The CIO of a government agency recently contracted me to write a customer satisfaction survey for the over 1,500 users in his organization. As part of the initial research, I asked to see the database of incidents logged in the online system. The CIO’s initial response was, “You can look, but I can probably tell you what’s in there.”

Some of the CIO’s assumptions were confirmed and others were shown to be incorrect by the analysis of three important statistics:

  • How many times has the help desk been used?
  • Who is using the help desk?
  • What problem types are reported most often?

This client’s database had been in place for over nine months. In addition to entries made by intranet users, the help desk and support analysts added records for all (or most) of the incidents initiated by phone, e-mail, or the “hey you!” method.

Until recently, however, no one had bothered to run any reports on the data. My message this week is: Analyze your data before it gets too old.

Help desk traffic
When the CIO conducted his first annual survey, he asked his users what method they used most often to contact IT support for assistance. Figure A shows the results.

Figure A
Survey results show intranet support capabilities are vastly underused in this organization.

Based on those results, the CIO directed the IT support department to create a Web-based intranet site for requesting technical support. Nine months after the online help desk went into production, 5,500 records had accumulated in the database.

One of the first reports I ran from that database included the number of unique users—people who had used the help desk to initiate a request. That detail (473) caught the CIO’s attention. “That’s impossible,” the CIO said. “That’s a lot of requests per user.”

So I ran a list ranking the unique user IDs by the number of entries entered. The CIO recognized the top 10 as IT people, most of whom work the help desk. Once the records entered by all IT personnel were filtered out, the resulting average number of requests per user (between four and five) was more in line with the CIO’s expectations.

The CIO was disappointed that more people weren’t using the system. On the up side, however, another report showed a respectable 70 percent of all help desk tickets were closed on the same day they were opened.

Populating the Issue Concerns column
The other report that spurred immediate action focused on the selections made for the Issue Concerns field. In this Web interface, the user opens a dropdown list and chooses a one- or two-word description from a list of predefined options. This selection is important because it routes the request to the appropriate person’s attention.

There are two challenges with this kind of help desk form. First, a propeller head usually writes the list for a target audience of fellow propeller heads in IT tech support. Second, even with well-written descriptions, some users have a hard time describing technology.

Figure B shows this statistic from the first nine months of help desk operation: When faced with a list of 30 or 40 options to describe a problem, users chose Software-Other, Unknown, and Hardware 29 percent of the time.

In most cases, the people who field those requests have to forward them on to the appropriate people, adding one more “hop” in the journey from request to completion.

Figure B
In this report, there were too many support problems that were not adequately described.

Where there’s smoke
When we reviewed the records stamped Software-Other or Unknown, we noticed a dozen or so obvious trends in the comments.

Those trends led us to some concrete fixes that would help users identity the problems they were having. For example, the CIO formed a small task force to review and rewrite the options under the Issue Concerns dropdown, with the goal of making the most common issues easy to find and select.

The updated list included new additions for many of the issues previously relegated to Software-Other and Unknown, ranging from “Computer running slow” to “Reimage workstation.”

Many options, such as Hardware, were expanded or renamed to make them more descriptive and easier to find.

Enhancements to the help desk interface mark the beginning of this CIO’s yearlong plan to enhance the quality of technical support in this organization. More changes will no doubt come about based on the results of this year’s customer satisfaction survey.

The quality of your help desk

What metrics do you use to monitor help desk performance? Please post your comments below or write to Jeff.