The quantity of daily generated data is tremendously growing. They are stored even longer and in more dematerialised devices.
This evolution impacts every area of the economy and paves the way for the use of “big data”. The recruitment sector is not aloof of this strong trend, even though it tends not to be as “big-data-soluble” in Africa as it is elsewhere…
Big Data: what a big deal!
The following figures can illustrate this situation that also concerns Africa:
- 90% of current data were created during the last two years only;
- In 2014, Gartner Inc. predicted an 800% increase of the data to be treated within 5 years;
- In 2013, Twitter generated 7 terabytes of data and Facebook 500 terabytes every day; e.g. 1,000 GB, the equivalent of hundreds of trucks of those old 3.5 disks…for those of you who knew them.
As the cost for data access, acquisition and treatment tends to considerably decrease, companies in some African countries are more and more showing interest in big data as a tool to improving productivity and profitability.
“Big data” is the terminology employed to indicate a set of huge data that cannot be treated through the use of classic database management tools. An organisation is able to quickly capture, treat and analyse massive heterogeneous and changing quantities and content as well as rapidly extract various relevant and analysed data by using a combination of technologies, architectures, tools and procedures at a competitive price.
Research conducted by IBM and IDG Connect shows that five African countries, namely South Africa and lagging quite far behind, Nigeria, Kenya, Egypt and Morocco, are leading in terms of having the ability to operate big data solutions on a large scale nowadays. This explains why Cape Town, South Africa, hosted in November 2013 the first Big Data Congress which aimed at raising awareness and promoting this new concept in Africa; 2,000 people attended the event of which 150 were CEOs of strategic companies of the sector.
It’s a concept that‘s been used so far by the distribution, marketing, industry and finance sectors. HR and recruitment departments could be next in exploiting it, taking into account the specificities of the African market.
Sandra K. Johnson, Head of technology at IBM Africa, confirms that Africa’s potential as far as big data are concerned is enormous; the concept of Big Data has taken a unique form in Africa, compared to other parts of the world: “most of the external data that we extract are generated from a single source: mobile devices. However more than nine millions of new mobile subscriptions are registered every month in Africa, representing exponential but complex data to analyse”.
Concrete evidence of recruitment through Big Data
HR Department of an SME specialised in telemarketing, Nairobi, Kenya.
This company’s growth is developing but it experiences a high turn-over rate of his key staff members.
The Director enters his HR Manager’s office and addresses him in a very angry tone of voice: “Who are these salesmen you’ve just recruited? None of them has a call centre or sales background, or even took a sales training. It does not look like they want to make a difference! I’m not even talking about artists among them!”
The HR Manager calmed his boss down and explained he was trying a new recruitment method. He asked him to give him six months to observe them and promised to limit the experience to these 10 newly recruited people if it was not conclusive.
Before the six-month observation period ended, the Director enters the HR Manager’s office again, this time with a big smile on his face. He shows him the results of hundreds of salesmen for the last trimester: 3 of the newly recruited sales staff were in the leading group of the 10 best performances and the remaining 7 were among the 20 best trimestral performances. Even better, none of them resigned; whereas usually 20% of new staff resigns after only 4 months of employment.
People Development did not witness this episode that could have also occurred in Nairobi, Cairo or Rabat, countries where big data are progressively entering companies’ key growth areas.
The fact of using big data for recruitment, also referred to as “talent analytics”, may give very interesting…and surprising results!
What has really happened in our little story?
Our HR Manager must have noticed how classic recruitment criterions are limited in terms of guaranteeing sales agents’ performance in this specific company, as they are often influenced by personal experience and the Director’s subjective opinions.
He then decided to use a specific methodology that allows for the enterprise to give more accurate and successful prognostics for newly recruited staff and for a higher retention rate of more than two years. Guided by the intuition that generally employees who reach or go over 100% of their targets have often several similar characteristics, he hired a recruitment agency specialised in collecting and analysing this type of data.
Huge quantities of data were gathered and cross checked from the targeted staff category within the company and with competitors located in the same geographical area over several years. They surprisingly revealed that employees who followed short sales training, had previous experience in telemarketing and are considered as supposedly having this strong desire to succeed that we often notice in sales people, were not necessarily the best performing and more loyal employees in the company. Employees with no previous telemarketing experience, who have a creative profile, reside within less than 5 kms from their workplace and are members of a few online social networks appeared to be more performant.
Our HR Manager extracted results of the targeted sales people, objectively evaluated each of their profile and took them through different aptitude, motivation and personality tests to finally realise that the recruitment criterions that were initially used did not take into account key employment principles that are sometimes considered as being secondary. He mainly found out that in this particular profession, employees become quickly saturated if they had a previous sales experience in a stressful environment. Also most of employees who live in a suburb area and have to go through huge traffic twice a day to reach their workplace in Central Nairobi, resign after 4 or 6 months.
From predictive recruitment tools to recruiting robots
Let’s start with profile aggregators, for example. They are powerful search tools which trace and aggregate every single web data of targeted people resulting from multiple sources, including social networks, which are then combined in one unique profile.
They help conduct profiles’ research and matching and sometimes integrate a third characteristic such as filtering which, with the use of prediction tools, helps trace all recent online profile updates or social network activities of a given person, as signals that he/she is looking to change employment.
These are just examples of some of the tools used nowadays and that show how old and new recruitment methods are different in terms of efficiency.
Understandably, recruiters fear that this robotic takes their place for the initial recruitment steps but also for the whole recruitment…and selection process.
Methods of e-recruitment that scan thousands of resumes when a recruiter looks for an employment site or a social network on a search engine such as google, are now outdated.
Algorithms go even further now. Thanks to their artificial intelligence, they are supposedly able to predict the behaviour of a future candidate and his/her capacity to fit in the company even before he/she is hired. According to researchers from Minnesota and Toronto universities who were published in the Harvard Business Review, these algorithms could even be more reliable than humans in the search for the ideal candidate! Algorithms are not perturbed by unnecessary details because they are constant and have no feelings.
In addition to this, the future is already here: an artificial intelligence could be installed in a robot which will then be in charge of recruiting, as it was experimented in Australia. Research conducted by Roland Berger shows that three millions of jobs could disappear in France by 2025 because of the automation of the economy; are we then, as human recruiters, about to disappear as well?
Pressure on recruitment costs and the necessity of saving time on observation and analysis play an important role in this trend.
As recruiters, the question we now need to ask is if, generally speaking, we are really benefitting from such approaches if we aggregate collateral risks without seriously thinking about it.
We must be wary of sorcerers’ apprentices!
An analysis of the environment allows us to have a balanced approach when studying this substantive phenomenon.
First, most of HR departments and recruitment specialists work in environments where using big data for proactive HR processes and to identify future staff needs would be like killing a fly with a barrel; especially that job and competency planning (GPEC) data and procedures have not been totally exploited.
One HR consultant we had an interview with when preparing this article was being ironic when he said: “It would already be a good start if my clients were using the small data and all the information they had sitting in their dashboards!”
Secondly, we can only appreciate the benefits of big data if we accept to have our personal data traded. However ethical questions would still need to be asked, even if we think that we have no more choice but using big data. This would be enhanced by the abyssal voids of the African legislation. In northern countries, laws like “Computer science and freedom” in France, that give you the possibility to request from any organisation or enterprise the data they hold on you as well as to ask for their correction or suppression, are still not easy to apply. Imagine how much it would also be difficult to apply them in an African context!
Employers already use information like the city of residence, meal times, hours during which candidates make postings on social networks to determine their age, bedtime and more generally “desirability” indicators.
In a different context, with maybe more serious consequences: one of the main short term insurance companies in South Africa is a pioneer in using predictive analysis based on the exploitation of mass data to rationalize the treatment of customers’ reclamations but most importantly to detect fraudulent ones.
Now, imagine that this company, with less effort, has access to specific customers’ information, on the internet, that reveal that some of their clients ordered over a 12-month period a significant amount of alcohol online, or that they conducted web search on liver or heart diseases. These customers would automatically and arbitrarily be classified in the high risk category, would have their insurance policy price increased or maybe even be considered as uninsurable!
In addition to not having a legal framework protecting the circulation and use of data, the challenge in using big data, especially in the recruitment field, may come from a lack of analysis of data generated through this practice. Systems that are specialised in the use of predictive tools often read collected data as they appear without taking the time to analyse them.
However we could miss essential information or give more importance to some interrelationships than necessary if we blindly trust algorithms’ readings. Humans with a good knowledge of the subject are still much needed to make sure we use the right statistical model.
The role big data will play in Human Resources beyond the recruitment stage cannot be denied. According to Sean Mclean, Director of relations with universities at IBM Africa: “Big data have become a serious business – but we’re lacking the necessary management, analysis and transformation competencies. Africa is in a good position to play a key role in the big data and analysis sector, thanks to its emerging technological markets and the rapid growth of innovation.” For this expert, African decision makers in the education sector should act now to ensure the appearance of a new generation of talents which will help get the most out of big data and make the continent a reservoir of human resources in that field.
This optimism is explained by the fact that the need in using big data has created more than 4.4 millions of jobs worldwide in 2015 and that only one third of this need in qualified human resources will be met.
Robots have not won yet!Africa, Africa Recruitment, Employment, Future trends, Social network