There is more data stored electronically now than ever before. This includes financial data, sales transactions, emails, instant messages, and employee activities, to name a few. Real value can be derived from analysis of this data, enabling companies to identify correlations, to predict trends, and to eventually grow the bottom line.
What can analytics do for HR?
In the field of human resources management, data-driven HR has also started to generate a lot of interest. As the use of analytics is becoming mainstream, more companies have started using analytics solutions to identify and develop talent in different areas of the business.
The application of data analysis to people's career is part of what is commonly known as "people analytics". It entails the creation of a large number of box scores of employee performance which are used to assess, for example, a job applicant's psyche and intellect or an individual's potential as a leader or innovator. Each model, based on unique algorithms, maybe tailored to a specific company, department or job type.
It has been said that in 10 years' time every company will use people analytics solutions to manage its workforce, and in so doing, company practices will change by the day, based on insights from using these analytics. If this prediction turns out to be true, companies currently neglecting people analytics as a source of people management will start to lag behind.
The following are some examples of how analytics solutions are being used in managing human resources.
- Recruitment: Analytics solutions can be used to help identify talent and improve efficiency in the recruitment process
- Talent development: Analytics has also helped some companies identify leadership characteristics in order to determine the most effective people to manage the workforce, and help maintain a productive work environment
- Retention and turnover: Analytics can also be used to help predict employees that might become a retention problem, allowing management to act before it is too late
- Fraud detection: Algorithms are also used to detect internal fraud and identify evidence of employee's misconduct, and the findings are then used to make people-related decisions.
What are the benefits?
Research has shown that the intuitive way in which recruiters and managers determine the potential of a job applicant and employee potential is rife with unconscious biases and blind spots.
The basic premise of the 'people analytics approach' is that accurate people decisions are the most important and impactful decisions that a company can make. In other words, a business cannot produce superior business results unless its managers are making accurate people decisions. Decisions can only be accurate if they are supported by data.
What are the tips?
Whilst analytics solutions can certainly arm decision-makers with better information and analysis, they cannot and should not be used to remove the human dimension in the decision-making process.
Here are the top tips when using analytics solutions to make people-related decisions.
1. Understand the business needs
- Identify the business needs of the company
- Consider the availability of existing data relating to those business needs, and how the data might be integrated, and
- Consider the capabilities of the HR, IT and Finance team in assisting with the process.
2. Information governance
Establish a set of multi-disciplinary governance policy, procedures and controls to manage the collection, retention, use, security and access of data. The governance policy should:
- Identify the broad nature and types of data that may be collected in line with the company’s ethics and branding, and
- Identify the legal and regulatory issues that may arise in each jurisdiction where the analytics project will be rolled out, including legal obligations relating to data use and collection, cross-border transfer of data, incident reporting, collection of sensitive data (eg health information, identity card information, remuneration and bank account details), and the right to access and retrieve data.
3. Personal data collection
- Identify the methods of data collection, data use, and other data protection issues
- Carry out a privacy impact assessment if the company is to collect additional data from staff monitoring
- Establish privacy policies and guidelines to deal with staff monitoring and data collection
- Check that the company does not collect excessive data, and
- Strike an appropriate balance in collecting the necessary data to carry out the data analysis without causing mistrust within the workforce.
4. Choose your vendor and consultant carefully
When choosing a vendor or consultant to help the company gather the data and perform the data analysis, consider whether the vendor or consultant is able to:
- Demonstrate compliance with relevant local anti-discrimination laws in using historical data and metrics, building profiles and making recommended actions without putting the company at risk of unlawful discrimination
- Demonstrate compliance with local privacy laws with robust data protection policies, guidelines and training for its staff on data protection, and
- Manage the risks of algorithm bias.
Agree with the vendor on issues such as:
- Ownership of the data, the insight and the results
- The IP rights of the algorithm
- Treatment of the raw data
- Anonymizing the data, and risks of de-anonymization.
Engage external lawyers to vet and audit the process.
5. Workforce planning
If analytics is used for recruitment or promotion:
- Determine the types of personality traits and strengths required in the team
- Check whether the team requires the same skill-set or different skill-sets, and
- Check whether the use of analytic solutions would create a homogenous workforce and hamper creativity.
6. Employee relations
Once the company has determined the scope of the analytics project:
- Establish an internal communication strategy with stakeholders and employees to ensure the company is sufficiently transparent with its approach in using analytics and with the types of data it proposes to collect from its workforce, and
- Seek buy-in from the employees.
7. Unlawful discrimination
- Identify bias in historical data to minimise risks of bias in algorithm
- Use appropriate metrics and historical data that would not lead to systemic bias against certain categories of individuals, and
- Apply intervention action that would not result in unlawful discrimination.
People decisions should involve people when making them. Whilst analytics solutions can certainly arm decision-makers with better information and analysis, they cannot and should not be used to remove the human dimension in the decision-making process. Human experience is key to putting matters into context and understanding the implications to the business.