Taxonomy principles

The purpose of this article is to provide guidance on how to plan your taxonomy for Atlas. Please be aware that this is mostly just a roundup of advice you will have recieved during your IA and taxonomy workshops, but may be useful to refresh your memory.

• Use the templates as a talking point
• Focus on the top-level terms
• Align terms with existing data
• A term set can be used for multiple purposes
• Avoid deep hierarchical structures by using combinations of term sets
• Create layers that are useful and meaningful
• If the term set does not work, then it does not have to be used
• Ensure someone (or a team of people) are responsible for managing terms and monitoring their ongoing use
  
  

Use the templates as a talking point

As part of your engagement with ClearPeople we can provide templates with examples to help show how the term sets can be used. Each customer will need to build their own specific taxonomy that will make sense for the business, using the Atlas term sets; Activity, Entity, Location, Department, Subject and Information Type.

We do not recommend that you use the templates directly, they are just to provide realistic examples and provide discussion points for your own taxonomy.

Focus on the top-level terms

When creating the taxonomies, invest enough time in getting the initial layer right. All the lower level sub-terms can be added later when they are needed.

For example, under the Subject term set, agree the types of subjects that will be relevant such as Expertise, Initiative, Topic, Sector and review where content logically fits under these headers first. Once agreed, then further categorisation of each type may be useful:

• Sector
  • Energy,
  • Transportation,
  • etc.

Further breakdown will only be useful if the volumes of content you have for each sub-term warrants the breakdown for refining searches:

• • Energy
    • Coal
    • Gas
    • Oil
    • Solar
    • Wind
      

So if you have a low volume of Energy content, the sub-terms may not be required.

Align terms with existing data

Ensure any existing taxonomy terms used in other enterprise systems are considered for suitability and reuse. Apart from simply being a good place to start the conversation, existing taxonomies are also likely to be known to users and familiarity will help them to understand the new system. 

For example, you should align your Departments taxonomy with your existing HR / Active Directory data so that when viewing People Directory information it aligns with the content taxonomy terms.

A term set can be used for multiple purposes

Atlas provides several core term sets, these can be used for multiple purposes rather than having to create new managed metadata terms of hierarchical structures.

For example, rather than adding a government initiative that affects all sectors to every sector, create two term sets within Subject;

• Subject -> Initiatives -> Government -> Net Zero

AND (combined with)

• Subject -> Sector -> Transportation
• Subject -> Sector -> Energy

Rather than...

• Subject -> Sector -> Energy -> Net Zero
• Subject -> Sector -> Transportation -> Net Zero
  

This helps in multiple ways;

1. When searching for a term to tag your content, you avoid having multiple "Net Zero" terms and being unsure which to choose,
2. When searching for this content it could be returned using "Net Zero", "Transportation", "Energy", or a combination, as all three tags are applied to the content, rather than having to choose to see the "Net Zero" information for one sector or the other.

Avoid deep hierarchical structures by using combinations of term sets

Use multiple term sets to create relationships between content rather than a single tree structure. 

For example, rather than have an Entity structure that describes a Client by the lead Sector, Department and Entity type all in one tag, use tags from Subject, Department and Entity together.

• Subject -> Energy
• Department -> Corporate Law
• Entity -> Client -> ClearPeople

Rather than...

• Entity -> Energy -> Corporate Law -> Client -> ClearPeople

Create layers that are useful and meaningful

Create term set groupings that make sense to how people locate or need to locate information.

For example, if it makes sense to search for all countries in Europe, then create a node called Europe in the Location term set and list the relevant countries under that node.

Items to think about:

Entity

• Could an entity be a customer and a supplier? Is this common? If so, perhaps have under Entity...
  • Entity type
    • Customer
    • Supplier, etc.
  • Entities
    • Customer name
    • etc.
• Do you group customers by any specific categories, alpha-numerical, 0-9, A, B, C, etc.? If so, include the relevant layer(s) in your terms.

Location

• Do you need to return content based on where you are located, where your work is done and/or its legal jurisdiction? If so, create appropriate groupings for each and ensure the purpose is clear, e.g., Nigeria office (perhaps adding the actual town, Abuja, versus Nigeria project).
• Do you need look for content by a geographical area, e.g., all work we've done in Africa, or is it only ever country specific, e.g., Nigeria? Create the geographical layer across all continents so that the term set is consistent.

Department

• How are your people tagged in your HR system and Azure Active Directory? How deep is the categorisation? Keep it aligned as best as possible.
• How specific is content and the associated volumes to teams within a department? E.g. an organisation that has a Call Centre for customers may have a Department of Customer Support as well as Support team 1, team 2, team 3, etc. Including Customer Support in the term set is likely to be sufficient for content tagging.

Subject

• Does your organisation work in different sectors, areas of expertise, on different initiatives both internally and externally influenced? In essence the 'what we do'.

Activity

• What deliverables are under in relation to the subjects? Useful to think of activities as the specific time bound piece of work done that relates to one or more subjects.

Information Type

• Are you listing types that in fact are document formats, which is automatically created for you, e.g., Excel?
• Do you need to group up types as each area of your organisation has ones specific to them rather than a flat structure?

If the term set does not work, then it does not have to be used

In some cases, the term set might not be useful for how an organisation operates or classifies its information.

For example, if a business does not produce items for specific organisations, then the Entity term set may not be useful.

Ensure someone (or a team of people) are responsible for managing terms and monitoring their ongoing use

The initial 'one off' load of terms is done as part of the installation service order.

The SharePoint Term Store Management tool can be used to manually maintain terms. They can be updated via customer created scripts if you wish to create such scripts, but ClearPeople do not supply or support those.

Ensuring people are responsible for the ongoing daily upkeep and tracking of how terms are used will ensure the taxonomy remains useful to the users.

For example, if a term (or set of terms) are not being used, then someone needs to be able to validate that and deprecate (or delete) as appropriate.