Be sure to distinguish among the words content, information and data as described here.
For more speaker notes, see the PRACT-1 deck
A content model is a representation of components or pieces that make up a body of content. Before I go into all of the details of this definition, I would highlight the picture on the right. You might be wondering why we have a picture of a golf hole. There’s a reason for that. Any piece of information has some sort of content model, a way that it’s represented; the different components of that content put together make up a model. So in this particular case, there are different elements of this model. You have a set of numbers that represent different yardages. You also have pictures that give you a visual representation of how this particular golf hole is laid out. You have trees along the left. You have a lake along the right. Then there are various points on the hole that are represented to give you a sense of the distance from one part of the hole to the next part. All of this is what helps the golfer, or the person using this particular content model, make decisions or decide what needs to be done with the information to take action. Going back to our definition, you can also look at a content model as the folder, repository metastructure, or broader enterprise information set. You could think of any piece of content in your enterprise just as this sort of golf hole is laid out. How are your documents laid out? How are your repositories laid out? What are the different components within those repositories and within those documents? That’s what really makes up your large enterprise content model. Where you run into problems is where that content model isn’t consistent. If I took this golf hole—and this one in particular is represented in yards—and showed it to a golfer who’s used to looking at distances in meters, you start to run into some challenges. The same sort of things exist within enterprises when you represent information in different ways. Content models are how you need to see and think about information. So when you look at a document or a picture like this golf hole, you need to think about what the components are, how they break down, and how to potentially reorganise them to optimise them for your content technology.
Many organisations will attempt to establish metadata strategy. And this is usually done in a larger context of a content strategy or a content management strategy. But considering metadata in and of itself is an extremely important part of that. And this is a process of identifying and understanding the different metadata types and their purpose. So as we talked about when I showed that recipe example, that metadata model, it wasn’t just about identifying what the metadata is, but also what’s the point of it? What’s the purpose? What exactly is this metadata going to be used for? The second point that I would make, that it might be the most challenging part of a metadata strategy, is the synchronisation and adoption across a department, a project, and ultimately the entire enterprise, so what something is called, how it’s labeled, and what that means. The consensus building ends up being perhaps the most challenging part of any metadata strategy and eventually actually implementation of the metadata. And that’s because it is very difficult to get different departments to agree on what something should be called. And there are some ways to solve it if you can’t, in fact, come to a consensus. And that’s where the thesaurus comes in, which we’ll talk about a little later. Finally the understanding of the people, processes, and systems, how they’re interacting with metadata and vocabularies, who owns that metadata, and planning for the maintenance and changes of the metadata — these are all very important parts of a metadata strategy and a content strategy overall.
There’s a number of benefits to establishing this strategy early on. Consistent use of the metadata structures across the enterprise makes the metadata more powerful. If the metadata is not consistent, the effectiveness really gets diluted. You won’t be able to consistently find information. And, of course as I’ve mentioned numerous times, the information and systems will be more interoperable and be able to exchange content more easily if those structures are consistent. The disambiguation I mentioned earlier is also very important. Understanding how the metadata changes will affect downstream processes is also very important. If someone changes metadata, that might kick off a different process or it might change where the content ends up being displayed. It’s very important that your content managers understand the ramifications of tagging content in certain ways. Identifying gaps are also very important because, odds are, you have some implicit metadata already in your organisation of course. So knowing what the gaps are usually is accomplished by understanding the end use of the content, how people are searching for it, how that content might need to be used within a system or on a website. That’s often where you realise you need certain types of metadata in order to accomplish that. So now’s the time during the strategy to identify those gaps. And then of course communication. Part of the key of establishing a strategy is making sure that everyone’s on the same page and that you communicate it out to the organisation. And with metadata, people outside of that initial circle that’s establishing a strategy might actually find a lot of use for that metadata that they hadn’t thought about before there actually was a metadata strategy. Understanding the level of effort to tag or index content is also something that tends to come out while establishing a strategy. You need to be prepared with the adequate level of resources and motivation to really tag an index content. Because it is something that can be very laborious and require a good deal of time and effort. And we’ll get into some more detail on that in a moment. Also establishing someone in the group with a centralised knowledge of the metadata process, this could be a single individual or it could be the whole group that was involved with establishing the strategy. Either way, it’s important to have that central source that people can go to if they have a question about how to apply metadata, how it’s being used across the enterprise. Actually implementing the strategy is something we’ll get into in more detail in the specialist track.
For more speaker notes, see the PRACT-1 deck
In our last section before wrapping things up, let’s talk about automated collection of metadata or auto-tagging. Little bit of content around this problem, for human beings, of course adding metadata often means quite a bit of work. The people who are assigned to tag, sometimes they don’t really see the benefit of getting involved. It becomes to them a chore or laborious process. They might even say, “Well, it’s not my job, so why should I do it?” In this case, I would use an example from many implementations of content management systems whereby explaining the result or the ramifications, or most importantly, the benefit of tagging is a good thing to explain to the people who would be doing the tagging. Because if they see the benefit, and you, in fact, explain what’s in it for them, so to speak, it might actually help and incent them to get involved with the tagging. And in some cases, let’s not forget, it’s not just anybody who should be tagging. In some systems, it’s a very specialised skill, not only to necessarily use the system and to tag, but also perhaps around the subject. You might need a highly specialised subject matter expert to tag things accurately. Commonly, human beings provide incomplete or inaccurate metadata, especially if you have a large number of people tagging. It’s very difficult to not leave it open to interpretation. People will tag things differently based on what department they’re in, et cetera. So given the challenges in, not only incenting people to tag, but also in keeping what they do complete and accurate, the question arises, well, is there a way to get machines to add metadata for us, consistently, accurately, and to alleviate some of the burden from people within enterprises? Let’s explore that question a little bit now.
Here’s an example of how someone might go about indexing a scanned image. So you have a picture here in this example on the right, where you see a little thumbnail on the far right of a document that’s been scanned by some sort of imaging technology. And an indexer might have to fill in information about this particular document. In this case, we have a claim number. We have a policy number. We have the last name of the person who’s being insured. So obviously this is some kind of insurance document. And then there’s also a claim date and a random notes field. So this data, when it gets entered into the system, is usually stored in a separate database, and it’s associated with the image file so that if someone wants to pull it up, they can type in any of these items, whether they’re looking for a particular claim number or a claim date or based on someone’s name. And then the system would pull up the document based on this particular metadata that’s within the index. The notes field is a little bit more complex, simply because the information isn’t as precise or discrete as the other fields are. So the notes field might actually have to be indexed by a text-based search engine to become more easily searchable. We’re going to talk a lot about search technology a little bit later, and how it might actually deal with a free form field such as this notes area.
Making Content Easy to Find<br />DC2010 – Pittsburgh, PA<br />Betsy Fanning<br />AIIM<br />
Who is AIIM?<br />The leading industry association representing professionals working in Enterprise Content Management (ECM). We offer a Membership Value Program Focused On:<br /><ul><li>Market Education
Professional Development</li></li></ul><li>About AIIM Standards<br />ANSI Accredited<br />ISO TC 171, Document Management Applications – Secretariat<br />ISO TC 171, Document Management Applications, SC2, Application Issues – Secretariat<br />U. S. TAG (Technical Advisory Group) to ISO TC 171 Administrator <br />Industry Standards Developer – AIIM Recommended Practices (ARP)<br />Open Source Standards for Document Management<br />Liaison Relationships<br />
What is ECM?<br />The tools and technologies used to:<br />Capture — move content (in any form) into your repositories for reuse or retirement<br />Manage — move it around the enterprise to drive key applications and processes<br />Store — put it in a logical place for easy access<br />Preserve — long-term archival and storage<br />Deliver — get to the right audience on the right device <br />…documents and content related to organization processes.<br />
What is content?<br /><ul><li>Content comes in a variety of formats:
Semi-structured content such as HTML</li></li></ul><li>What is Expected?<br />Information should be easy to discover or locate<br />Information access is about helping users find documents that satisfy their information needs<br />Remember, someone may be looking for something they’ve never seen or touched before<br />Information should be easy to tag or assign the metadata<br />
Organizational issues<br />Which of the following organizational issues have you experienced with your SharePoint implementation?<br />40+% no planning or strategy<br />26% lack of information management expertise <br />N=362 SharePoint using or implementing<br />
Know What You Have<br />In order to improve information access, you need to know<br />How much content you have<br />What types of content you have, and its relative value<br />What content needs to be archived, retained, or deleted<br />In order to undertake a successful ECM/WCM/RM/Search implementation or improvement effort you need to know:<br />What documents you possess <br />Who “owns” the content in order to determine proper security, roles and permissions<br />Who or what creates content in order to properly tag/index and otherwise contextualise and enrich content<br />Ultimately, you need to create an overall Content Model<br />
What is a Content Model?<br />Components or “elements” that make up a body of content<br />The folder or “meta”-structure of a repository or enterprise information set<br />The document types <br />Associated metadata<br />Elements within a (structured) document<br />A framework applied to content to create relevant information<br />Making those related pieces useful to the people who need it<br />This is how you need to see and think about content<br />
What is a Metadata Strategy?<br />Identification and understanding of different metadata types and their purpose<br />Synchronisation and adoption across a department, project, and ultimately the entire enterprise; <br />Agreement on terms, labels, and meanings<br />Understanding of people, processes, and systems applying and interacting with metadata and vocabularies<br />Understanding who owns various metadata and structures<br />Planning for maintenance and changes<br />Source: Ed Stevenson, Really Strategies, Inc.<br />
Benefits of Strategy<br />Consistent use of metadata structures across the enterprise makes the metadata more powerful<br />Information and systems become more interoperable<br />Lesser chance of ambiguous terms when metadata and its purposes are defined, helping to ensure quality in the metadata<br />Understanding of how metadata changes can affect downstream processes<br />Identification of gaps in what should have more metadata<br />Communication of metadata information to others who may find uses for the content outside its original area<br />Realistic appreciation for level of effort to “tag” or “index” content<br />Establishment of someone or some group with centralised knowledge of the metadata processes<br />Source: Ed Stevenson, Really Strategies, Inc.<br />
Governance<br />Which of the following governance policies do you have in place for SharePoint usage?<br />55+% trying to address team-site sprawl<br />22% guidance on classification and metadata<br />16% or less on retention, legal discovery - and emails!<br />N=391, Using or implementing, May 2010<br />
Why?<br />Digital content is expanding at almost unmanageable rates<br />New information worldwide has been increasing on average 30% a year (doubling every three years)*<br />Getting access to the right information is an increasingly acute challenge for enterprise employees and customers alike<br />Better Information Organisation leads to better Access<br />*<br />http://www2.sims.berkeley.edu/research/projects/how-much-info-2003/<br />
ECM Drivers<br />When you consider your document and records management projects and priorities, what is the most significant business driver for your organization? (Check only ONE)<br />Efficiency and business process: 45%<br />Compliance and risk: 28%<br />N=680 Non-trade,<br />
ECM Drivers<br />Thinking about the compliance benefits of ECM and Records Management, which of the following are the TWO most important compliance drivers in your organization? <br />#1 Customer/supplier litigation <br />#2 Financial reporting and audit<br />N=680 Non-trade,<br />
Metadata and ECM <br />Metadata often acts as a “great unifier” in the area of content technologies and enable them to work together<br />Many content management systems depend on solid library and categorisation services order to add significant value<br />Essential for organising any large content corpus<br />Required for meaningful records management<br />Critical to effective findability<br />How you choose to design the repository, and how the system you choose can use certain repositories and content structures, greatly influence the business value you can realise<br />
ECM Drivers – content types<br />How are the following content types managed and archived in your organization?<br />40% with documents in ECM/DM/RM system (scanned and electronic)<br />15% storing emails in ECM/DM/RM, <br />29% in EMM system<br />N=604 Non-trade,<br />
ECM Drivers - Electronic<br />How confident are you, that if challenged, your organization could demonstrate that your electronic information (excluding emails) is accurate, accessible, and trustworthy?<br />Electronic (not emails) <br />41% Slightly or not at all confident <br />N=607 Non-trade,<br />
ECM Governance<br />Who is the highest person in your organization who has specific reporting authority, or management ownership, of document and records management?<br />28% have a CIO who really is a CIO<br />Plus 11% with a CRO<br />39% have no board-level ownership<br />N=645 non-trade<br />
SharePoint - Use<br /> How would you describe your use of SharePoint in the following ECM areas? <br />Most people manage documents in SharePoint<br />Lots are using it as a portal to other systems<br />RM low but set to rise<br />Capture and emails v. low<br />N=436 SharePoint using or planning<br />May 2010<br />
ECM DC Use<br />Many content technologies are now offering Dublin Core standard repositories and content formats out of the box<br />SharePoint uses Content Types<br />Tied to business process or document type<br />Shared across site collections<br />DC is used with file formats – PDF and PDF/A<br />
Sometimes humans provide incomplete or inaccurate metadata</li></ul>So a question arises:<br /><ul><li>Is there a way to get machines to add metadata for us?</li></ul>Source: Taxonomy Strategies<br />
Indexing a Scanned Image<br /><ul><li>A person adds data to the record of a scanned document
That data is (typically) stored in a separate database, associated with the image file
Later, the scanned document can be retrieved by number, name, or date
The Notes field will likely need to be indexed by a text search engine to become searchable.</li></ul>Source: QMainFrame<br />
PDF and Metadata<br />General information about the document, i.e., title, author, creation and modification dates<br />Used to help search for documents in external databases<br />PDF metadata may be stored in document information dictionary or a metadata stream<br />
PDF/A Metadata<br />Requires the use of Extensible Metadata Platform (XMP)<br />Proprietary, but open format – soon to be ISO <br />Used for metadata creation, processing and interchange<br />Based on restricted form of Resource Description Framework (RDF) – W3C standard<br />Fosters re-use, re-purposing across domains<br />Enables metadata capture, preservation, and propagation across devices, applications, file formats<br />Not limited to a specific schema<br />
Document Information Dictionary<br />Title<br />Author<br />Subject<br />Keywords<br />Creator<br />Producer<br />CreationDate<br />ModDate<br />Metadata entries are optional and are deleted if not provided.<br />