Monday, January 24, 2011

Success Stories and Predictive Analytics

Thanks to Andy Spence (@andyspence) for asking me the question, "Do you know any success stories from predictive analytics in HR?

I do know some success stories and I hear and read more and more everyday. But before I discuss the successes, I would like to define what "predicative analytics" mean. As defined by wikipedia:

Predictive analytics encompasses a variety of techniques from statistics, data mining and game theory that analyze current and historical facts to make predictions about future events.

In business, predictive models exploit patterns found in historical and transactional data to identify risks and opportunities. Models capture relationships among many factors to allow assessment of risk or potential associated with a particular set of conditions, guiding decision making for candidate transactions.

Predictive analytics is being used more and more in companies to make predictions about behavior, investments, and performance as it relates to a firm's human capital.

Here are a few examples from my experience:

1) A company needed to understand what its labor needs would be 5 years out as certain jobs were extremely hard to fill. By using historical data, the company was able to predict the number of engineers it would need at certain points in time based on past movements in and out of the organization. This information allowed them to PLAN and act in a proactive manner rather than a reactive one when it came to recruiting.

2) Another company needed to reduce cost in recruiting and improve new hire performance. By analyzing historical recruiting and performance data the company was able to create a candidate profile and an interview guide that was based on the knowledge, skills and abilities of the companies current top performers. This insight allowed the company to reduce it's cost per hire and increase it's average new hire performance rating. Productivity also had an increase as well.

3) Another company needed to analyze HR investment choices. Due to a limited budget the company needed to determine if it should invest in sales training for the organization or a new rewards program for call center employees. By using historical data and constructing a model based on the outcome of increased market share, the company was able to prove that a new rewards program would have the highest impact on market share.

In order to really get great insight, start bringing all your data sets together to see what kind of story it begins to tell. For example by analyzing your customer satisfaction data, financial data and engagement data, you can predict with a high degree of certainty what will happen to your financial results if your employee engagement scores decrease by X amount. It is important in this scenario to understand what is driving engagement with your employees so that those drivers can be tracked and continuously improved upon.

What has your company been doing in the area of predictive analytics?


Micky Jay said...

Some interesting stuff on your blog. I have just discovered it and am flicking through the older posts (with the intention of reading them in more detail later on).

Our team is interested in predicative analytics so we can take our HR metrics that one step further but it is hard to find concrete examples of how it is specifically done. While there are common measures for things like turnover or recruitment efficiency, it seems as though predicative analysis is "use the data you have to make predictions".

Now this all seems mysterious and crystal ball gazing, but are there any hard rules (or at least some rules with fuzzy edges) on what predicative analytics is about? Some of what I have read is simply forecasting (ie if turnover continues at this rate we will need X people in 3 years time). Other information is very vague and hints at predicative analytics importance without giving much away as to the processes.

We were hoping to look at things like "if we look at these 4-5 sets of data, we can show higher risk of separtaion in this gorup of people". That way we can strategically intervene rather than prepared reactivity.

Is there much out there on the hard and fast rules of predicative analytics? Even general rules of thumb would be helpful.

Unknown said...


I have not seen any hard and fast rules for predictive analytics. I do know what makes HR analysts successful as far as skill sets and I do know that you have to know your business really del to be able to analyze the right data. I think the main thing you have to do is...Ask the right questions or analyze the problem using the data you have.

In your example on those that are at risk for leaving...that is actually one of the easier predictions to make song engagement, performance and turnover data.

I would be happy to chat with you on that.

Thanks for stopping by and reading....