Earlier this month I wrote a post in which I likened IT metrics to piecrust. And not because they are both delicious. (Sarcasm intended.)
Like piecrust, creating metrics seems like a simple task. Yet many people go through lots of failed versions before getting it right, and some folks never manage to produce a good crust (or metrics).
Mark Tauschek, director of IT Research for Info-Tech Research Group, likened metrics to another food, broccoli, during his presentation at this week's Midmarket CIO Forum in Orlando. The broccoli analogy is more obvious than my piecrust one. Both metrics and broccoli are often reviled, yet both are undeniably good for you.
Tauschek ran through some of the benefits of employing IT metrics:
- Provide insight into IT's performance and build credibility with business units. The key is being proactive, said Tauschek, "showing the business how you are doing before they ask."
- Justify staff and spending levels. CFOs are less likely to balk at funding requests from IT departments that are in line with their peers.
- Assess and report on operational and project performance, which should help IT teams improve and create a culture of continuous improvement.
- Understand and articulate where IT creates value.
Despite these benefits, Tauschek said IT organizations sometimes resist metrics. Smaller organizations tend to object more than larger ones, he said, insisting that business units already have a good handle on what IT does and they don't want to waste scarce resources in defining and monitoring performance metrics.
Even if this is true (and chances are, it's not), Tauschek said the control provided by metrics can help IT organizations manage growth. Also, at some point business executives will want to see some metrics. "Most organizations capture some kinds of metrics. It's hard for IT to justify not doing it if everyone else is doing it," he said.
Let's say you are interested in implementing an IT metrics program. (Eighty-two percent of organizations surveyed by Info-Tech are.) How do you start? Tauschek's suggestions:
- Select three to five performance issues that you suspect need improvement.
- Create a benchmark. Compare against peers, if possible. If not, use data showing historical internal performance.
- Determine how data will be collected and metrics generated. There may be existing tools to help with this task, such as the built-in tracking and reporting tools included with many help desk applications.
- Collect and track them regularly.
- Determine desired reporting schedule, taking needs of the business into account. Help desk metrics should be analyzed a minimum of once a month, for example, while metrics around budget and staffing levels may be needed only once or twice a year.
- Determine how to report them. I found this the most interesting part of Tauschek's presentation. He advised choosing a presentation style that will appeal to your audience, keeping in mind some folks like text while others prefer visuals and/or numbers. Share the current metric value, the change from the previous period and the comparison to your benchmark. Make sure you include context and provide a call to action. For example: Don't just say IT resolved 175 help desk tickets durng the past month. Instead, note that the number is down 10 percent from last month, but 30 percent of staff time is still devoted to resolving tickets. If 40 percent of tickets were related to operating system issues, suggest that providing OS training to users could cut down on the number of tickets and boost IT productivity.
Tauschek also ran through some top-level metrics and possible interpretations. Noting that operational spending comprises more than 80 percent of typical IT budgets, he said operational budget as a percentage of revenue is one of the most critical metrics to benchmark against peers.
While results higher than those of peers can indicate overspending, this may not apply if the business is gearing up for a growth or transformation initiative. (This is where context comes in.) Complex environments usually have higher operational costs, so a figure at the high end of the scale may also show IT is too complex for the current business model.
If the figure is well below peer results, this could mean IT is performing well and does a good job of managing the available budget. Or the business may be running in reactive mode, with IT under-serving the business and just doing its best to maintain current performance.
Audience questions usually yield some good insights, and Tauschek's presentation was no exception. A CIO in the crowd asked about "better dead than red" metrics, ones that reflect so badly on IT that no IT leader would want to reveal them. Tauschek said it's better to proactively share these kinds of metrics with the business, especially if you can show they are improving, because the business will eventually want answers anyway.
"You don't want to have the CEO come up to you in the hall and say the help desk is terrible," he said. "If you don't have the right answer, that's problematic. If you know, you can work toward improving it. And you can tell the CEO you have a plan for improvement."