Findability Primer by Information Architected - the IA Primer Series
7 likes5,991 views
The document discusses the importance of findability in the digital age. It defines findability as "the art and science of making content findable" and distinguishes it from simple search functions. Findability utilizes various technologies and techniques to help users efficiently locate relevant information among large volumes of digital content. These include tagging, taxonomies, semantic search, and natural language processing. The document provides an overview of different findability component technologies and their applications.
Findability Primer by Information Architected - the IA Primer Series
- 1. IA Primer Findability The Art and Science of Making Content Findable Information Architected Phone 617-933-9655 http://www.informationarchitected.com
- 2. IA Primer: Findability Why Findability is Critical Today Content without access is worthless. With the advent and maturity of the Internet, what was once exclusively the domain of libraries and the private collections of enterprises is now a broadly understood issue. Case in point: Moments ago, I entered the word “Findability” into an online search tool that indexes the Internet. More than 543,000 individual bodies of content were retrieved. Eureka – Findability solved, right? With a simple search, I am able to retrieve “all” of that content. No. The rules of the game have changed significantly. Different approaches and models for organizing and accessing content (e.g., file-shares and simple word-based query tools) are dated and no longer adequate. Providing access to content in a shared repository with a loosely defined naming convention all too often results in smaller sub-collections that are no more easy to navigate and retrieve content from with any degree of effectiveness and timeliness. Applying simple search to the problem does not make the situation much better. There are many reasons for this, all discussed in this report, but the simplest to point to as means of introduction to the issues is the scenario introduced above : A simple search for “Findability” on the Internet results in 543,000 hits found. While the search may narrow down the overall collection, the sheer volume of content renders this naïve form of selection useless. It does not provide a level of granularity low enough to meet my particular needs and perspectives. While the example above is based on the World Wide Web, the situation within large organizations is not much better. The volume of content in any single organization is hardly near Internet-scale, but nonetheless large enough to overwhelm the user in a similar manner. It is irrelevant whether retrieval results in 543,000 documents found, or 543 found. The point is that both scenarios represent a situation that is unmanageable by the user. In addition, the ability to create content within the enterprise in myriad formats (e.g., e-mail, video, instant messaging, blogs, wikis, PDFs, word processing files, scanned images) far outpaces our ability to effectively find that content. Information Architected Phone 617-933-9655 http://www.informationarchitected.com Page 2
- 3. IA Primer: Findability That said, it must be appreciated that the challenge of findability is twofold. First, users need the ability to locate all relevant content. This is no small feat. Finding content digitally is possible only if pointers to content or the content itself is in native digital format, made available for indexing by search, and/or accessible by information organization and access techniques. The lack of such functionality in the enterprise is at the heart of user frustration. The second element to Findability dictates that relevant content be presented in a manner that is conducive to the information seeker and the business at hand. Given the volume of online content that exists within enterprises today, it is no longer enough to simply retrieve all content in response to a simple query. The amount of time knowledge workers spend reviewing “irrelevant material” is growing at a phenomenal pace because the amount of digital content being stored is also growing at a phenomenal pace. Effective Findability retrieves content in context. Therein lies the crux of Findability. It cannot be attained simply by search – even a powerful search. Findability provides intuitive interaction between the user and the content. It provides multiple ways at getting to content, each tailored to a specific type of retrieval need, which includes necessary controls over security. Defining Findability To achieve Findability levels that equal or exceed those on the Internet, organizations need to address Findability directly and strategically. But most enterprise users struggle and become frustrated with in-house search and access tools that are far cries from the capabilities found on popular websites. Far too many organizations do not have an ap- preciation for the difference between applying a search engine to an online collection of content, versus applying “findability”, via a strategically designed and deployed informa- tion architecture. This lack of distinction between standalone search and Findability leaves many to believe that if they have a search tool (or in many cases several search tools), then Findability is being adequately addressed. The reality is, however, that applying search tools in an ad hoc manner often does not result in effective Findability which in turn often leads to a belief that the ineffectiveness of Enterprise Findability is the fault of a poor search engine. In reality, the search tool may be more than adequate for what it has been designed to do, but the design behind its deployment is flawed. Information Architected Phone 617-933-9655 http://www.informationarchitected.com Page 3
- 4. IA Primer: Findability Difficulties with effectively cataloging and retrieving unstructured information are eternal problems. Not only has this been a pressing issue for generations, but there is evi- dence to suggest that we will always seek better, more efficient approaches toward the ultimate goal of perfect recall and keen precision. So what is Findability, if not simply search? Succinctly put, Findability is the art and sci- ence of making content findable. The science referred to in this definition is library science, at the heart of which lie the issues of precision and recall. (See the callout at the end of this report section for more detail.) Effective Findability should provide reliable and complete retrieval of content based on user need (recall) while simultaneously returning only that content (precision), thus eliminating the review of irrelevant content by the user. This facet of Findability potentially utilizes tools and technologies such as tagging, tax- onomies, and indexing. The art referred to in the definition is really two arts: language arts and the art of user interface design. A Findability strategy should include a level of understanding of the language of the con- tent (this can be multiple written languages as well as multimedia “languages”). In this Information Architected Phone 617-933-9655 http://www.informationarchitected.com Page 4
- 5. IA Primer: Findability manner, the content can be analyzed and a level of sensitivity to the content is devel- oped that can provide intuitive, instinctive, and informed decisions regarding how and if content is related to a user’s focus – far beyond an indexed search or word-based search. Findability technologies can include any number of linguistic and semantic analyses. These functionalities allow the system to be more responsive to the user, to provide a level of insight perhaps not expressed by the user query, and/or add a level of insight not yet spotted by the user community. The art of language is powerfully complemented by the art of the user interface. Effec- tive Findability provides access and interaction with content in a variety of means. The interface can be static or highly dynamic and auto-generated. Findability systems can incorporate sensitivity to individual user habits, preferences, and depth of knowledge. In this manner, the Findability experience can be potentially different not only for each user, but for each user’s interaction with specific bodies of content, or content in a given business setting. This is personalization to the nth degree, which is now practical be- cause of the inherently dynamic nature of e-based approaches to content access. Find- ability interfaces and levels of functionality are invoked dynamically based on the user and the business issue at hand. Distinguishing Findability from Search A fundamental difference between Findability and its predecessor and component, search, is where the burden of effort lies. Under Findability, the burden of intelligent con- tent processing is placed on the content itself. It is the content that is made “findable” in any number of scenarios. It no longer takes asking the “right question,” or having inti- mate knowledge of what you are seeking in order to execute an effective search. Find- ability strategies establish content that is aware of its potential value in any number of settings, and leverage art and science to respond to these settings. The Findability requirements of a system designed to support e-Discovery, for example, are grossly different than one designed to support inventory control. Each of these busi- ness scenarios should be supported with an effective approach to search, but through a common strategy that does not duplicate the underlying content and/or basic functional- ity components of a Findability strategy. Without a Findability strategy, it is likely that an organization, with the best intentions, will create multiple individual search platforms to support its needs. Indeed, our research found that multiple search engines are the norm in most organiza- tions. Search (and by association Findability) is all too often viewed as application- specific, resulting in search silos in many enterprises. There is a separate search solu- Information Architected Phone 617-933-9655 http://www.informationarchitected.com Page 5
- 6. IA Primer: Findability tion for e-mail, Web content, wikis, Blogs, ERP systems, CRM systems, intranets, File shares, and so on. This leads to user frustration with enterprise search. (See Section 4, State of The Market for more detail.) One might argue that Findability is not separate and distinct from search, that it is sim- ply a methodology or approach to deploying search. This is incorrect. Again, search is a fundamental component to Findability, but it is not the sum total of Findability. (See the following section, Findability Technologies for more detail.) Understanding the Findability Dilemma of Precision vs. Recall The ultimate Findability challenge is to strike a balance between precision and recall. Recall focuses on the robustness of the system. Its ability to provide access to any and all content relevant to the user’s focus, i.e. everything relevant, has been recalled. On the other hand, effective Findability should also provide precision. Precision focuses on pinpointing only content that is relevant to the user’s focus. Taken individually, these goals of Findability are relatively easy to achieve. Total recall can be achieved by return- ing every document for every search. The problem here, of course, is the time wasted reviewing irrelevant content. Precision can be very tightly tuned, returning documents that only expressly contain an exact word, for example. The problem this creates is the exclusion of “smoking guns” simply because their relevancy is not specifically spelled out in exact terms. In addition, this approach does not support in any manner software-based approaches to trend analysis or dynamic discovery of relevant content. A Findability strategy comprised of several technologies and methodologies can offer a balance between these two somewhat diametrically opposed metrics. The ability to ad- just these parameters within the context of specific user interactions can be extremely powerful. Information Architected Phone 617-933-9655 http://www.informationarchitected.com Page 6
- 7. IA Primer: Findability Why Findability is Critical Today As introduced in the preceding section, Findability is not a single technology but instead a genre of technologies that can be integrated and deployed in an orchestrated fashion to provide a point of interaction between the user and content. Under a comprehensive Findability strategy, the burden of finding is placed on the content, and not on the user. While it should be appreciated that Findability is more about a well-defined and exe- cuted strategy model than it is about technology, it is nonetheless prudent to have a working knowledge of technology alternatives before embarking on a strategy develop- ment and information architecture. For too many organizations, Findability is synony- mous with search, or is limited in functionality due simply to an ignorance of what alter- native technologies can provide In most organizations, technology tools and techniques are grossly underutilized, accounting for much of the current disappointment among en- terprise users. Before organizations can embrace the state-of-the-art in Findability, they need to have an appreciation of what is and is not possible. Findability is to a very large degree about technology. As a means of introduction, pri- mary point technologies that can comprise a Findability strategy are overviewed and positioned in this section of the report. It is important to note that many of these point technologies are provided through a single product. The point of a Findability strategy is to first determine the features and functions required for the organization, select the smallest number of products that meet all of those needs, and then combine those products under a single interface. Enterprises may opt to deploy multiple Findability platforms, but this approach should be driven by business strategy, and not be merely an accommodation for multiple sets of siloed tools. Defining and Positioning Findability Component Technologies The list provided below is not exhaustive, but provides a good primer on the technolo- gies and techniques most prevalently available today. • Agents • Behavioral Search • Clustering • Concept-based Search • Controlled Vocabularies • Data Dictionaries • Emergence/Trends • Entity Extraction • Federated Search • Free-text Search • Multimedia-based Search • Natural Language Processing/Query • Ontology • Parametric Search Information Architected Phone 617-933-9655 http://www.informationarchitected.com Page 7
- 8. IA Primer: Findability • Relevancy Ranking • Semantic Search • Sentiment Analysis • Tagging • Taxonomies • Text Analytics • Thesaurus Agents are a form of user-created perpetual queries that run in the background. The underlying functionality of the agent is determined by the type of search engine(s) it supports, and can include parametric search, full text search, and concept search. These software- based queries notify their respective owners about the existence of or changes to an area of interest. Behavioral Search, also known as social search, is a form of emergence/trend analysis that specifically monitors users’ interaction with a Findability platform. Search terms they use, paths they take, and, most importantly, the ultimate action they make (e.g., download a specific document), are collected and analyzed. In this manner “popular” approaches and content are determined. This insight is used to provide dynamic links to the “right” content and/or affect relevancy ranking. Clustering dynamically categorizes collections of content. Clustering looks for commonalities and differences between individual bodies of content. Many different types of algorithms are used to make these determinations. There is no standard set of functionality provided in a clustering tool. Therefore, clustering tools should be evaluated individually, and se- lected based on an approach that best mimics the way a given community would cate- gorize its content. Clusters can be defined dynamically (e.g., discovered through behav- ioral search or user tagging) based on a user’s queries, or in static fashion wherein a body of content is analyzed and a group of clusters emerge. Clusters can be used to provide finer precision to a search either in the back-end or as a front-end navigational aide. Concept-based Search is a form of free text search that uses text analytics and semantic analysis to locate and retrieve content based on its overall meaning or on concepts contained within it, as op- posed to the occurrence of specific words and/or phrases. Concept-based search can utilize clustering models. Controlled Vocabularies are a finite set of terms imposed on a collection of content. Controlled vocabularies can be deployed at various levels of granularity. For example, at the low end, the controlled vocabulary can limit the values, or entities, entered into a document’s meta-tag (see Tagging, below, for more on this). At the high end, a controlled vocabulary can be used Information Architected Phone 617-933-9655 http://www.informationarchitected.com Page 8
- 9. IA Primer: Findability to dictate how subjects are discussed within a body of content. Controlled vocabularies can be integrated with thesauri and data dictionaries. Data Dictionaries do not contain any data or content themselves per se, but provide a centralized road- map to the schemas or formats of individual repositories. In this manner, a single query can map into several separate collections of content despite inconsistencies in naming conventions and tags. Data dictionaries can be fundamental to concept-based search and federated search. Emergence/Trends Analysis is a genre of technologies that monitor and analyze the actions of many users and dis- cover the trends that emerge. Folksonomies, link analysis (e.g., PageRank), and behav- ior search are examples of functionality based on emergence/trends analysis. Emergence/trends analysis leverages the wisdom of the user community as a whole to influence the ranking of search results, clustering, and the categorization of content. (See the AIIM Market IQ on Enterprise 2.0 for more detail on Emergence/Trends Analy- sis.) Entity Extraction is a form of text and imaging analytics where the software identifies particular values within a body of content (e.g., a document) and extracts the value automatically for fur- ther processing. Entity extraction can be used to automatically tag content (i.e. values extracted are used to populate the documents’ meta-tags). Extracted entities can be used to display a list of values associated with a collection of content (e.g., a list of all corporations named in a collection can be provided automatically). Entity extraction can also provide a high degree of precision (e.g., a user query into a collection of docu- ments asking for “who is the parent company of ACME Corp?” can result not in a list of documents that are relevant to the query, but the specific name of the parent company, automatically determined and extracted from the content.) Content values extracted through these tools can also be automatically passed to a decision support tree or workflow tool. Federated Search/Universal Search is the ability to provide a single search interface across multiple repositories that each have their own search interface. Federated search is a single point of search. The user interface provides a means to capture a request, pass it to the underlying search tools, collect the returned content from each underlying search, de-duplicate the list (in some cases), and present the list of returned content to the user. Relevancy ranking may be a feature of this system, but the one employed will be the ranking algorithm of the fed- erating tool. (See the description of relevancy ranking in this section of the Market IQ for more detail.) The federating engine does not enhance the search capabilities of the un- derlying search tools. Thus, results can be somewhat inconsistent across sites, as the individual sites may utilize different approaches to search and concept-based search. A form of federated search known as universal search provides a different approach. Under this method, a single engine is positioned as the preferred search processing en- Information Architected Phone 617-933-9655 http://www.informationarchitected.com Page 9
- 10. IA Primer: Findability gine. This engine bypasses others and directly indexes all of the content collections it- self. When a query is posed through the universal engine, it executes across all sites providing a single response similar to federated search, but the end result is completely under the domain and quality control of the one universal engine. While this eliminates any potential inconsistency associated with federated search, it does place all emphasis on the functionality of the universal search tool and its effectiveness in each repository. Free-text Search is a popular form of search on text content. At the highest level, what they have in common is functionality that can locate and retrieve unstructured text con- tent (e.g., a word-processing file) based on a user query that specifies topics and/or words of interest. Many methodologies and specific technical approaches exist to proc- ess the query and the content and render a response. Differences among free-text search tools are significant enough that they warrant investigation and appreciation be- fore strategically positioning one into a Findability strategy. The methodologies and technologies used range from simple word indexes to sophisticated pattern recognition and text analytics. Multimedia-based Search refers to any number of approaches to searching and retrieving non-text-based content, including images, videos, and audio. Multimedia-based search is often performed through tagging and parametric searching. At the high end, however, multimedia search can provide technology that analyzes multimedia content in its native form. This in- cludes facial recognition, image recognition, video recognition (including story board frame change), and audio recognition. Natural Language Processing/Query is a form of text and semantic analysis that in- terprets user queries for content in a natural language. Natural Language Query elimi- nates query command languages. The approaches used to provide natural language query vary widely. At the low end, the process would string the words in the query to- gether with a Boolean “or.” At the high end, methodologies similar to those used in concept-based retrieval would be leveraged to interpret the meaning of the query, and not simply retrieve results based on stated words. Ontology is a form of organizing a collection of content and intelligently linking content resources by declared relationship types. Ontologies can be used to create knowledge maps, a network of relationships that tracks how sources of knowledge relate to one another. For example, a "lives at" link or "works for" link in ontology would be used to track these re- lationships for listed individuals. Ontologies are similar to thesauri and taxonomies, but support the definition of contextual, customizable, self-describing relationships. An on- tology is the framework of the semantic Web. It permits intelligent navigation through a body of content and can be used to support entity extraction. Parametric Search (metadata-based search) is search based on meta-tags. If content is associated with one or more tags (i.e. data values associated with a body of content but not necessarily stored within the content itself such as a date, author name, or contract number), parametric search enables the Information Architected Phone 617-933-9655 http://www.informationarchitected.com Page 10
- 11. IA Primer: Findability query and retrieval of the content based on user specific values within the meta-tags. For example, “Retrieve all documents published between January 2007 and January 2008,” is a parametric search leveraging a published date tag. (See Tagging for more detail.) Relevancy Ranking is a genre of methodologies and technologies used to display a list of retrieved content in some type of prescribed order. Relevancy ranking is adjunct functionality to a multi- media or full-text search, as it lists retrieved content in descending order of the level to which it is pertinent to the query statement. There are no standard approaches to rele- vancy ranking, so the inclusion of this functionality in a Findability strategy warrants an analysis of all potential tools to establish the exact approach each one uses to deter- mine ranking. Methods range from ordering content by publishing date to establishing a relevancy quotient of content to query based on semantic analysis, text analytics, and/or clustering techniques. Relevancy ranking encourages broad recall (i.e. retrieve anything and everything remotely relevant to a user query), but simultaneously provides precision by allowing the user to interact with the content in this order, and thus make an informed personal decision about when the level of relevancy has dropped off. Semantic Search (Semantic Analysis) incorporates the analysis of meaning in text, identifying concepts and their relative im- portance to the subject of the content and to each other. These utilities can form the ba- sis of concept-based search, natural language processing, cognitive understanding, and text analytics. Sentiment Analysis is a form of semantic analysis that determines the perspective (i.e. positive or negative) a body of content has on a particular subject. Sentiment analysis can be integrated with agents (e.g., alert me when something negative is posted about my subject) or rele- vancy ranking (e.g., rank positive sources first) and could be utilized in situations such as e-discovery, where an automated and timely determination of positive versus nega- tive sources can be insightful. Tagging in its simplest form is the intersection of databases with content management. It is the ability to assign meta-tag values, also referred to as facets, to any and all stored bodies of content. Tags can range from simple things like an author’s name, to user-specific values such as “favorite bookmarks.” The challenge of tagging is determining what tags to define, and then populating the tags. (Entity extraction can help in this regard.) The value of tags lies in how they are utilized within a Findability strategy to support taxonomies, ontologies, parametric search, and relevancy ranking. Tags can be used a primary search vehicle (see Parametric Search), or as an adjunct to a full-text search, providing added precision. Tags can also be used in an emergent fashion. For example, folksonomies are a form of tagging in which individual users tag content themselves as a way to sort, store, and retrieve content, but simultaneously the tags used across all users are analyzed to look for common themes, rendering a folksonomy. Information Architected Phone 617-933-9655 http://www.informationarchitected.com Page 11
- 12. IA Primer: Findability Taxonomies are a hierarchical or poly-hierarchical listing of topics or subject categories. A taxonomy may not contain a definition of the topics it covers, but only the hierarchical relationship of the topics to one another. A taxonomy can incorporate content from a thesaurus and an ontology. There are no standard file formats or approaches to taxonomy construc- tion. Findability strategies can include technologies that create a taxonomy (based on text analytics and semantic analysis), provide a framework in which a taxonomy is manually built, and/or automatically associate individual bodies of content with a node of the tax- onomy (automated categorization), based on text analytics and semantic analysis. The taxonomy can be used as a navigational front-end to a body of content. It can also be used as a back-end to search (providing more precision in retrieval), or as a form of input to a clustering engine. Taxonomies can also be leveraged as a front-end to search, providing structure to the way retrieved content is arranged and presented. This taxonomical display of retrieved content in response to a query, provides greater preci- sion and a customized, in-context navigational aide through a body of retrieved content. Text Analytics is a combination of semantic analysis, linguistics, entity extraction, tag- ging, pattern recognition, lexical analysis, and other forms of artificial intelligence used to infer meaning from bodies of textual content. Text analytics is a form of data mining on textual content and can be an integral part of a Findability strategy, complementing search and/or extracting data from documents. It provides business intelligence (BI) and data mining within the realm of textual content. Thesaurus refers to a network of words and word meanings and relationships used to put concep- tual definitions into context. It defines a lexicon and the relationships between words and phrases in that lexicon. A thesaurus can be used to enhance the intelligence and effectiveness of a taxonomy and/or full-text search by providing insight into word mean- ings and term relationships. A thesaurus, when integrated into a Findability strategy, can enhance precision and recall. Thesaurus construction is governed by ANSI standard Z39.19. Information Architected Phone 617-933-9655 http://www.informationarchitected.com Page 12
- 13. IA Primer: Findability Developing Your Findability Strategy It is likely that users in your organization instinctively understand that Findability is criti- cal to their own goals and success, as well as those related to the overall organization. Yet, Findability within organizations is severely lacking, especially in contrast to the con- sumer Web. Users of commercial tools and sites such as Yahoo, Google, Amazon, and eBay often wonder why such Findability functionality does not exist in the corporate world. Compounding this is the fact that technology solutions and related strategies are no longer viewed as the sole purview of IT. Business users are consumers: educated enough to be dangerous. They are now confident enough to know they can demand more, and become frustrated when they cannot simply have outside-the-firewall experi- ences inside it as well. If only it were that easy. Part of the user dissatisfaction with enterprise Findability stems from information man- agement solutions in organizations that have developed organically over a period of years. Also, the long tenure of a particular Findability technology —search — can poten- tially confuse strategists within the firewall. A belief that Findability is synonymous with search leads some to conclude they have a Findability strategy if they have search technology. All too often, however, search is linked to an individual application, which leads to fractured Findability. Moreover, applying search tools in an ad hoc manner often does not result in effective Findability, which in turn often leads to blaming the search engine for the failure. In real- ity, the tool may be more than adequate for what it has been designed to do, but the de- sign behind its deployment is flawed. Findability introduces a platform of functionality that goes well beyond search. Many Internet-based search experiences today transparently include approaches to relevancy ranking, integration with many types of taxonomies (from stackonomies to folksonomies, deployed at multiple levels), behavioral search, text analytics, and sentiment analysis. But most organizations do not have a centralized, strategically developed approach to Findability. Under such a methodology, multiple layers of Findability technologies are orchestrated and coordinated as a service, capable of being deployed against any number of content sources and repositories. A well-executed Findability strategy is the antithesis of approaches that link multiple sets of search and management tools with individual repositories. That strategy only causes user frustration due to the need to jump from one research environment to another, and merely provides an illusion of content integration. Though perhaps not the only or most efficient manner in every case, the provision of a singular, navigational front end and omnipresent search tool that collectively aggregates Information Architected Phone 617-933-9655 http://www.informationarchitected.com Page 13
- 14. IA Primer: Findability disparate content resources can deliver the simple and effective single point of access many users desire. The provisioning of enterprise Findability experiences that equal or even surpass those on the Internet requires a rigorous needs assessment and strategy. But therein lies the challenge to the business and IT communities. While the business side must determine the right levels of functionality required, IT must develop approaches that simplify the delivery of it and minimize the number of front-ends. Understanding myriad sources of unstructured information (e.g., e-mail) requires the orchestration and coordination of multiple disciplines and technologies working in concert. The development of a Findability strategy has to be addressed deliberately and uniquely within each organization. It is not a product that can be purchased, but instead an envi- ronment that requires analysis, design, and ownership. Development of an effective Findability strategy requires multiple disciplines. We defined Findability as the art and science of making content findable. The Findability strategist cannot simply be someone in IT who can support the technical deployment of a “search tool,” but someone who is part library scientist, part linguist, and part interface designer (or someone adept at bro- kering these talents into a single team). The Findability strategist must also be technically savvy. This individual (or team) should have an understanding of a full range of Findability technology alternatives, what each offers individually, and how they complement one another. This individual (or team) must also have an appreciation for business objectives, including legal and security concerns associated with information access. From a technology perspective, a sound Findability strategy includes components such as automated classification, taxonomical and facet-based interfaces, ontologies, tag- ging, and search. The search function is potentially complex, comprised of many fea- tures and functions such as natural language processing, concept analysis, and user sensitivity. This is the more difficult aspect of designing and implementing the technol- ogy to support a Findability strategy. Each component should complement and lever- age the capabilities of the others in a seamless and intuitive manner. The end result or "total solution” should be an interactive, personalized user experience that provides a single, albeit dynamic and heuristic, approach to finding content. It must be stressed that the full value of investments in search technology will not be re- alized by purchasing the "best" search engine, but by integrating that engine into an overall Findability strategy. In the end, you want to give your users the ability to effi- ciently recover information so they can use it in innovative and targeted ways. Inside the enterprise, however, this is no small task, particularly in light of users’ expec- tations. What the Internet has taught us is that “search” can occur across “all” content, applications, and repositories. Whether looking for text files, images, or videos in one collection, across multiple collections, in known repositories or discovered in unknown sources, the general approach or interface should be the same. What Internet- Information Architected Phone 617-933-9655 http://www.informationarchitected.com Page 14
- 15. IA Primer: Findability enlightened users can fail to realize is that while results equal to those achieved with commercial tools on the Web are obtainable inside the firewall, accomplishing this is not as simple as replicating what the commercial sites do merely by deploying the same technology. Planning Advice Each Findability strategy is potentially unique. For example, the steps taken to organize a collection of digitized video and audio files for purchase and download by a general- interest consumer community may not work for a collection of scientific findings needed by marketers and developers of chemical compounds hoping to increase collaboration and the rate of innovation. What remains the same in each instance, however, is the re- quirement for some form of Findability, working in a coordinated effort with search, navi- gation, taxonomy, personalization, and interface. Although the potential technical com- ponents remain the same, the individual features, make-up, and execution should differ. Begin the development of a Findability strategy by examining Findability scenarios across your organization — or at least to the level of sponsorship. (One of the major is- sues affecting Findability is a lack of ownership of the strategy. The level of planning, change management, time, and technology investments that may be required will not be supported or funded without a corporate sponsor or owner. Therefore, it is important to identify early on who that individual or team of individuals is. Their position within the organization will dictate the scope of the strategy development. ) Using the sponsor’s domain as a framework, assess the needs for Findability across the organization. This is accomplished through a combination of process analysis, content audit, and user needs assessment. The process can be facilitated if the needs assess- ment and rationalization of findings is done within the context of content, community, and context. Focus on why the content is created in the first place, who needs access to it, why they need access, and all possibilities of how to get answers from the content. Content The questions below address the issues that need to be addressed regarding content: • How is the content created and by whom? • Where is the content located and how many databases and repositories is it in? • What categories of information exist? • Do different types of content require or have different tags? • Who are the owners of content? • What is the volume of content? • What similarities and differences exist within the body of content? • How dynamic is the content collection? Community The types of issues that need to be addressed regarding community are: • Who is the intended community of users? Is it one community or several? • Are all users created equal? Do the needs of some supercede the needs of others? • What interface models remove ambiguity or fit well with the users? Information Architected Phone 617-933-9655 http://www.informationarchitected.com Page 15
- 16. IA Primer: Findability • What attributes make the content relevant? • How proficient are the users in understanding the content? Context Pay particular attention to context. While understanding content and community is im- portant, it is context that ultimately defines the intersection of the two. Context places the reality of content and the needs of the community within the demands of business goals and objectives. The types of questions that need to be addressed regarding con- text are: • What is the ultimate goal of the Findability strategy (e.g., to enable customer self- service, ensure compliance, facilitate e-Discovery, foster expertise identification and brokering of tacit resources, facilitate collaboration, expedite knowledge discovery, as- sist targeted retrieval, detect brand sentiment, mine business intelligence — or all of the above)? • What is the value of content after it is retrieved? What do users do with content after it is found? • How is the relevancy of content determined? • What access rights need to be supported? It is important to note that security can be a major component of a Findability strategy. With increased Findability often comes a need for greater control over access. In fact, security should be viewed proactively and may require iterative assessments. The need for security may not be apparent initially. Prior to the availability of effective Findability, some content within an organization may naturally “hide itself” from inappropriate use or discovery. That is to say, ineffective Findability functionality may have rendered the un- secured content nonetheless difficult to find and thus “self-protecting.” Although security can be viewed as a feature of the content itself, it is better to view ac- cess rights from the context perspective. This allows you to appreciate that security rights may be dynamic and contextual, but it will also let you weigh not only what secu- rity is required, but what security measures may in fact be unnecessary and actually im- pede a Findability strategy. Security should be viewed as integral to the entire Findabil- ity strategy. By understanding the context of Findability, the strategy can also support "process intel- ligence.” Imagine a “search tool” that does not simply return content that satisfies a user's query, but actually interfaces with inventory systems, marketing campaigns, or customer support tracking systems to dynamically trigger or advance processes, as well as alter or greatly personalize how content is retrieved, ranked, and displayed. In the end, it is not a single component that determines the best approach to Findability, but the intersection of content, community, and context. In more technical terms, this is the development of an information architecture, an underlying foundation that manages all content (whether in a single repository or not), and provides a flexible, adaptable in- Information Architected Phone 617-933-9655 http://www.informationarchitected.com Page 16
- 17. IA Primer: Findability terface to the content to support the rapid, accurate identification and retrieval of con- tent. It should be appreciated that content, community, and context can break out into multi- ple levels. Content does not have to be physically positioned as a single collection or subjected to a single records plan. There may be separate and distinct communities within an organization, each with very different needs. Such unique combinations of content, community, and context must be examined, at least to see whether there is merit in specifically and uniquely supporting them separately. Does each instance need its own search tool, relevancy ranking scheme, tagging system, and taxonomy, or can functionality be shared and “personalized” dynamically and deliberately in context? With the Findability strategy clearly defined, you can begin to assess technology alter- natives. For the search component of the architecture, seek out technology providers that stress ubiquitous integration across business applications with highly customizable and context-sensitive interfaces. Look for integrated text and data analytics, contextual search and navigation, multimedia search, intelligent value/entity extraction and linking, social and collaborative search/tagging, and transparent and appropriate security as re- quired by your strategy. This approach to providing “search” inside the firewall is rigorous. It requires far more work than simply selecting a search engine, or accepting the search tool that comes embedded inside another application. But the effort likely results in the production of a enterprise asset — an ability to leverage corporate content, to eliminate reinvention, poor communication and miscommunication, increase innovation and responsiveness, support compliance and e-Discovery, and facilitate collaboration. But, remember, con- tent without access is worthless. Information Architected Phone 617-933-9655 http://www.informationarchitected.com Page 17
- 18. IA Primer: Findability About Information Architected Information Architected is a consultancy focused on the intelligent use of content, knowledge and processes to drive innovation and help you thrive in a digital world. We offer three distinct lines of service: Analysis (case studies, market reports, whitepapers and podcasts) Consulting (strategy development, needs assessment, ROI calculations and solution architecture development) Training (from 1/2 day overviews to multi-day in-depth training). Our founders have over four decades of experience and industry recognition as ex- perts and thought leaders in facilitating the successful creation of business strate- gies that leverage technologies and practices including: Business Process Man- agement (BPM) and Workflow, Enterprise 2.0 and Collaboration, Enterprise Content Management, Information Architecture, Taxonomy, Enterprise Search, Innovation Management, Knowledge Management and SharePoint. Our Information Architecture, Search and Taxonomy training provides insight into technology alternatives and business strategies to leverage these. Is Your Information Architected for Findability? Contact us at 617-933-9655 or info@informationarchitected.com to discuss how we can re-build your Information Architecture to enable effective Findability today. Information Architected Phone 617-933-9655 http://www.informationarchitected.com Page 18