“Judgments made in the first ten seconds of an interview can predict the outcome of the interview. … The problem is that these predictions based on the first ten seconds are useless. … We make hasty judgments based on the observation of the slightest interaction, under the influence of our existing prejudices and beliefs. Then, without realizing it, we change, and from then on, we do not try to evaluate the candidate, but hunt for evidence that supports our initial impressions.”
These are very honest words from László Bock, former HR manager at Google, which can be read in the book “The Google Secret” also published in Hungarian. Yes, this is the László Bock who, as Google’s Senior Vice President of HR, revolutionized the workforce selection processes.
Later in the book, he describes the strategies and processes they used at Google to avoid the pitfalls of recruitment and selection: he talks about the templates used, team interviewing, and last but not least, the importance of data and artificial intelligence.
How do we use data and algorithms for more effective recruitment and retention?
During job interviews, it is often the case that the interviewer's personal liking determines whether or not a candidate is selected. However, in many cases, gut feelings and associated qualities do not reflect reality, so the new employee ends up failing to meet expectations.
Nowadays, more and more companies are trying to put their processes on an objective basis, in which behavioral elements previously played a predominant role, and which can now - thanks to technical developments - be rationalized with various analytical and algorithmic tools.
László Bock has used statistics and research results in his HR solutions. The leader swears by analysis, and thanks to his pioneering work, scientific methods are now used in all organizational decisions at Google, including HR. Google has a separate division dedicated to HR issues, and experts at PiLab (People and Innovation Lab) research what makes employees happy, what form of reward they value most, and what is needed for leadership development.
Companies that recognize the need and importance of data analysis work on the following main components:
- Large amounts of data collected from various sources.
- Various analyses and reports.
- Artificial intelligence (Machine Learning) based algorithms that help with understanding and prediction.
How can domestic companies also apply this practice in their daily work?
Most companies use very little data in their interviewing process, and most often only take it into account to a small extent. Yet companies have a lot of data about their employees. Such data can be, for example, data stored in the CRM system in the case of retailers, which reveals a lot about the “operational” characteristics of individual employees, or even production line data in manufacturing companies. This historical data can be incorporated into the selection process, which can be used to filter out those applicants who – whether based on master data or behavioral data – will not spend a long period of time at our company, no matter how sympathetic they are to the interviewer.
Based on the above, let's take an example of a company that has one or more databases that contain information about the experiences and employee behavior of its current and former employees. Such information could include educational background, previous work experience, salary level, distance from home to work, or even the reason for leaving a particular job.
However, data alone is not enough, because without the right resources and methods to connect them with each other and with existing processes, we cannot take advantage of our information advantage. Data can help us in this case if we can draw useful conclusions in the selection-recruitment-retention processes based on it.
As László Bock has stated several times, he placed great emphasis on having many analysts and statisticians in the HR team. This allowed for a much more focused focus on filtering out the distorting factors that are the results of the strange game of the human mind. These distortions were tried to be examined and weeded out by using and analyzing data extensively. As in the case of data, most companies have great reserves in reporting solutions that could be exploited much more effectively.
By connecting these databases, clarifying, organizing, and analyzing the data, Azure Machine Learning's predictive algorithms can show how likely and how well an interviewee with similar parameters, i.e., similar qualifications, working in a similar salary range or shift, will succeed in the organization, using the example above.
Nowadays, not only in the case of reporting tools, but also in Machine Learning, companies have everything at their disposal to start on the path shown by Google. After all, it is now possible to move towards decision-making supported by artificial intelligence with zero investment – for example, with the help of the Azure Machine Learning service.
From here on, it is up to the company to decide how much to rely on human instincts and how much on statistical probabilities. Of course, human intuition and experience cannot and should not always be excluded, but analyses based on data and with the help of algorithms can greatly facilitate a company's life. Last but not least, HR expenses can be rationalized, such as the costs of recruiting, selecting and retaining employees.
Once we have chosen what data we will work with and have understood our data through analysis, all that is needed is a leadership decision to get started. The Machine Learning for Business workshop, jointly launched by Training360 and BCS Consulting, helps companies quickly find a solution to the business case presented above and similar ones, and put together all the necessary elements to get started.