Information Quality Metrics Workshop
Objective of Meeting
Few organisations know how much poor quality information is costing them. More and more
people are spending more and more of their time dealing with information. The poor quality
of that information is often accepted as a fact of life. The result is that employees
waste time, and management makes bad decisions. Globally, this probably adds up to
trillions of dollars annually. But no one knows how much, or what to do about it.
There are no standard metrics for information quality. A widget factory can measure the
volume of raw materials used, and the number of widgets produced, and can calculate
process efficiency. Everyone can understand the calculations and the results. But this is
not the case for an information factory, which is what many organizations are nowadays.
Vendors offer new products, and CIOs plan changes, but the potential improvements are
usually assessed using methods that are ad hoc and subjective. Standard metrics
would help vendors and CIOs to evaluate and compare different options. They would enable
ways of improving information quality to be found.
The purpose of the workshop was to:
- Review the business and technical environment
- Look at the human and computer actors
- Produce specific requirements for information quality metrics
- Survey existing techniques and discuss their relevance
Summary
The workshop investigated the requirements for Information Quality Metrics, using the
Business Scenario technique. It identified the following objectives.
- Improve project times
- Improve production costs
- Improve production quality
- Increase customer satisfaction
- Reduce the time taken to find a piece of unstructured information
- Reduce the time taken to build an interface
- Reduce the time taken to perform an information operation, such as indexing
- Reduce time taken to get data updated
- Increase probability that decision-makers have the information they need
The workshop then briefly surveyed existing art, in particular the Total Data Quality
management (TDQM) approach. This approach is soundly-based and can be used to measure data
quality, to estimate the cost of lack of quality, and to apply statistical process control
methods to improve quality.
There is consensus among practitioners on the principles of the TDQM approach, but the
details are not standardized. Evidence from the workshop suggests that it is not widely
deployed.
More seriously, its principal application is to structured data. However, the majority
of information in enterprises today exists in unstructured form: in email, notes,
spreadsheets, documents, etc. Unstructured information now presents by far the most
serious quality problem.
This problem requires new approaches to information quality.
Outputs
The output of the workshop will be published as an Open Group Business Scenario,
following validation by the participants and further cycles of input, review, and comment.
Next Steps
Workshop participants will draft a charter for a new Forum to address the Information
Quality problem. This will be presented for discussion and approval at The Open Group
conference in New Orleans, October 18-22, 2004.
Links
Information Quality Workshop Presentation
|