There was once a man who ran a restaurant. He was making steady income. Business was booming and he knew all his customers by name. Then one, fine morning, a young customer walked into his eatery and said, “Hello! I’m a data scientist. How would you like me to analyze your customers and help you understand how to make your sales better?’
‘That sounds great!’ said the owner, who was actually not keen to give a try. So off went the data scientist, collecting a ton of data about who was visiting the restaurant and why, and then a couple of weeks later, he came back with the answer. ‘So I have an insight for you!’ he told the excited restaurant owner. ‘I can tell you with statistical certainty that the primary reason people visit your restaurant is because they really love your food.’
“But I know that already. What about the people, who don’t love my place. How do I best understand why they want to try my masala tea?” said the mystified owner.
‘Ah! For that I need to delete all the people who really love the place and run my model again’. So off went the data scientist to crunch his numbers and he returned a week later. ‘OK, I re-ran the model and the results are astounding’, he said. ‘What have you found?’ asked the owner, with anticipation.
‘I can tell you with strong statistical likelihood that, in the sample of people that don’t love your restaurant, people buy your food because they are hungry!’
‘Obviously’ said the owner. But what if you don’t like my restaurant and you aren’t hungry?’‘Give me another week!’ said the data scientist, and he went off again and deleted more people from his data set, and re-crunched his numbers.
‘So it’s likely that the others came to your place because it is a Sunday and they don’t want to travel far ’ he told the owner on his next visit.
‘I don’t mean to be rude’, said an exasperated owner, ‘but assuming that someone does not love my food, isn’t hungry and it’s not a holiday, why are people coming to my restaurant?
‘Hmmm, well, taking all of those people out, it’s actually only one person’, said the data scientist, ‘and he thought the food was free!’
Go you get the drift?
Machines can’t do what human beings can. So does that imply that we can’t programme computers to do things by themselves!
The ability to learn is the hallmark of intelligence. Transferring that ability to machines sounds like a huge leap forward in making machines smart and intelligent.
Machine learning, broadly defined is a sub-set of computer science that automates learning to enable pattern recognition and predictions. ML is increasingly being applied to predict human behavior (say buying behavior on a site like OLX or Quikr) and make labour forecasts and predict money movements.
Machines have long been used for organizing huge pools of data. Now, with the ability of providing recommendations and decision-making assistance to humans, their applications in making human life easy, have multiplied. In recruitments, ML can be used in findings a good match between a candidate and an advertised position and subsequently using that information in making a sound hiring decision, according to HR experts.
Google has just added a job feature to its search engine that uses a slew of ML tools with the help of algorithms that predict which candidates are most likely to succeed in their new role and predicting their retentive value to the employer.
JP Morgan is apparently one of several financial institutions that has put into place algorithms that can survey employee behavior and identify “rogue employees” before any criminal activity takes place, an obviously more insidious form of attrition with dire consequences—watch the interview with Bloomberg Reporter Hugh Son as he discusses these new safeguards with Bloomberg Technology.
ML is capable of analyzing different sources of data to find patterns that cannot be studied with simple spreadsheet analysis. Think of the many unknown reasons why one person loves his role, while the other detests it, and given the next best opportunity won’t hesitate to quit. By working with a diverse set of variables –age, qualification, salary, job satisfaction etc., - machine learning algorithms can delve into the humongous amount of data produced to uncover key drivers of employee engagement.
According to Bersin by Deloitte’s Human Capital Trends, in 2016, 51 percent of companies correlate business impact to HR programs, up from 38 percent in 2015. Forty four percent use workforce data to predict business performance, up from 29 percent last year. However, only 8 percent rate themselves as doing an ‘excellent’ job.
In 2017, predictive analytics in HR at organizations has matured further as companies are getting sharper at collecting and processing big chunks of information (Big data analytics). Predictive analytics strategy is being implemented a across HR functions.
Although this is likely to cut labor and eliminate some roles that HR professionals performs, on the plus side, it will enable HR professionals to begin to make more data driven decisions in the areas of recruitment, workforce management, and employee engagement, reason experts.
ML and AI are fundamentally changing the role of HR within an organization. Shape up or get shipped out of the race to automate.
Get Started !
Schedule a demo with EasyHR to experience ease of use and how an enterprise payroll software could be simplified.