Why gut feelings will no longer cut it in human resources...
Experts say businesses that want to land the best people or hold onto their top staff should start getting serious about using data analytics in the HR department. Nick Heath reports.
Google and other large US companies have boosted their business as a result of analysing how they manage their workforce, according to Competing on Talent Analytics, a recent report in the Harvard Business Review.
Report co-author Tom Davenport, president's distinguished professor of IT and management at Babson College in the US and director of research at the International Institute for Analytics, spoke to silicon.com about how HR departments should follow in these companies' footsteps and tap into analytics.
"HR is among the last area in business to be penetrated by analytics, but it seems to be coming on in terms of identifying talent that is a good fit with your company or valued employees who are likely to leave," he said at the recent SAS Premier Business Leadership Series conference in Las Vegas.
While many companies already collect the data needed to carry out basic HR analytics, such as levels of staff training or time working for the company, few are using it to its full potential according to Davenport.
"Many companies have the basic data in place and the question they are now facing is, how do we use this to manage our business better?" he said.
"Maybe 10 per cent of large companies are carrying out some variation of [data analytics]. The time is right for HR organisations to say 'Let's add some value'."
In the long term, Davenport believes that companies will need to use HR analytics to keep up with their competitors, predicting that half of all businesses will be using analytics in the HR department within 10 years.
Identify individual or company-wide performance
The simplest way that a HR department can use data analytics is by using a couple of datasets as a measure of workforce performance in certain areas, according to Davenport - such as measuring levels of employee satisfaction by analysing employees' willingness to recommend their office to others as a good place to work.
Analytics can be used to ensure that businesses achieve concrete benefits from their HR spending. By using quantitative analysis, Davenport said Google has been able to reduce the number of interviews it puts new employees through.
"They used to do 10 and now they are...
...closer to five. They looked at the number of interviews that were associated with a new employee and the performance of that employee.
"They found that after about five interviews the employees didn't seem to get any better, and also that some potential employees dropped out of the process when they were doing 10, as they just got frustrated."
Similarly, food distribution company Sysco has improved staff satisfaction and reduced attrition levels by introducing management practices that have been successful elsewhere in the business. The processes were identified by looking for correlations between data such as supervisor effectiveness and workplace climate, and measures such as staff productivity and retention rates.
Davenport said Sysco's analytical approach to staff management had saved it "substantial" sums in retraining costs by reducing levels of turnover among its customer service representatives.
Find new ways of predicting staff performance
By collecting and analysing a large amount of data about staff members, businesses can identify more accurately who is likely to be a high-flier or an underachiever.
"Start measuring everything about an employee," Davenport said. "You are not sure what factors are going to be associated with high performance, so you need to gather almost everything you can think about."
"Google gathered a large amount of bio data for a small sample of employees. At Google the factors that became the best predictors of high performance were 'Have you ever started a non-profit organisation?' or 'Have you ever set a local or national record?'. Those are not things that appear in the average HR database.
"Once they found out about these things they produced an online questionnaire asking about these factors - it changed the whole process."
Get more out of individual employees or departments
Analytics can proactively identify those staff or departments that need special attention from managers, helping to reduce attrition rates and improve performance.
"Google analysed the staff who accounted for the bottom five per cent of their performance levels with the belief that they put a lot of energy into hiring these people and maybe there was something wrong about the process," he said.
"They found that many of them were...
...quite capable and were poorly assigned or had a conflict with a particular manager, and once reassigned they did very well."
Predict the future staffing needs of the business
Businesses can also forecast future challenges, such as skills shortages or workforce turnover, by analysing historic or current data.
"At the top end of analytics you get into the talent supply chain - identifying what might be the needs in the future. For example, looking at when people are likely to retire or how many people in a speciality area are coming out of schools with skills in those areas," Davenport said.
"An example is Lockheed Martin, which has a quite extensive HR database, and so can answer questions like how many C++ programmers do we have in this business unit, and compare it with their strategic objectives."
This approach can also be used to predict staffing demands in real-time - for instance, to calculate the hours that staff should work the next day, based on work patterns and demand on the previous day.
Davenport added that this kind of "big picture thinking" in the use of HR analytics is still relatively rare.
The wrong way to do HR analytics
A common way that businesses get HR analytics wrong, Davenport said, is by not applying it to staff at all levels of the company.
"A lot of companies only do it for junior level people - they consider management somehow exempt. It's a little hypocritical to say we can't begin to analyse these higher level management skills," said Davenport.
Data analytics should also not be thought of as a wholesale replacement for human judgement, he said.
"It is possible to get carried away. One company digitised all the CVs for executive assistants, converted them into a score and ended up hiring based on this score. The comments from the executives being supported were 'These are some of the worst hires we have ever seen'.
"If you only go by analytics and never think you need to meet the people it is a recipe for disaster. You need to use both your human intuition and analytics."