Forrester: The Critical Role of SharePoint Information Architecture

Executive Summary

SharePoint buyers expect intuitive navigation, contextual search, and easy administration out of the box. But such benefits depend on how content is structured, labeled, and categorized, and they require a nuanced understanding of how different audiences will navigate and search for information. The information architecture (IA) behind a SharePoint deployment has lasting consequences for the end user experience and for Web site management. Information and knowledge management (I&KM) professionals should use their SharePoint implementations as an opportunity to set solid information architecture in place that turns today's information overload into tomorrow's valuable information assets. The upshot? Information workers will finally be able to find the critical information they need to do their jobs.


For the past 10 years, information architects have worked through how to organize and present information on corporate intranets. Common best practices and design guidelines have emerged, including: 1) prioritizing directory lookups, news, and financial and HR information on the home page, and 2) offering task-driven or process-oriented navigation — such as how to orient a new employee or how to move offices — in addition to functional navigation.(see endnote 1) Organizing and controlling the information on an intranet has historically fallen to a small team of stakeholders who update the site map, scope the search engine, and design the navigation. This manual approach does not scale well to large enterprises with diverse needs.

Many enterprises unveil SharePoint to facilitate developing their intranets — better employee communication and shared access to team information.(see endnote 2) But unlike a simple intranet or collaboration solution, SharePoint also includes portal, Web content management, and business intelligence capabilities. A project plan focusing on quick deployment of SharePoint workspaces may overlook critical information classification tasks necessary to make SharePoint effective as an enterprise intranet and knowledge management vehicle.

In particular, SharePoint has some distinctive elements that affect an enterprise's information architecture, such as:

clip_image001A lack of site hierarchy. SharePoint content is stored in a SQL Server database, not in a hierarchical file server. SharePoint sites are managed in one or more "site collections." By default, the content in each Office SharePoint Server 2007 Web application lives in a single site collection and is stored together in the same database. Enterprises typically divide their content into multiple site collections due to performance, storage, and management concerns. A single site collection cannot be stored in multiple databases.(see endnote 3) Thus, the absence of a treelike site structure defies traditional navigation of content from root to leaf.

clip_image001[1]Distributed administration. Site collections can be thought of as secure containers that hold content of a similar stripe. A site collection administrator has full access to everything in the collection. Administrators can manage security, create elements such as libraries and calendars, and organize content how they see fit. This distributed model means that as SharePoint sites grow virally, IT and the business may struggle to balance control and chaos. As a result, large enterprises must decide what they should make mandatory and consistent across sites, and what they can delegate to project-, team-, or department-level administrators.

The bottom line is that SharePoint is more than just a portal server. Its wide coverage of information management tools requires a dedicated, cross-functional approach to governance.(see endnote 4) Given that these capabilities are integrated, I&KM pros have an opportunity to manage content with greater rigor and with more end user participation than has been possible before.


The primary information architecture mechanism for MOSS 2007 is the site collection framework. Microsoft describes site collections as native containers of Office SharePoint Server 2007 sites, and "the unit of ownership, quota and security management."(see endnote 5) What's determined and tracked at the site-collection level?

clip_image001[2]Application administration. Site collections affect operations like usage tracking, backup/restore abilities, storage quotas, and security boundaries.(see endnote 6)

clip_image001[3]Branding, content, and features of subsites. A site collection's Web parts, master pages and layouts, workflows, content types, and templates control the common "look and feel" and functionality of its subsites.

clip_image001[4]Navigation and search. SharePoint's navigation or site browsing structure as well as search scopes, keywords, and search "best bets" are set within site collection boundaries.

Site collections offer extensive opportunity to manage metadata at multiple layers (see Figure 1).(see endnote 7)

In short, site collections are the lynchpin of SharePoint information architecture. The way information is structured and stored affects its governance, security enforcement, disposition, accessibility and more.(see endnote 8)

Figure 1: High-Level View Of Information Architecture Of Site Collections


Add Time For Information Architecture Tasks To Your Project Plan

It's common to organize the sites by department structure and then department function (e.g., Purchasing>Contract negotiation) because existing security groups are often modeled with this hierarchy and it's familiar to end users. But information architects should make the most of their intuition about human behavior and skills in interface design, content analysis, and technical know-how to challenge that status quo as needed. Some companies create site collections based on product names, client names, or project names to offset the tradeoffs of hosting each department's content in a separate site collection.

Site collections are just one piece of SharePoint information architecture. After determining how to structure SharePoint, I&KM pros must decide how to distribute universal information to multiple roles and groups, how to harmonize local and global metadata properties, and how to implement search. To get started answering these questions:

clip_image001[5]Ask your users. Determine the boundaries of your user base: Does it include clients, partners, vendors, the whole enterprise, or a limited subset of knowledge workers? Do geographic or functional boundaries matter? Interview a sample of users to understand what content they need and how they access it today.

clip_image001[6]Analyze your content. Audit existing content stores to understand where high-value content lies and how it is organized. What content will be migrated to SharePoint, and how will you integrate what is not? How much content is duplicated? Is it templated and carefully managed throughout its life cycle? The answers to these questions will inform your decisions around content types, information management policy, and metadata fields.(see endnote 9)


A rigorous approach to information architecture in the design phase is critical to facilitating flexible information delivery and access. SharePoint administrators translate the output of the design stage (e.g., paper prototypes and wireframes) into URL namespaces via "managed paths". Depending on circumstances, they might allow a single site collection under a specific path or allow users to create multiple top-level sites under a specific path.

Other mechanisms for contextual information access and delivery are:

1. Audience targeting. Audience targeting enablesI&KM pros to define a subset of end users by certain common criteria, such as a shared project or interest in a topic. Administrators can hide or show Web parts or target any item in a SharePoint list — like a news item — to defined audiences.

2. Search configuration. MOSS 2007 search can search across site collections, crawl shared drives and Web sites outside of SharePoint, map co-workers by "social distance" and retrieve data in line-of-business applications. Further, search administrators can pick "authoritative pages" and assign best bets to popular search terms to optimize relevance.

Audience targeting and advanced search need clean, coherent metadata to run properly. Without significant commitment to taxonomy oversight, these capabilities will not work.(see endnote 10)

Rigorous IA Is The Silver Bullet For Business Content

While a rigorous approach to information architecture benefits a structured portal architecture, the benefits can also be extended to user-generated business content. Just as users struggle to find information in a portal setting, they also struggle to find relevant business content. Spreadsheets, presentations, documents, and a host of other content is generated and thrown into a sea of hard drives, file servers, and email folders, more often than not never to be seen again.

The aggregate cost of lost content can be tremendous. Applying a structured taxonomy to business data has long been one of the keys to tapping into its value. Yet the burden metadata tagging puts on users has led to disappointing adoption because most are accustomed to very lightweight storage tools like file servers (see Figure 2).

While users embrace simple file servers, finding information after the fact presents a challenge since file servers only store two pieces of descriptive metadata: file name and folder label. Frustrated by the inability to find information on file servers, organizations invested in content and knowledge management systems. These systems provided the ability for extended application of metadata to content. However, the user experience suffered (see Figure 3).

Automating the application of metadata when business content is created, rather than asking users to manually apply metadata after the fact, may be the silver bullet. By leveraging rigorous information architecture principles, users can create SharePoint sites directly in an existing portal architecture. For example, a user starts developing finance-related content by starting a workspace from within the Finance section of SharePoint. The custom site can adhere to best practice workflow and approvals, and it can inherit metadata related to finance or the specific author herself. Thus, users interact with the system much like a file system without additional metadata input into the workspace (see Figure 4).

When content and people using SharePoint are classified in multiple ways, there is unlimited potential for end users to find dynamic connections between content and people that were not preconceived by content creators. For example, teams in different regions may generate sales collateral for the same product. If that content is tagged with controlled metadata values, then a new teammate can find all existing sales-related content and expertise regardless of regional boundaries. The database structure behind SharePoint offers a hint of a future world less burdened by file formats and content storage.(see endnote 11)

Figure 2: The User Saves Content To A File Share And Classifies It By File Name And Folder


Figure 3: User Classification Required By Knowledge Management Systems Leads To Low Adoption


Figure 4: Content Is Instantiated Within SharePoint IA And Inherits Metadata



make the most of an Information ARCHITECTURE BLANK slate

Many organizations are looking at SharePoint as a foundation for better management of organizational unstructured data. SharePoint has technical capacity to organize data in compelling and usable ways. The key to success is to create a strategy that allows users to quickly access and create information that is broadly reusable within your organization. The strategy will begin with an intelligent information architecture that is reflected in your site collection plan.

clip_image001[7]Extend the benefits to user-generated business content. The same logic that applies to finding information in a portal environment can be extended to business content. Make your portal and information architecture a jumping-off point for creating business workspaces that drive best practices and inherit key metadata.

clip_image001[8]Plot the life cycle of diverse content types. Some SharePoint content is ephemeral and ad hoc; some is long-lived and essential to key business transactions. Investigate the tradeoffs of using SharePoint to manage high-value content from its creation to disposition. In particular, assess the impact on existing records management, risk and compliance, and storage procedures.

clip_image001[9]Actively curate content. SharePoint is not a hands-off, self-service system. Enterprises that intend to start off slowly with straightforward collaborative information sharing often end up with anarchy if elements like storage quotas and search scopes are not vigorously monitored by a central team. Assign appropriate resources to managing SharePoint sites and workspaces.

clip_image001[10]Consider add-ons to achieve your goals. Microsoft has embraced a partner network to augment its out-of-the-box functionality. Some enterprises buy additional tools like Autonomy's IDOL, FAST ESP, Dow Jones' Synaptica, Interse's iBox , or SchemaLogic's Enterprise Suite to compensate for SharePoint's shortcomings in search, autoclassification, and taxonomy management.


A new type of database is on the horizon, where the "data" is content

SharePoint is part of an emerging class of information management tools from diverse vendors that are structured to treat content in a way similar to how data is treated in a database. This architecture allows fundamentally more structure for managing content that is currently largely unmanaged. In the future, as content moves through the enterprise, semantic meaning will be added, like an envelope with many postmarks. Getting there will be anything but easy. Just because the tools exist doesn't mean the structure will build itself. Don't take lightly the opportunity a blank slate offers. Careful planning is required, and plans will need to adapt as new lessons are learned.


1. When thinking about portal usability, go beyond knowing users' titles and where they sit. Too often, user profiling, segmentation, and usage analysis fail to surface insights necessary for the user-centered design so critical to portal adoption. Organizations must avoid focusing on surface-level, obvious distinctions between groups of users (e.g., "Sally works in marketing and is 34 years old. She is technically savvy but has little time to spend surfing the Web."). Instead, develop user profiles based on business tasks or goals that users must achieve. See the January 12, 2006, "Too Much Portal, Not Enough Portal Strategy" report. Back to Text

2. Microsoft Office SharePoint Server (MOSS) 2007 adoption has been strong. In a survey last spring, 65% of respondents who planned to implement some part of the Microsoft Office System indicated that their timeline for implementing MOSS 2007 was either immediately or within the next six months. Source: March 2008 North American And Western European Enterprise Microsoft Office 2007 Adoption Online Survey. Back to Text

3. Microsoft discourages the use of content databases larger than 100 GB. Source: SharePoint Server TechCenter ( Back to Text

4. With the introduction of MOSS 2007, Microsoft moved SharePoint well beyond its traditional roots in portal and collaboration. SharePoint now includes broad, robust middleware capabilities. Achieving business value with SharePoint investments requires methodical strategic planning to minimize risk and maximize potential benefits. This requires information and knowledge management professionals to reach out and work with numerous other roles within the organization, and likely requires additional new headcount. See the June 24, 2008, "SharePoint Success Will Take A Village" report and see the January 9, 2009, "Governing SharePoint In The Enterprise" report. Back to Text

5. Microsoft has provided a fictional case study (Fabrikam Industries) to illustrate common principles of information architecture for Moss. Source: Information Architecture for Fabrikam Industries Intranet ( Back to Text

6. Microsoft recommends a 100 GB maximum for site collection content for optimal performance. To ease backup and restore functions, content may need to be moved into a new site collection so that it can be moved into a separate content database. Back to Text

7. An article on Microsoft's TechNet introduces Office SharePoint Server 2007 site collections, sites, and subsites, and recommends a method for recording site structure decisions. Source: SharePoint Server TechCenter ( Back to Text

8. Information and knowledge management professionals incorrectly assume that information classification is all about making information easier to find. This narrow vision ignores other critical reasons for classifying information — such as ensuring security, implementing a retention policy, and optimizing the use of storage. See the October 2, 2007, "Information Classification Must Reach Beyond Knowledge Management" report. Back to Text

9. Content types are a key building block of SharePoint site planning and document management. Content types manage the metadata and behaviors of content such as documents, list items and folders in a centralized, reusable way. Microsoft defines a "content type" as a collection of settings that can be applied to a certain content category. These settings include any associated information management policies, templates, workflows and metadata. Source: SharePoint Server TechCenter ( Back to Text

10. In an effort to eliminate information silos and increase content reuse, enterprises invest in big-ticket information management software such as search, portal, and enterprise content management (ECM) systems. To reach such an ambitious goal, it's vital to enrich the content that flows through these systems with more meaningful and structured metadata. Enriched content is simply easier to isolate, promote, find, and control. But it's challenging to incorporate taxonomy projects in the whirlwind of tools and processes to create, manage, and access content in the enterprise. Here's a guide to building taxonomies that don't sit on a shelf; they work inside key information systems to make them hum. See the November 17, 2008, "How To Build A High-Octane Taxonomy For ECM And Enterprise Search Systems" report. Back to Text

11. The discipline of cataloging in library science has always distinguished between the intellectual content of an object (i.e., the "work") and its form or realization. For example, the theme of Macbeth (the work) may be instantiated in a satire, a play, or a movie, and the relationships among these instantiations should be preserved. Back to Text