5 Analytics Mistakes You Need to Avoid

More than any other technology, analytics is driving business transformation to new heights. From spotting sales opportunities to optimizing marketing campaigns to understanding customer behavior, analytics has helped a countless number of businesses to turn data into dollars.

Still, for all its power and potential, analytics remains far from foolproof. Seemingly simple mistakes can lead to lost or misleading results, fooling enterprises into making critical business decisions based on defective insights.

Are analytics blunders hurting your organization's bottom line? Here's a quick look at five analytics mistakes that can be easily avoided.

1. Failing to unite analytics with business

When analytics teams are closely aligned with business units, both sides gain a deeper understanding of enterprise priorities and processes. "This positions them to conduct ongoing long-term and broad-based investigations that can produce holistic, durable solutions and solve higher order problems," explained Greg Bowen, senior vice president and CTO of Dell Digital, Dell Technologies' internal IT team. "Strong functional and contextual knowledge also enables analytics teams to create dynamic solutions that can pivot quickly as questions and situations change," he added.

Greg Bowen, Dell Digital

Too often, analytics teams are deployed as service bureaus that take ticket-based requests, Bowen noted. "This is not only an inefficient way to leverage analytics, it can also be frustrating for the business because the process feels slow and frustrating for the data scientists," he said.

2. Leaping to incorrect conclusions

Organizations should be thoughtful and systematic about how data is collected and analyzed. "Often, we draw conclusions about an issue before we look at the data," said Carin Lightner-Laws, an assistant professor of marketing and supply chain management at Clayton State University. It's okay for business leaders to formulate hypotheses or conjectures before data is examined. "However, we should hold off on making decisions until after we have thoroughly reviewed the analytics and mitigate bias, as much as possible," she explained. "Data should tell a story and lead to effective decision making."

Data collection and business analytics should be viewed as both a science and an art. "Data analytics should drive decision-making as long as we use proper data collection techniques and mitigate bias [and] pre-determined conclusions," Lightner-Laws observed.

Carin Lightner-Laws, Clayton State University

3. Underestimating the change management demands created by an analytics deployment

Empower the key business stakeholders whose decision-making processes will be directly impacted by adopting analytics technology, suggested R. Ravi, a professor of operations research at Carnegie Mellon University's Tepper School of Business. "Recognize that analytical models can expose well-hidden inefficiencies in business units and that this exposure, as well as other harsh facts uncovered by analytics, must be carefully managed," he advised.

R. Ravi, Carnegie Mellon University

4. Missing the goal

Given the time and complexity required to deploy an analytics project, it's easy to miss the planned objective, observed Benjamin Smith, senior manager, consulting, at Clayton and McKervey, a public accounting and consulting services firm. "Losing sight of the end goal can lead to a bunch of numbers on a page and lackluster adoption," he explained. "If you aren't answering relevant questions and sparking an emotional response connected to your business goals, people will fail to act, lose interest, and walk away from the project."

Clearly define the goal at the project's start and constantly revisit it. "Turn the results of your analysis, or your dashboards, into a story by visualizing the data and scripting out the actions users need to take as a result of the analytics," Smith recommended. "For example, as CPAs, the most valuable asset we have is our people," he said. "No one wants to see increasing employee turnover, [but] presenting it as a graphical representation can change the way people perceive the information."

Benjamin Smith, Clayton and McKervey

5. Missing or poor internal support

Trying to deploy an analytics project without proper enterprise support can lead to inaccurate and incomplete insights. "Data flows throughout organizations are very complex, and any data governance solution will eventually require buy-in from numerous stakeholders," said Bryce Snape, senior director at FTI Consulting, a global business advisory firm. All relevant C-level executives and department heads should be included in data analytics planning. "Involving them in the process early on and helping them see the value the [analytics] tool will bring to their current challenges, will lead to a higher rate of success,” he noted.

Takeaway

As with any disruptive technology, it's inevitable that mistakes will be made during initial deployments. "The best way to repair any damage is to take a step back, take a deep breath and come up with a coherent strategy before diving in," Snape advised. "Lay out the objectives, timeline, available investment, and relevant stakeholders."

Find quick wins to show progress yet be realistic about the effort it's going to take. "Don't hide from the hard conversations," Snape added.

For more on data and analytics strategy, follow up with these articles:

IT Disappoints Business on Data and Analytics

The Best Way to Get Started with Data Analytics

Why Everyone's Data and Analytics Strategy Just Blew Up