Introduction
In the 2009 paper, “Pimp My Website: Tech Tools to Redesign and Reinvigorate Museum Web Sites on a Budget” (Masri & Grossman, 2009), we focused on redesign tools and tricks to refresh the graphical look of museum Web sites for little to no budget. Here, we explore and provide how-to’s for inexpensive refreshes of museum Web sites using an equally important method – reorganization.
While undertaking a redesign of a site provides an obvious graphical change and often garners the most buzz among staff, institutions can also create significant visible and usability impact by undertaking a site reorganization.
Reorganizing your site’s navigation, content hierarchies and paths to access content using your existing graphic design and content can be done in-house, without a large investment of time or money, and with tremendous benefit to your institution.
Best of all, the process affords you real face-time with your Web site users and your museum’s audiences so your institution can better understand, optimize for and deliver their needs and goals.
Audience Understanding: Needs Analysis
To achieve optimal usability and organization, you must build your site around an articulated, well-understood audience.
Sounds simple enough, yet there is a wide-spread tendency for sites to lump users into overly-broad categories such as museum visitors, members, funders or researchers. A Web site audience is not a homogeneous group of people with the exact same interests or requirements, and within categories of users (such as members), there are subsets of needs typical for engagement with a site.
Thus, it is vital to ensure that your site is developed and designed in a way that is user-focused, not organization-centric. When a site is designed with your users’ needs at the forefront, you create a welcoming interface that connects immediately and encourages a deeper level of engagement with the content.
Hunters and wanderers
Since you won’t have a crystal ball to know the type of information people are looking for on each visit, it’s important to always design sites for two types of users: Hunters and Wanderers. Hunters know exactly what they are looking for and either go straight to the search box to type in their exact need, or rely on understandable navigation to drill down to information. Wanderers may have an information need in mind, but are more interested in browsing and exploring the possibilities of what a site has to offer. They like to see fascinating content called out and presented as items that may interest them, as well as to find overviews of site data to whet their appetite for more.
To handle the needs of Hunters and Wanderers, your site should emphasize redundant navigation: provide your Hunters with distinct navigational links, but give your Wanderers items of interest to pursue. We’ll delve deeper into these navigational challenges and how to optimize for all types of users in the following section on Organizational Tactics.
User needs are in a constant state of flux
A single user is not always a Hunter or a Wanderer, but may change from one to the other depending on mood, need and importance of the information. In addition to different ways of looking for content, the types of content a user needs can change at a moment’s notice, too. For example, on any given day, a museum member could come to your site searching for anything from events schedules, donation forms and trip planning tools to items from your collection.
Therefore, it’s the developer’s job to get to know the characteristics, desires, priorities and habits of a site’s users in order to create a user-centric site that caters to a diverse set of needs and goals.
The case for audience analysis
Think you can't afford to take the time or spare the budget to find out what users think? The reality is that you can’t afford to offer a site that doesn’t understand and deliver precisely what your audience expects. Usability testing can reveal poor usefulness or performance of a product, but only audience analysis shows how to fix it. The good news is that it can be done for little to no cost.
It is typical for developers feel pressure to get a site live quickly, and it is certainly true that taking the time for user analysis and site organization adds extra time to the schedule. That time can translate into salary/budget needed for audience analysis, usability testing and other planning tactics.
However, can anyone afford to launch a site that is difficult to use, in which users cannot achieve their goals? What you save at the beginning of the project will be spent many times over post-launch if you organize based on false assumptions of your audience. And the best part is that simply asking users to assess your site can add invaluable information to the process, for free.
Imagine the cost and breadth of changes you will face – not to mention the alienation of your users – if you organize your site and content around showcasing marquis pieces from your collection when what your audience really wants are better scheduling and planning tools to help them experience these assets in person.
The cost of failure in redesigning or launching a site that doesn’t meet your users’ needs is far more than the time and funds taken to understand your audience at the start of a Web project. Thankfully, audience analysis can be scaled to the size and budget of the project. It’s better to do a 3- or 5-day analysis than no analysis at all.
Organization Tactics
Organizing your Web content can be a large undertaking, but one that can be shared among your staff members. In order to properly organize this content, you not only need to understand why the content is organized as it is (audience needs), but you also need to understand how content can be organized.
However, before you embark on the why and how of content organization, you need to take stock so you know exactly what content you are dealing with. A content audit is a fairly simple, yet very important process that will get you pointed in the right direction.
Cataloguing exactly what content you will be working with will help crystallize logical content buckets. This step is also an important catalyst in the decisions about what content may be stale and ready for removal, as well as what new or augmented content may be necessary to speak to key branding messages, user expectations and site goals. Depending on the size of your site, a comprehensive audit can be a tedious task, but it is time well invested since you will walk away with a much greater understanding of your content and the scope of what needs to be organized. We’ve provide a helpful framework to follow in the Re-architecting how-to section of this paper.
Once you have a handle on the full spectrum of your site’s content, there are a variety of organizational techniques to consider - but which is the most appropriate for your project?
In order to see the differences between the various techniques, it’s best to look at the target audience (institution vs. visitor) and the organizational method (passive vs. active). When combined, these four entities create a quadrant-like approach whereby each quadrant maps to an appropriate organizational tactic.
Fig 1: The methodology matrix shows where organizational tactics fall in relation to the type of user and the activity level of the method employed
It should be noted here that the two target audiences are defined as internal institutional stakeholders (your staff and other museum decision makers) or external visitors (those who do not have an affiliation with the institution).
The organizational methods can be differentiated by understanding what is changing during the reorganization process. An “active” organization model involves a looser process that allows dynamic arrangements created by users, such as related search results, whereby a “passive” organization involves content that stays within its initial structure, such as a static Contact Us page. It is important to note that both methods can be utilized during the organizational process as well as afterwards, once the site goes live.
Types and methods
Institutional, passive
Wireframes are a fairly easy and inexpensive way to organize content by institutions in a passive manner. What this means is that the content is organized in a way that is driven by an institution’s needs or structure, and that the organization of this content changes very little during the process, if at all.
For example, an institution may decide that there is a need for an “About the Institution” section. It is highly doubtful that this content will change either during the process or once the site goes live due to the fact that this was an institutional decision to create and populate in the first place. This isn’t necessarily a bad thing – one of the upsides is that site users, both internal and external, will always be able to find specific information. Wireframing is a way to quickly lay out this information and is the most popular method of the four. It also tends to be the most traditional and easily understood method.
External visitor, passive
To organize content from a user’s perspective in a passive manner, an institution can rely on user testing and interviews. This becomes a passive method due to the fact that once users have completed their testing and/or interviews, they no longer take an active role in the organizational process. Often an institution will take this information and use it to create wireframes, again employing feedback from the testing and implementing it into the content’s structure.
This method of organization can be done for free, simply by inviting users to give their opinions on an existing site, and this often provides valuable insight into what site visitors need and want. The key is to glean the appropriate feedback and utilize it. Gathering feedback and then ignoring it during the subsequent wireframing process can undermine the rationale for gathering the feedback, so make sure that your test users are representative of your target audience.
Institutional, active
Card sorting is a relatively under-utilized but highly effective, budget friendly way to gather information in an active manner. During the process, users are given blank index cards with various section/page names and asked to organize these cards in a hierarchical manner that makes the most sense to them. This exercise can be done with institutional users or external users.
The idea is to understand how users would expect to find content and how they look at content relationships from a different viewpoint as they are actively participating in the reorganization process. As with user testing and interviews, care must be taken in gleaning what is appropriate to focus on in terms of relevancy and feedback. The best course of action is to have several users do the exercise and take notes during this activity, making sure to document how the user is arranging the information and with what thought processes. Encourage the users to verbalize what’s going on in the thought-process and why they’re choosing (or not choosing) to tie together particular content relationships.
External visitor, active
The final organizational approach that is gaining popularity is the taxonomical approach. Taxonomy is, simply put, the classification of objects, and Web site content is no different. In a taxonomically-organized site, the content is organized dynamically as the user visits the site either through real-time modification of navigational structures or by users grouping items together in order to display content in a way that works most effectively for them. In order to achieve this functionality, content is tagged with keywords or linked to other content based upon some sort of relationship that exists. Then as users navigate the site, they may find themselves forging a path that is no longer based on the site’s navigational structure, but on the items of interest to them.
Some popular examples of this include purchase suggestions akin to Amazon.com’s upselling methods (i.e., “Customers who bought this item also bought”), Drupal’s content tagging and display (CCK) module system and the “Most Popular Stories” on sites such as http://www.cnn.com. In fact, more users are browsing Web sites in this taxonomic manner (60%) than by traditional navigational methods (Straub, 2005), and often many users are entering sites not through the homepage but through a link from a Google Web search. According to Pearson Education, searching on-line is the second most popular Web activity (89%), just behind emailing, at 92%. (Pearson, 2008) If this many users are entering your site via search engine, it’s imperative to know what term or terms brought them there in the first place.
Fig 2: Examples of dynamic content relationships
Re-Architecting How-To’s
Methods for gathering user feedback
Interviewing even a handful of representatives from each audience segment (versus hundreds of users) is a great option to ensure you hear real-life critique and kudos within even the smallest budget and time constraints. In fact, noted usability expert Jakob Nielsen has shown that that best results come from testing with no more than 5 users, allowing you to conduct as many small tests as you can afford throughout the reorganization lifecycle (Nielsen, 2000) .
There are also very inexpensive user testing services (such as http://www.usertesting.com), which let you choose the demographic profile of your target audience, how many users you want to test, and what tasks you want them to perform on your site for as low as $20 a user. These reviews are delivered as audio and video recorded screencasts of actual users' visits to your site. Keep in mind that you’ll need to be prepared to handle the data and the feedback, compile results, and determine conclusions and actionable items based on that feedback.
There is a great deal written already on best practices to follow for practical usability testing (Kaufman, 2006) as well as a variety of resources published about basic usability concepts and practices. Remember that you can have very insightful results just by testing with a small group of users. You can also choose to test only a specific section of your Web site, allowing the feedback to determine whether or not you need to increase your survey size or Web areas covered.
How to create user personas and audience matrix
So, how do you start to get inside the head of your audience? As noted above, the simple act of talking to your current and prospective users and listening to their needs is a powerful and inexpensive method to gain insight into the external view of your brand and the needs of the user population.
From your audience interviews, you can then craft powerful user personas to help you visualize the goals and needs of your end-users as you develop. Personas are a set of fictional, representative user archetypes based on the behaviors, attitudes, and goals of the people interviewed in your research phase. Personas serve as stand-ins for groups of people who have common goals and characteristics, so to fully identify with your audience, create personas that have names, personalities, pictures, personal backgrounds, and, most important, goals for what they want to accomplish on your site. For more information on crafting personas, see http://www.stc.org/intercom/PDFs/2009/200901_22-26.pdf(Filippo, 2009).
Fig 3: Sample persona matrix
Armed with interview feedback, demographic details and personas, you can now create an Audience Matrix that marries the needs and goals of your users to your site’s content. This matrix allows you juxtapose audience demographics and goals with the top reasons they visit the Web site. The resulting chart will illuminate patterns in the types of content most accessed.
Fig 4: Sample audience matrix worksheet
Fig 5: Audience matrix showing the overlap of content types most accessed
This key content becomes the primary informational tool you will use as the basis for wireframe creation for the home page and other specialized pages on your site, such as forms, search results and other important pages requiring specialized layouts.
Mining user testing and user feedback
When you look at what users actually do on your current site (user stats) and overlay this data with what they say they want from your site (user interviews), you can create an excellent framework for user-centric reorganization.
Pour over your stats. There are probably 50 or so pages that are the most used. Make sure this content includes the things you want folks to be checking out, and figure out what is missing from the top 50 that should be added based on the user interviews and feedback.
After completing all of your stats review and user interview sessions, it’s time to sift through this treasure-trove of feedback and extract the most useful items. Depending on how you collected your observations, this can be a formal quantitative analysis or more qualitative in nature. Enter the stats observations and interview results into a spreadsheet, then analyze by grouping similar observations and analyzing trends. Each of these groups should be described by a short sentence, defining the problem and its impact on the user experience and the severity of the issue.
Remember the adage of the squeaky wheel that gets the oil – be careful not to get sidetracked by an individual user’s complaint or personal preference that is not corroborated by other feedback. By focusing on the group consensus, you keep your focus on the broad-reaching and impactful reorganization changes to your site.
Conducting a Content Audit
Taking inventory of your existing Web site content may be a bit time consuming, but it is a fairly straightforward task. Click through your site and record exactly what you find using a simple Excel spreadsheet. Examine everything piece by piece, as thoroughly and quickly as possible. Include a row for each page or piece of content and use columns to capture content attributes. Columns can be tailored to your specific needs, but should include, at the very minimum, Page Description or ID, URL and an indication of hierarchy. To download a sample audit worksheet, visit http://mw.beancreative.com/Site_Architecture_Worksheet.xls.
Fig 6: A sample content inventory worksheet
You can also enlist the help of an automated spidering tool; however, the resulting list of URLs may not be all that helpful if your page titles are not unique and do not explain enough to identify content.
As a time-saving tip, assign sections to various staff members and get everyone involved.
Wireframing do’s and don’ts
When wireframing there are some key things to keep in mind.
Do keep the number of top-level sections small. Remember, Web sites are hierarchical and the navigation should be as well, so pick your sections thoughtfully and utilize sub-level navigation to group content items. However, don’t be so limiting that everything falls into a couple of gigantic categories. Avoid the “kitchen drawer” syndrome.
Do use broader terms for the primary navigation, with more descriptive terms for secondary and tertiary navigation. At the same time, don’t be so esoteric in your classification that top level groupings no longer make sense.
Do spend some time on your wireframes – these will be the organizational blueprint your Web site will be built upon. Rushed wireframes can be difficult to implement if you suddenly realize you didn’t plan for certain scenarios. On the other hand, don’t spend so much time on them that stakeholders get confused and mistake certain elements for design entities.
Do focus on creating the wireframe and what it represents; however, don’t worry if you don’t have the latest, most expensive software to create them. Sketches can go just as far for a fraction of the price. Also consider some free on-line tools such as HotGloo (http://www.hotgloo.com) and Gliffy (http://www.gliffy.com).
Running a card sort
As explained earlier in the paper, a card sort is a quick and inexpensive method for finding patterns in how users expect to find content on your site. There are two ways to run a card sort exercise:
Open Card Sorting: Participants are given cards showing site content with no pre-established content/navigation groupings. They are asked to sort cards into groups that they feel are appropriate and then describe each group. Open card sorting for re-architecting and usability improvements allows participants to create the structure that is most intuitive for them.
Closed Card Sorting: Participants are given cards showing site content with an established initial set of primary content/navigation groups. Participants are asked to place cards into these pre-established primary groups. Closed card sorting is particularly useful for gaining additional feedback after an open card sort.
Once you decide which method is right for you, figure out what content you’re testing. Content for the card sort should be representative of the site (or the part of site that you are investigating) so that natural groupings (taxonomies) can be formed. For example, you might want to use a card sort to test the organization of content across your entire site, or you might want to tackle specific processes, such as guiding the museum visitor process. Make a list of each content item (such as “Museum contact information,” “Hours of operation,” “membership form,” etc.) and write it on a card (a 3 x 5 index card, a sticky note or other small piece of paper) with a simple name and short explanation, if necessary.
Fig 7: A sample card sort exercise aides in user-centric reorganization
Depending on the size of your site or the specific content being tested, expect to have about 30 to 100 cards (keep in mind that if you have too many cards it may be difficult for testers to effectively group them – consider a closed card sort in this case). Be sure to have some blank cards and a pen handy in case in case participants want to add something you didn’t include.
Now, bring in your test subjects (a cross section of the audience from your Personas and Audience Matrix) and ask them to sort the cards into groups that make sense to them, having them verbally articulate why and how they organized their groups. It is recommended to survey approximately 15 people as evidenced by the detailed analysis at http://www.useit.com/alertbox/20040719.html (Nielsen, 2004). When they’re done sorting the content, users give each group and sub-group a name that makes sense to them.
This exercise can help answer many questions that your institution needs to tackle throughout your reorganization and re-architecting phase. For example, there will likely be some areas that users disagree on regarding groupings or labels. In these cases, card sorting can help identify trends, such as:
- Do the users want to see the information grouped by subject, process, business need, or information type?
- How similar are the needs of the different user groups?
- How different are their needs?
- How many potential main categories are there?
- What should those groups be called?
Card sorting is a cheap, fast and effective information architecture tool. And because it involves user input (instead of a designer’s gut feeling), the resulting structures and patterns put you on the right path to creating intuitive structure and organization. For more information on running a card sort, see http://www.useit.com/papers/sun/cardsort.html(Nielsen, 1995).
Taxonomical approach
Working with taxonomic schemas can be a daunting task, but there are some things to keep in mind. First, true taxonomies are relationships that exist within a hierarchical structure. Often the term gets used to denote various types of structures and relationships, but the origination of the system by Carl Von Linné in the 18th century was based on parent-child relationships, so for purposes of this paper, a taxonomy must contain such relationships. For example, “car” is a child of the parent term “vehicle” – every car is a vehicle but not every vehicle is a car. Coincidentally, this allows a taxonomic approach to co-exist with a well-planned hierarchical navigational structure.
Next, one must consider the classification of their content from our two audience perspectives – the internal user and the external user. The internal (institutional) user is accustomed to utilizing the site via navigational system, but the external user is more likely to have a better experience on the site if allowed to contribute in some way, to a personal taxonomic schema. As previously mentioned, this can be presented to the user in a variety of ways, but some examples include content tagging, related items associations, and listings of content that other users viewing the same article also viewed.
Content tagging isn’t new, but it is a great way to show users with the click of a link what other pieces of site content they might be interested in based upon a subject. The Web site needs to be able to support this sort of technology, but there are plenty of free development tools and CMS systems that quickly allow site administrators to build these sorts of associations as they add and edit content. The utilization of these relationships is left up to the user; this creates a useful, yet non-intrusive, method for content browsing. One such robust and widely used system is Drupal – it is free to download, and the Drupal-community of open-source specialists boasts more than 650,000 user accounts with over 2000 developer accounts. (http://en.wikipedia.org/wiki/Drupal)
Google also does an excellent job of this with its Google Directory at http://directory.google.com. Here you can see the hierarchical relationship of data within the directories. Clicking Arts, for instance, further breaks down into a myriad of categories, including Museums, which then is broken down in further sub categories such as Art Museums. These paths are forged using a traditional breadcrumb approach; however, paths can also be created from metadata and content relationships presented in the form of search results. Either way, users are given options to assist them in finding the content that is of interest to them. This is the key component to taxonomical approaches – they are user-consumed arrangements. They may be suggested or pre-set by the institution, but ultimately the users are the ones who utilize them in the manner most efficient for them.
Fig 8: Google’s directory with dynamic content relationships and groupings
The final two things to consider when setting up the taxonomical relationships are that the vocabulary should be “expandable and understandable”. That is, a site’s taxonomic terms should be chosen wisely and be broad enough to encompass other terms and continue the relationship hierarchy, or specific enough to end the chain. The terms also need to be familiar to your users. A system of tagging is only useful when it is comprehensible. By creating an initial set of broad categories that users can comprehend, you’re ensuring that your taxonomy is working as efficiently as possible.
Conclusion
From user testing, audience matrices and content audits to card sorting, wireframing and taxonomic approaches, there is a variety of ways to create significant impact for your end-users via reorganization.
Not only is there a variety of options and combinations you can implement quickly and inexpensively, but because of their minimal time and cost investment, you can also test your implementation of the reorganization as you plan and build.
Be sure to schedule break points during your reorganization process to run your ideas past your user audience. Test early and often so that you don’t get too far down the road and commit to something that may not be ideal for your end users.
Again, keep things simple and cost-effective; you can show users everything, from a draft outline of your revised site architecture, to wireframes that function as a Web site paper prototype, and ultimately observe them beta test on your development server before launch.
Re-architecting your Web site with a user-centered approach can be done with minimal resources for maximum impact. Optimizing your site around how people can, want, or need to interact with your institution is a far more intuitive, accessible and user-friendly technique than forcing the users to access your site according to your institution’s assumptions. Best of all, this process gives you valuable face time with your audience constituents and can impact and inform all aspects of your museum’s messaging and mission, both on-line and off-line.
References
Filippo, Elizabeth (2009). The Road to Personas. Intercom. Consulted January 20, 2010. http://www.stc.org/intercom/PDFs/2009/200901_22-26.pdf
Kaufman, Joshua (2006). Practical Usability Testing. Last updated February 13, 2006. Consulted January 10, 2010. http://www.digital-web.com/articles/practical_usability_testing/
Masri, Layla and Emily Grossman (2009). “Pimp My Website: Tech Tools to Redesign and Reinvigorate Museum Web Sites on a Budget”. In D. Bearman and J. Trant (eds.). Museums and the Web 2009: Proceedings. Indianapolis: Archives & Museum Informatics, 2009. Last updated April 29, 2009. Consulted January 27, 2010. http://www.archimuse.com/mw2009/papers/masri/masri.html
Nielsen, Jacob (2004). Card Sorting: How Many Users to Test. Last updated July 19, 2004. Consulted January 13, 2010. http://www.useit.com/alertbox/20040719.html
Nielsen, Jacob (1995). Card Sorting to Discover the Users’ Model of the Information Space. Last updated May 1995. Consulted January 13, 2010. http://www.useit.com/papers/sun/cardsort.html
Nielsen, Jacob (2000). Why You Only Need to Test with 5 Users. Last updated March 19, 2000. Consulted January 18, 2010. http://www.useit.com/alertbox/20000319.html
Pearson Education (2008). Most Popular Internet Activities survey, July 2008. Consulted January 21, 2010. http://www.infoplease.com/ipa/A0921862.html
Straub, Kath, Ph.D., CUA (2005). The answer you’re searching for is… “Browse”. Consulted, January 21, 2010. http://www.humanfactors.com/downloads/jan05.asp
