 1999 - 2003.
The UI logo device and uidesign.net wordmark are trademarks of uidesign.net