CASE DESCRIPTION

Most of the problems associated with the use of the CallOnUs knowledge base are related to poor data administration. As stated by McNurling and Sprage (2002), challenges of managing information resources derive from problems associated with ambiguity of language and undefined structure. They find that in the world of data administration, there are several tasks: (1) clean up of data definitions; (2) control of shared data; (3) management of data distribution; and (4) maintenance of data quality (p. 214).

Data Definition

Data definition refers to the consistency of definitions used in the logging or input of data. Any IC staff member with access to CallOnUs can create a knowledge document, which requires the entry of between two and eight keywords. IC staff and end users use these keywords to search the system when trying to resolve problems. Each staff member has the freedom to use any words he thinks are best for the document. For example, to describe a common problem with an operating system, the terms Windows95, Win95, Windows 95, Win-95 or Windows-95 are used. Because of the multiple terms that can define a problem, numerous incident solutions can result from using the full search query and the user has to read all of them to find his or her particular answer.

At the Web site, users experience similar problems. When using the search engine, they can create a query by subject or problem symptom. However, this too takes several tries because employees creating knowledge documents use many different keywords to describe the same problem. As a result, users do not retrieve all of the information available in the knowledge base applicable to their problems and the information retrieved is often not helpful. This creates a situation where a user prefers to call the IC for help rather than solving the problem alone.

Control of Shared Data

The CallOnUs knowledge base is used and maintained by all employees in the Information Center. When an employee is trying to answer a user's question, he will spend a few minutes locating the documents that will help to solve the problem. Because of the lack of standardized data definitions, there are occasions when he will not be able to find the exact solutions. If he cannot find an answer in the knowledge base, two things can happen. One is to ask other employees who are also providing help desk support. This is common because many employees are frustrated with the knowledge base and rely on each other's knowledge. If none of them knows the answer, then the employee will inform the user that he will try to find a solution and will call her later. When no solution is found, the employee then proceeds to create a document outlining the problem and solution that he was able to recommend.

Management of Data Distribution

As a result of the hospital having technical support staff in other units aside from the IC, there has not yet been a consolidation of documents. Each unit maintains its own knowledge base system. One of the problems associated with this set up is that there have been many occasions when one of the units was not able to find a solution to a problem for which the IC had already found an answer. This means that whenever an independent unit faced a problem that it could not resolve, an employee would call the IC for help.

Maintenance of Data Quality

As stated by Fisher and Kingma, "Data quality is one of the critical problems facing organizations today" (2001, p. 101). It has been estimated that error rates in industry are as high as 75% (Redman, 1998). According to the Data Warehousing Institute poor customer data quality costs U.S. companies $611 billion a year (2002). The metrics often associated with data quality are: accuracy, timeliness, consistency, completeness, relevancy, and fitness for use. Table 4 provides a description for each of these variables

Table 4: Definition of Data Quality Variables

Variables

Definition

Accuracy

Lack of errors, the data conforms to a real world value of fact

Timeliness

Data is not out of date

Consistency

The data represented is the same in the entire data collection

Completeness

The data is represented in all of its degrees and variables

Relevancy

The applicability of the data to the particular situation at hand

Fitness for use

The data is represented in a format that best serves the user's purpose

Because each employee is able to document the solution to the problems that he faces when supporting users, there have been multiple documents answering the same question. The simple addition of documents that solve problems that have already been logged contributes to the problem of retrieving these documents in subsequent periods. Differing levels of expertise have further led to the creation of documents that provide inefficient or inadequate solutions to a problem. When an employee proceeds to provide support, he may be making things worse in the long term by giving solutions that could have negative effects on the overall performance of the system. This in turn leads to other problems.

Help Desk Support Flow

Employees at the IC log all incoming phone calls, e-mails, or walk-in visits into the CallOnUs system as incidents. When a user makes a request, the employee searches the knowledge base several times with a variety of keywords in an attempt to resolve the customer's request. If the incident is resolved, the employee documents the solution in the knowledge base. If the incident is not resolved, the employee forwards it to the appropriate queue for higher-level support. The request sits in the queue for up to three days until an employee attempts to solve the problem, or until the customer contacts the center again inquiring on the status of the problem. The time lag occurs because no one at the center is responsible for responding to questions left unanswered by other employees. It is only due to the unrecognized efforts of some employees that these requests are eventually resolved. If the customer contacts the center again, one of two things will occur. First, the employee will look for the incident report or start the resolution process with the customer again. This process includes searching the knowledge base again. Alternatively, the employee can forward the incident to the next level of support in another queue if he cannot resolve the incident. If the incident is resolved, the employee documents the solution. Alternatively, due to a lack of communication, the incident may be logged a second time and then two different employees may try to solve the problem, thus wasting staff time. The process repeats itself each time the customer calls back. The tension between the customer and the employee can escalate with each callback. Figure 2 shows the steps in the resolution process.

click to expand
Figure 2: Incident Flow Diagram

Support Over the Phone

If the user cannot find an answer on the Web site, they can call, e-mail, or walk to the Information Center. Unanswered calls and busy signals have created customer frustration, but the staff is unaware of reasons why calls are missed when staff is available. Neuma Adams, the head nurse in the radiology department, recently sent Horton a written complaint detailing the days and times her department placed support calls. She followed up with a phone call that stated: "When are you going to do something about this? We are all too busy to deal with these kinds of tech problems. We need your department to get things done to help us." Horton was unable to provide anything to her other than an apology.

Often, the employees are on the phone for extended periods with customers trying to solve customer callbacks on previously logged problems. Because of these long calls over unsolved incidents, no one is available to answer incoming calls regarding new incidents. Frustrated clients hang up and send inflammatory e-mails to Horton. He is not sure how often this happens because the phone system does not log hang-ups or missed calls. He does not know when the peak call periods occur or who is making the most support calls because the phone system provides no metrics. Employees are increasingly becoming frustrated and accusing each other of not "pulling their own weight" with regard to solving incidents when they are first logged into the system. Two of his most experienced front line employees left, and their replacements also left two weeks later.



Annals of Cases on Information Technology
SQL Tips & Techniques (Miscellaneous)
ISBN: B001KZAZTK
EAN: 2147483647
Year: 2005
Pages: 367

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net