Last year, I received a request from a reader to speak before a regional SHRM convention. She had read my articles about how to reduce large-scale turnover by using a process called “data mining”. Data mining works by extracting knowledge hidden among application forms and pre-hire tests and using that information to reduce turnover.

The Offer
The representative asked me to speak in exchange for the opportunity to meet and greet HR participants. Initially I said, “No”. Then I reconsidered. If she could find one large local organization I could use as a data mining case study, I would speak.

I’ll skip over the technical details, but the case-study process would have been pretty straightforward: the client would give me a spreadsheet containing pre-hire data about employees and whether or not they terminated. Security was not an issue because names weren’t important...just a list of pre-hire data and whether the employee turned-over. I’d do everything else.

After the analysis was complete, I would give the client a formula that would predict the probability any new applicant would stay or term. The HR department would have a way to actively choose between two equally-skilled applicants based on who would stay and who would leave. Viola! A sound method for reducing turnover!

Oh, yes. I would also do the whole project for free.

Does Data Mining Work?
Almost anyone who has ever studied a graph to make a decision has used a simple form of data-mining. Now imagine computer-analyzing hundreds of graphs and combining the data from all of them. That’s the principle behind data mining. It’s rare among HR, but more wide-spread than most people know. For example:

A credit card company in Britain estimates they saved over a million pounds each year. A Japanese retailer tripled profitability. A British mortgage company earned an additional 50 million in revenue. Other applications include crime analysis, predicting course attendance, detecting fraud, improving customer relationships, evaluating program effectiveness, and so forth.

Yes. Data mining works. Yes. Data mining has been widely applied and accepted by organizations around the world. No. HR departments are not among the world leaders.

Is Turnover Worth Fixing?
I don’t mean to make turnover sound simple. It is anything but that. In fact, turnover often has three components: 1) the employee; 2) the direct manager; and, 3) the environment. For simplicity, though, we’ll just focus on the employee.

Adding together low productivity, agency fees, staff expense, training, lost time, litigation, down-time, customer service, management time, and so forth…the cost of turnover comes right off the bottom line!... Anywhere from three months salary for a simple job to 300% of annual wages for a high-level manager or professional.

In financial terms, a 1000 person organization with a 3K average monthly salary would have a 350 annual million payroll (1000 x 3K x 12 months). Now add a 30% annual turnover rate. A few calculations later, we would see turnover hemorrhaging from 2.7 million (300 x 3K x 3 months) to 32.4 million (300 x 3K x 36 months) every year!

But let’s be conservative. We’ll estimate turnover at 10 million annually. Now…here comes a person offering a solution that could potentially cut turnover a minimum of 25% (i.e., from 30% to 23%) or about 2.5 million dollars EACH YEAR. Whoa! That must be expensive! What’s he asking for? A big consulting fee? A percentage of the savings? Nothing! Free! Nada! Just have a clerk compile data and send it to him in an XL spreadsheet.

Weigh the risk: about 5 days of clerical work days versus gaining a couple of million every year. But what if he could not save 2.7 million each year? What if he could only save 500,000? Senior executives still pay attention to that kind of money. And, operations managers probably would appreciate less employee turnover hassle. It’s a no-brainer, right? Well, not among HR managers in this neighborhood.

Believe it or Not!
Four times eager HR staffers initiated long conversations with HR VP’s whose companies suffered high turnover. Each time their VP of HR said collecting the data would be too much work!!! Impressed? I’ll be their senior management would be.

I can almost hear the company President, “Let’s see, Bill… you had a chance to earn a couple mil for the investors without doing anything different? All your department had to do was supply a consultant some anonymous data that should already be in your files? He’ll do the work for free? And, five days of temp wages would have cost how much?....Ummm…Go back to your office and fire yourself for being silly and shortsighted!”

Looking Ahead
Today’s HR department is in a dilemma. On one hand it seeks credibility and authenticity. Hardly a month goes by when another SHRM article advocates the value HR brings to the organization…But, more often than not, these folks are writing their own press releases. In practice, HR seems to prefer a life of reactivity and paperwork.

There is an implicit understanding when a department renames itself, i.e., that the name reflects its contribution to the organization. One hundred years ago, there was no such thing as a department devoted to employees. Line managers hired and fired their own staff. Gradually, paperwork, benefits, and automation led to the need for a Personnel group to handle all the mundane activities associated with employment. Then, a few years ago, Personnel renamed itself “Human Resources” as if to imply a higher and better contribution to the organization. Unfortunately, although the name changed, their basic activities and reactive legacy did not.

Universities have entire libraries and disciplines of sound information advising better ways to hire and manage employees; how to understand and use employee data; and, how to fine-tune the organization. However, like the SHRM members in our example, they seem reluctant to step outside the box and look at employees as a manageable resource.

What kind of strategic-orientation does your Human Resources department have? Is it willing to embrace new technology to improve the quality of human capital? Are hiring, performance appraisal, and training coordinated to work together? Does it work to reduce or eliminate organizational obstacles to employee performance? Does it fully conform to best practices as outlined in the Uniform Guidelines on Employee Selection Procedures?

Most importantly, can it make contribution to the bottom line? If not, Karnack sees a strong possibility of outsourcing in its future.