I think we have all been there, searching for documents, you know the document exists but without the document number you are having trouble finding it; you are having trouble searching for a group of documents that relate to your subject matter, and you are noticing that you are returned spurious and erroneous documents in your search results. Considering what you have found versus what you cannot see, it does not fill you with confidence does it!

Generally Document Management Systems (DMS) rely upon document metadata to enable users to search for the documents required, and it is within this metadata that we find the deficiencies.

1. Incomplete Metadata

On a recent project I was involved with an issue came to light that the user could not search for all documents related to a purchase order when looking in the Project DMS. Our investigation revealed that the Purchase Order details where incomplete in the metadata. You can imagine the disappointment when we learnt that this metadata field had been deemed as “optional”!

Determining which fields should be mandatory and which fields should be optional may require a wide group of stakeholders, and it might be a contentious discussion, but it is one which needs to be had.

2. Incorrect Metadata

Document numbering standards come in many different forms with numerous ways of constructing a document number code, some with intelligence built into the document string (whereby at a glance someone who is familiar with the standard can determine which Project, which Facility, or Purchase Order, or Discipline, or Document type it relates to); other standards are more about just establishing a unique number, however, these simpler numbering structures are more dependent on getting your metadata correct.

It is common in my experience to see incorrect metadata in the area of document type

classification and revision status, perhaps users are confused about whether the document should be a Report or should it be a List? Should they start with version 1 or A?

3. Inconsistent Metadata

Most menacing is inconsistent data, you have probably come across many different ways to represent the same thing, which can make searching challenging. Some systems are intelligent enough to cope, others require you to be a wiz at Boolean expressions to return all the variations.

• Australia and New Zealand
• Australia & New Zealand
• Aust & NZ
• ANZ
• Australia & NZ
• AUSTRALIA and NEW ZEALAND

Here are my 7 “first aid” tips for metadata improvement; I will be applying DRS ABCD which every first aider will be familiar with.

D – Don’t allow these problems in future submissions,We need to stop the problem from growing any bigger. Create a quality assurance workflow process to capture errors.

R – Responsibility, Determine who will own the issue and lead the clean-up effort

S – Standardise and agree on the use of abbreviations, If you are going to allow them at all!

A – Agree and enforce mandatory fields This should be enforced by your workflow

B – Build lookups to complete fields containing set values Enable the users to select values rather than typing

C – Communicate the new process with all stakeholders. This is to reinforce item 1.

D – Determine how to systematically address the documents already processed
Create plan and track the progress of your remediation. Build the activities into a new workflow or add them to an existing workflow.