Knowledge Management (KM) can be defined simply as the process through which organizations generate value from their intellectual and knowledge-based assets. Knowledge assets are often grouped into two categories:
(1) Explicit Knowledge:
Generally, everything and anything that can be documented, archived and codified. Examples include patents, trademarks, business plans, marketing research and customer lists.
(2) Tacit Knowledge:
The rest. Tacit knowledge is the know-how contained in people's heads. The challenge inherent with tacit knowledge is figuring out how to recognize, generate, share and manage it.
Most often, generating value from such assets involves sharing them among employees, departments and even with other companies in an effort to reach – or go beyond - best practice. For explicit knowledge, the focus can usefully be described as “connecting people to things”, whilst for tacit knowledge, the focus is “connecting people to people”.
Search technologies made simple
There are essentially two types of search technology: structured search and unstructured search:
1) Structured Search:
In a structured search (example Yahoo) the user clicks down through a directory of categories to find the material sought. The tree structure of the directory is called a taxonomy, with a root node at the top that applies to all objects and nodes below that classify more specific subsets of the total set of objects. A well-known example of a taxonomy is Carolus Linnaeus's Scientific classification of organisms. The root node is (implicitly) “organism” and nodes below are Kingdom, Phylum, Class, Order, Family, Genus, and Species.
In Yahoo, the root is “Directory” and there are 14 main nodes, including Society & Culture, Social Science and Reference. If I were searching for best man’s speech tips (which I was not so long ago), I could try (on Yahoo) clicking down the branch of the tree: Directory > Society and Culture > Weddings > Speeches and Toasts. Alternatively, I could try Directory > Social Science > Communications > Public Speaking.
There are obvious pros and cons to structured search. The main con is that I might head down several blind alleys on the tree before finding the most useful branch. The main pro is that – once I have found that branch – I am likely to find a whole collection of relevant material gathered together in one place.
For the librarian (maintaining the taxonomy), there is another key con. As our wedding example illustrates so well, there are often two or more places one could put any given information. Oh the agonies of choice!
2) Unstructured Search:
In an unstructured search (example Google) the user enters a series of keywords into a search engine, which searches an index of content (regularly crawled) and brings back results, ordered by closest match to the search string.
The main pro of an unstructured search are that I do not have to second-guess where the right branch is on the taxonomy tree, but rather leap in at the leaf I am looking for. The cons are rather less obvious and the main one is that some materials I might actually find very useful may not come up in the search. This can be due to my poor selection of search terms, deficiency in indexing / search algorithms or poor metadata in the content itself.
Should I implement search functionality and, if so, how?
Search invariably scores well on any prioritisation of intranet functionality and is generally “out-of-the-box” with your portal solution, so I would definitely recommend you include it in the scope of your project.
It may not surprise you to hear that the most effective search implementations allow the user to chose between structured and unstructured search options and to easily navigate between the two. For example, my unstructured Yahoo search on “best mans speeches” takes me straight to a relevant document, but also tells me where it sits in the directory. By clicking on the directory category, I can bring up all the other materials in that area (where I ultimately find the best resource for my need).
To implement the structured search part of your solution, you will need to develop a taxonomy structure for your organisation and the information resources your people need to do their jobs. This can be quite a challenge! For example, should an HR grievance policy be found under ABC Co > Human Resources > Employee Services Unit > Policy or under an ABC Co > My Employment > My Rights > Grievance branch?
My advice is to keep it simple and give it room to evolve and change. An ideal taxonomy should be flat and broad (having no more than three levels) and should suit the provider or creator of information rather than the user of it (as they are the people who will populate your library and you need it to be easy for them to do so).
To ensure the unstructured component of your solution is effective, you need to ensure firstly that people avoid jargon in the body of their documents (using instead keywords that users will recognise) and secondly that a high percentage of documents contain decent metadata. Metadata can be simply defined as "data about data". For example, the grievance document metadata might include author: Tessa Jones, job title: Employee Relations Officer, department: Employee Services, function: HR, subject: Employment, title: Grievance policy.
Can search help with tacit knowledge sharing?
Absolutely! Many organisations fail to recognise this. Connecting people to people (for that 10 minute telephone conversations that could save a week’s work) is often much more valuable than storing documents.
You should create a well developed yellow pages database, where people have entered augmented their white pages details (job title, email address, telephone number) with information about their skills, experience and interests. Then – when someone searches for grievance – in addition to (a) the word document policy, the results also include (b) a link to Tessa Jones’ Yellow Pages entry and (c) a link to the Employee Relations teamspace, where Tessa – and her line colleagues across ABC Co – collaborate on policy development and employee relations management.
Some final thoughts
The humble search function can be the most powerful agent for improved knowledge management your organisation has ever invested in. By extension, therefore, it can become the definitive “killer application” on your intranet portal. However, it is vital that the search capability can acccess all the information and people in your organisation and that result relevancy is high. This is not as easy as it sounds and requires proper planning and detailed work.