BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
Josh Bersin founded Bersin & Associates, now Bersin by Deloitte, in 2001 to provide research and advisory services focused on corporate learning.
He is an active researcher and industry analyst, a frequent speaker at industry events and a popular blogger. In a recent phone interview, Bersin discussed HR talent analytics with SearchFinancialApplications.
What is the recent history of HR talent analytics, and why is it growing in importance?
Josh Bersin: Most companies spend 50, 60 or 70% of their revenue on payroll. People account for the largest expense in most companies. Most of the decisions about who to hire, about who to promote, who to put into what job, how much money to pay somebody, how to assess whether somebody is a good candidate, what training to give people, are all made on somewhat of a judgmental basis, with a lot of good insights but not a lot of data.
Over the years, HR departments have struggled to rationalize the data they have about people because most of the HR technology -- the systems we use for HR -- is old and there are many of them. There is usually a payroll system, learning system, recruiting system, system for benefits and recognition. If you want to know a lot about a person, it is a big project to get the data in one place.
In the last three to five years, the technology markets have started to consolidate, and most of the new HR platforms are in the cloud. Companies have access to much more integrated data about their people than they ever did before. Thanks to books like Moneyball and other popular books on analytics, business leaders and HR leaders have said, 'Wait a minute -- why don't we use the data we have about our people to make some decisions?'
That has created a new discipline within HR which we call talent analytics or people analytics.
Why are many HR leaders becoming more interested in talent analytics?
Bersin: It's a way of taking data you have about your employees -- just like the data you have about your customers -- and doing much more business-oriented analysis on why certain things are happening and how to make decisions better.
For example, the biggest application of talent analytics everyone is interested in right now is retention. Why do top people leave our company? What can we see about the patterns of people leaving? What are the risk factors we can work on to try to make it a better place to work for certain critical skilled people?
The second one is engagement. What can we do to make the company more fun and productive and interesting to people?
One of the things that also has happened in the past five years is that companies -- Facebook, LinkedIn, Twitter -- have collected so much data about people that the HR department, or anybody who is doing recruiting, can go out and look into the market, with very little work actually, using a lot of off-the-shelf tools, and find people that potentially might be candidates to come work for your company. Those people may or may not even be looking for a job.
How do you get them to join? Are they the right people? What data do we have about the people who have succeeded versus not succeeded in a particular job that we can use to assess people better? All of those are really critical questions that now can be answered by analytics.
How many companies are actually using analytics to make personnel decisions, and why will that number grow?
Bersin: A lot of companies are starting to make very data-driven decisions in HR. I would say the trend started about three to four years ago, and now the average HR department knows that it is something they should do. About 5% to 7% of the companies we talk to are actually doing it. Most of them, about half to 60%, are still dealing with a mass of dirty data they can't get all in one place.
That is being addressed. Workday bought an analytics company, Cornerstone bought an analytics company, ADP just announced a huge bunch of new analytics capabilities, Ultimate has an analytics tool -- more of those core software companies that serve HR are buying or building integrated tools to analyze all the people data that their clients have.
A lot of vendors have access to data across many companies. They can say, 'Among all of our clients in this industry, these are the types of people that appear to be leaving, these are the types of people that appear to be staying.' And that's data that is starting to become available now to HR departments to make better decisions.
Does talent, or predictive, analytics only involve top, highly paid employees in a company?
Bersin: It kind of depends on the situation. If you are a hospitality company, a retailer or food service company and you have critical, or a lot of, employees quitting, it may not be costing you much to replace those people -- but it indicates something is wrong and takes a lot of management attention. This kind of data can be used for everybody. It can be used for anything.
What kind of fundamental problems can the technology solve for a company?
Bersin: Two things cost a company money: hiring the wrong person and then people quitting. When they quit you have to replace them; you have to recruit and replace.
Technology can help you figure out how to hire people who are more likely to stay and get rid of the things that are causing people to leave, which includes things like management practices, compensation, training, talent mobility. These are all the sort of things HR already handles, but they never really know the impact.
The average company in the U.S. spends $1,200 a year per employee on training. Some companies spend $4,000 to $5,000 per employee on training. Most of them have no idea what they are really getting for that money. They know they have to train and they know it makes people happy and more productive, but they don't really know the financial return and so they can't really optimize it too much. They can optimize it to a degree.
It turns out analytics data shows that in some jobs, failing to train people has a direct impact on them quitting.
One company we worked with did some analytics work and looked at why people were leaving and who was leaving. A lot of people were leaving in the first three to six months on the job. They found it was all about training. The training was too complex, or it was not the right training for the first nine months of the job . That data allowed them to change the way they manage a relatively low-paid workforce in a very significant way.
If a software company is trying to hire the world's best engineer, or trying to attract an engineer from another company, it would be nice to have more data about where he or she is, how to find the candidate, and who the candidate knows who are just as capable.
Where does the data come from?
Bersin: It comes from everywhere. A lot of it comes from inside an organization. Companies have a lot of internal data about people's job history, performance ratings, their training, demographics, their education. A lot of this data is outdated.
A lot of the analytics vendors, the new ones, are capturing data from the outside world and turning it into meaningful information that corporations can use.
Social networks like LinkedIn, Glassdoor, Indeed and Facebook collect data. We volunteer all this data to these websites and they share it. Sometimes they sell it and sometimes they offer it for free and sometimes they build tools and products around it.
When we go to Facebook and we do things and we produce data about ourselves, Facebook sells that data to advertisers.
That is happening in HR. If you are on LinkedIn, or any of those tools, LinkedIn is making a business out of the fact your data is there. We as individuals are more likely to have accurate data there than we are in our own corporate systems because the corporate systems don't get updated all the time.
What type of companies use predictive analytics right now?
Bersin: We did a big study called High-Impact Talent Analytics, and we have a new one coming out, Deloitte Human Capital Trends. Both of them pointed out that it is a relatively small group of companies that are very much ahead of the curve.
It is kind of a have and have-not market right now. A lot of companies are trying to catch up. There is a massive market of companies seeking to replace HR software.
Tech companies are very focused on predictive analytics, especially in Silicon Valley and New York City and cities where there are lots of tech workers. Many tech companies, in particular, are using predictive analytics because they have a big challenge with retention. Engineers have very fungible skills, and they can move from company to company pretty quickly and do the same work.
Building on traditional BI with predictive analytics
How to encourage predictive analytics adoption
Can talent management systems revolutionize HR?