The Process of Application and Data Development
This is the ninth in a series of discussions of quality guru W. Edwards Deming's Fourteen Points of Quality and their ramifications on data quality. Deming's point 5, "Improve constantly and forever the system of production and service," holds two messages: 1) Quality must be built in at the design stage, and 2) Improvement is not just a one-time project; it is a continual, unending process, involving everyone in the enterprise.
Last month I addressed the ramifications of quality point 5 for data quality improvement in business processes. This month I address its ramifications for the processes of application and data development.
Deming's fifth point of quality has strong implications for every IS organization that builds applications and databases, buys software packages or is developing a data warehouse. The key question is this, "Are all--including downstream--beneficiaries of the delivered application and information products generally happy with the results?"
Improvement Versus "Problem Fixes"
Deming attended an awards ceremony in which the highest award was presented to a man who saved his company $250,000 by discovering that labels were missing from bottles of vaccine that were about to be shipped. The second award went to someone who discovered contamination in a shipment before it went out, resulting in the condemnation of the shipment. Those awards were not for improvement of quality nor of the system, but only for putting out fires, Deming commented.1
Fixing program bugs is not quality improvement. Fixing the Year 2000 problem is not quality improvement. It is putting out the largest data quality "fire" the planet has seen to date. "Fixing" the software and databases to support the Year 2000 calendar change does not improve any business process--it only allows the business processes to be performed in exactly the same way on Monday, January 3, 2000, as they were Friday, December 31, 1999. The fix does not improve the software development process--it only puts out a fire that was entirely preventable.
Will there be award ceremonies for those who have "fixed" their organization's Year 2000 problems? Probably. Will there be award ceremonies for those who designed their organizations' databases with four-digit date fields, preventing the need for a Year 2000 project? These "Year 2000 problem preventers" are the ones who should be honored.
Many business data quality problems can be traced back to poor applications and database design. On the other hand, quality application and database design can be a major contributor to data quality improvement. Information technology is only a tool--it is not a "solution."
Information technology, however, can play an important role in business process innovation and improvement by assisting in preventing many types of "human error." Some human error can be attributed to not paying attention, fatigue, lack of training or other human reason. But much human error is actually induced by poor application and information presentation design that is not intuitive, cumbersome or forces unnecessary keystrokes and other ergonomic shortcomings.
If data quality is designed into the process and not inspected out, application and database design are a very important part of process improvement. The application represents the automated part of the business process. And the database is the data store that inventories the information product.
If data is defined from a narrow vertical perspective, it is virtually guaranteed to be defective for those knowledge workers outside the scope of those who define the data. The only way data can be defined in a way to satisfy all information customers is to define it from an information value chain perspective. In other words, data must have a consensus definition for both data producers and all knowledge workers who use the data.
Quality Improvement Steps
The information value/cost chains for data definition are the application and data development methodologies. These identify all processes that define information requirements and translate them to databases that store data and applications that create, update, delete and retrieve data.
1. Analyze the application development methodology:
- Is it defined and followed? Or is it defined and ignored? Or is it non-existent?
- Does it have a philosophy of teamwork between the business and information systems and between application development and information resource management groups?
- Does it focus on only the application for development? Or does it identify the value chain of which the application is a part and the interrelationships of the value chain?
- Does it address only benefits to the sponsoring business area? Or does it specify how the business objectives of the application directly support the enterprise mission as a beginning point?
- Does it define data requirements as the product of the application or does it focus on functional requirements as the product of the application? In other words, does the methodology differentiate the concept of performing work--the function--from the work product that has customers?
- Does it define data requirements only for the scope of the application, even though other knowledge workers in the enterprise have a stake in the data? Or does it identify the knowledge workers and downstream process owners to allow them to discover their requirements from the information products being delivered?
- Does it seek to define all application and data from scratch? Or does it focus on reuse of existing data and application infrastructure, along with developing application and database components that are reusable and sharable to eliminate all unnecessary redundancy?
- Does it include a phase for getting post-implementation feedback from both immediate beneficiary and downstream (outside the scope of the application) beneficiaries?
2. Conduct a customer satisfaction survey among business personnel of their satisfaction with applications and databases delivered within the past three years.
3. Analyze and listen to the feedback. The only way we can make effective improvements in application and data development is to listen to the "customers" who must live with the applications and databases long after the developers have moved on to "greener projects."
4. Analyze the history of business feedback, program and system problems.
5. Bring representatives from the business areas, applications development and information resource management to analyze root causes.
6. Identify potential causes and agree on the most likely cause(s).
7. Plan process improvements and implement them in a pilot development project in a controlled environment.
8. Study the implemented improvements to see if they achieved the intended improvement among all knowledge worker beneficiaries.
9. If so, roll out the application and data development process improvements and make them permanent.
10. Look for the next critical application or data development defect to eliminate.
To conclude, data quality improvement is not a one-time activity. Data quality improvement is a cultural mind-set that says the status quo is not sufficient. It is the actions and behavior that implement the philosophy of providing quality to all our customers, whoever they may be.
When I was the manager of application development at a large publishing firm, we implemented a procedure for documenting application problem events. The person "fixing" the problem completed a report called a "trouble report," describing what was done to "fix" the problem. But the most important section on this report was a "recommendation for problem prevention" section in which they identified a potential improvement that could prevent the problem from happening again. This section was the most difficult to complete--but the most valuable. It is this kind of mind-set that Deming's quality point 5 is all about.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access