Last month, Forbes published Portrait of an HR Data Analyst, which lays out the case for HR analytics, the skills required to excel at it, and the reasons people with these skills are in short supply. The article is one of many that urge HR to beef up our analytics skills, so that we can better use the information that accumulates at a faster and faster pace. Of the skills these authors discuss, however, one appears to be missing: the ability to evaluate data to confirm its quality, accuracy, and relevance before we use it to make critical decisions.

The Power Behind HR AnalyticsExperts tell us that 90 percent of the world’s data was created in the last two years (a relationship that apparently has been true for the last three decades). Quoting that figure, Analytics Week concluded that “companies have access to an unprecedented amount of information: insights, intelligence, trends, future-casting. In terms of HR, it’s a gold mine of Big Data.”

Some of it is, but some of it is fool’s gold, information masquerading as truth. The Internet and social media in particular have made “good data” more and more difficult to distinguish because everyone is an expert. There are no barriers to entry, no certifications required. Too often, we forget that and accept what we read without question—and then pass it on to others.

Perpetuating “bad data” didn’t start with the Internet. Take, for example, the allegation that seventy percent of large-scale transformation initiatives fail. That infamous statistic first appeared in Reengineering the Corporation, the book published by Michael Hammer and James Champy in 1993. Through the years, other change experts—among them Beer and Nohria (Breaking the Code of Change) and John P. Kotter (Leading Change, Heart of Change)—endorsed the seventy percent failure rate in some form or fashion. It became embedded in the management consulting literature until 2011, when a few change management contrarians decided to debunk the myth.

Dr. Mark Hughes of the Brighton Business School traced its evolution from Hammer and Champy, source by source, quote by quote, and concluded, “Whilst the existence of a popular narrative of 70 per cent organizational-change failure is acknowledged, there is no valid and reliable empirical evidence to support such a narrative.” In other words, the number was not derived from a controlled, scientific study. It was someone’s opinion or someone’s guess. (Hughes added a few words about “opportunistic business consultants” who may have promoted an exaggerated figure to sell their consulting services.)

Other change practitioners joined the discussion. “Change Whisperer” Gail Severini, Jennifer Frahm, founder of the Australian group Conversations of Change, and others attacked the claim as totally lacking in evidence. “Nothing to support it,” Frahm sums up. “No mention of where this fact has come from.” And yet businesses made decisions based on that fact. Some postponed, scaled back, or canceled transformation initiatives and others invested in change management activities, all based on a dubious statistic.

Here is the lesson for HR: While it probably didn’t matter whether 40, 50, or 70 percent of change projects fail—any of those figures would have earned the business world’s attention—HR can’t afford to give its business partners bad data. Tasked as we are with substantiating important business and talent management decisions, we must ensure that the data we use—from our own systems and records as well as publicly available sources—is reliable, up to date, and relevant so that the conclusions we reach are sound. That isn’t easy, especially when the pressure is on to move quickly, but much is at stake, namely, HR’s contributions to the business, not to mention HR’s somewhat fragile reputation.

The “hybrid skill set” of computing skills, HR knowledge, and business expertise will get a data analyst only part of the way to successful problem solving with workforce analytics. He or she also must learn to recognize the good and the bad in data, and to understand what makes it so. Here are some ways to achieve that:

  • Become familiar with the fundamentals of workforce analytics, its value and its pitfalls. Start with this Beginners Guide from Talent Analytics and then go deeper—take a course, attend a conference, and see how other HR organizations are using analytics to solve business problems.
  • Identify a few internal business problems—such as reducing turnover in key positions or identifying the characteristics that make sales representatives successful—and brainstorm with colleagues or professional communities to find ways to solve them with data.
  • Get on with the job of understanding the data available internally and making it pristine. Don’t limit these efforts to HR data, because the real power of analytics comes from combining HR information with business data. Engage IT professionals in an inventory of all relevant data sources and a review of how well they are integrated, identifying strengths, potential gaps, and sources of errors.
  • Work collaboratively with other functions. A trusted colleague from Finance, IT, Marketing, or an operating unit will have a different perspective—and may introduce data that HR is not aware of. Likewise, a second set of eyes to review assumptions, methodology, and conclusions may result in an angle or opportunity that otherwise might have been missed.
  • When it comes to external data, verify the sources used to build or buttress business cases, for example, and develop a list of trusted sources based on their credentials, experience, and history. Never accept a statistic unless it appears in more than one place.

Foundational skills are essential to become a good data analyst, but advancing to the next level requires the ability to differentiate good data from faulty data and rigorous, complex analysis from superficial opinions. That is the basis for informed decisions that fine-tune strategy, improve financial results, increase employee engagement, and enhance employee productivity and retention.

Change Management EBook