How hints in your help desk data reveal larger tech-stack issues

Help desk tickets like “can’t log in,” or “computer is slow,” or “I spilled a Stanley Cup’s worth of coffee on my laptop” might offer more useful information than you might assume.

When analyzed for metrics like volume, sentiment, and seasonality, help desk data can reveal larger tech problems and warning signals within the organization—including broken processes or workflows that could use a little automation.

Categories. Before you can spot troubling trends, you need informative, well-defined categories, according to Robert Rohrman, SVP of IT infrastructure at certifications and training body CompTIA. Increases in specific ticket types allow IT pros to zoom in on specific areas of need.

“If you just see a spike in ‘software,’ what is that really telling you?” Rohrman told us. Software, he added, could mean a marketing team’s platform, Excel, or video-editing tool.

To ensure that an organization’s help-desk menus “reflect current reality,” Rohrman recommends teams review categories on a quarterly to biannual basis, adding and retiring topics as needed. Rohrman’s just recently added a help-desk option for “AI,” with subcategories including customer-facing tools, agents, and token costs. Steve Sweett, chief operating officer at IT consulting and outsourcing services firm Buchanan Technologies, has used categories like “password resets,” “device failure,” and “wi-fi outages.”

“Microsoft Office issue” might require more specificity when digging into a root cause, he mentioned. And one other category will certainly be too broad: “If your biggest category is ‘other,’ that’s a problem.”

Along with Sweett, Matt Peters, CEO of IT-request automation platform Fixify, sees enterprise-level AI platforms, featuring appropriate data protections, as helpful tools for iterating on categories once help desk ticket details have been exported. (Peters recommends about 10 top-level categories, each with subcategories; Rohrman doesn’t like to go beyond two or three subcategories.)

Other metrics to track. AI tools can also be helpful, Peters told us, in spotting “top talkers”: the one or two people doing all the reporting. A project manager sending a blast of tickets, for example, may indicate the need for a kickoff meeting with IT before any new project gets underway. If an executive assistant has the most help desk contact, the tech team may want to meet and find ways to provide more proactive help (and keep execs happy).

Peters also cited the importance of measuring seasonality. If IT pros know when tech issues peak (say, during quarter-end, when sales teams might need certain tools to close a deal), they can get ahead of the challenge by addressing needs early or adding resources.

Sentiment, which many AI and help-desk tools track today, also provides a helpful metric, Peters (and Sweett) told us. Peters said sentiment analysis recorded at the beginning of the help desk experience can help IT pros pinpoint painful processes—say, onboarding—that require reconsideration.

Automation intimations. While studying volume, top talkers, and sentiment offers a solid starting point for troubleshooting persistent tech-stack issues, other help-desk data can suggest a broken workflow or one requiring automation. Rohrman looks for two types of tickets that signal an opportunity to fix a faulty process.

First, a request for an unusual tool—maybe an Excel file merger—may indicate an overwhelmed user shouldering a task on their own, and may demonstrate a need for the resident IT pro to understand an internal process better.

Second, a slow laptop has many root causes, but one possibility could be “an overloaded process that can be automated or at least offloaded,” Rohrman said, noting the importance of finding hints of problematic workflows in help desk tickets.

“People will not open up a help desk ticket and say, ‘I have an inefficient process. I want you to help me automate it,’” he said.

And Rohrman must always be on his toes, ready to spot inefficiencies as well as everyday tech issues and warning signals. Recently, he had a day featuring eight application log-in requests from users. That sudden increase led his team to investigate the root cause and discover that their single sign-on (SSO) application’s certificates had expired.

“It’s incumbent upon us on the IT side to keep up on the technology, to understand the various tools that we have, and identify those tickets where you get a hint of an inefficiency.”