Analytics can help HR teams make informed decisions by relying on data rather than best guesses or gut feelings.
HR teams use basic metrics to inform decisions such as hiring and terminations. Some metrics that HR commonly references include employee head count, salary increases, performance reviews and employee survey data. Each of these metrics typically relies on data from one source.
Some decisions might require data from multiple sources to determine the best course forward. For example, when determining which employees to promote, leaders might look at data in multiple systems, such as compensation, performance reviews and information about peers in similar roles.
Company leaders might also use current data to attempt to predict future opportunities or concerns. For example, a company might use demographic data, turnover data and performance data to predict who is most likely to leave the company in the next six to 12 months if no action is taken.
HR teams can use these types of data to inform programs and policies within their organizations. Analytics can help show the benefits of taking a certain action based on concrete data rather than people's opinions or guesses.
Using HR analytics is not as simple as turning on a switch, however. The HR team needs staff who understand analytics and a go-to expert to answer questions. HR leaders should also know what the data is, where it's stored and how best to use it to help people get the information they need. Without these items, analytics can be challenging and lead to mistakes.
Organizations can get several benefits from implementing HR analytics.
9 challenges to look out for
HR professionals should keep in mind the following nine challenges of using analytics in HR.
1. Lack of analytics expertise
HR staff might not possess the required knowledge about analytics, which could lead to issues such as losing credibility with the business when presenting incorrect or problematic data. For example, the data used as input for the analysis might contain errors, the calculations used in the analysis might be inaccurate, or a misunderstanding of the process might lead to an inaccurate interpretation of the results.
There are ways for the HR team to avoid these issues. Organizations can hire someone with data analytics experience, provide analytics training for existing HR staff or require that employees who are knowledgeable in analytics review any data before HR publishes results.
2. Legal concerns
It's important when using data to make sure that confidential information is not widely shared. For example, HR should not use an employee's Social Security number as a unique identifier. Also, the organization should not use the results of the data analysis to discriminate against a protected group.
The best way to avoid running into legal issues is knowing what data in the system is considered confidential based on the countries in which a company operates, and then restricting access to that data. HR team members can look up legislation online or work with their legal team to identify such data. Training can also help mitigate problems by making sure that everyone can identify confidential data.
3. Statistically inconclusive results
Without statistical knowledge, HR teams might miss indicators showing that the results are unreliable and use the analytics anyway. When this happens, organizations could take actions or build plans assuming the results are trustworthy. To avoid publishing data with inconclusive results, the HR team can consult with data experts within their company or get help from an external contractor. Training is also a valuable option; however, this might take time for someone to become comfortable with statistics.
4. Insufficient resources
Developing HR analytics takes time and resources. If a team is short on staff, finding the time to explore analytics and developing the skill set to truly take advantage of the results might be unrealistic.
5. Data in multiple systems
If the HR team uses multiple systems to manage HR processes, they might need to integrate the data to get any valuable analytics out of it. For example, a company might use an HR information system (HRIS) for common employee data and use different systems for recruiting, performance management and compensation planning.
In addition to HR applications, there might be a need to incorporate data stored in systems that IT, finance or other departments manage. Organizations might be able to merge subsets of data from other systems into the HRIS, or they can adopt an analytics tool to pull data from each system.
Many HR teams use spreadsheets to merge data from multiple systems. Applications such as Tableau or Microsoft Power BI, however, offer the ability to import data from multiple systems. Then, those tools can produce charts and graphs based on data from multiple systems.
6. Lack of direction
HR leaders, such as the vice president of HR, or other company leaders often ask for analytics without clearly articulating what they need or want. They might only know if it's what they require after they see it, which can lead to a lot of back and forth. This can be especially challenging when the person developing the analytics isn't trained or doesn't have a good understanding of the data that's available.
For example, say a CFO requests performance data for the finance department's employees. Initially, they might only ask for an employee's current performance review data. However, they might then realize that they need more information about the worker and ask for other data, such as the employee's attendance and compensation information.
Moving past basic HR analytics might take an investment in terms of technology and resources.
7. Stakeholder buy-in
Moving past basic HR analytics might take an investment in terms of technology and resources. Without support from senior management, it can be difficult to take advantage of more advanced analytics, such as predictive analytics, or to produce basic analytics without significant manual effort.
When trying to gain support, it's helpful to build a business case that identifies the investment required and the current challenges. Highlighting requests from the business that have been received, limitations with current systems, the time and effort required to respond to requests, and an estimate of the required investment gives senior management the data that they need to make an informed decision.
Stakeholders also might want HR to be able to measure the direct impact that analytics have on the business. This can be difficult to measure because there are multiple factors that influence a given outcome. For example, incorporating more analysis into the performance review system might or might not change the final rating a manager was going to give an employee. It's difficult to quantify how the additional data might have influenced the manager's decision.
8. Resistance to change
HR and business leaders within the company might be uncomfortable moving from their current state to data-driven decision-making. Using analytics could represent an unwelcome shift in how they work and force them to learn new skills. For example, leaders might receive data that they don't understand or trust, even if it is valid.
To reduce the resistance, it can be helpful to train leaders on how to use the data. Also, including commentary with the analytics can help explain what the numbers are showing rather than expecting those people receiving the data to interpret the results. Additionally, HR should identify champions within the company who support the initiative and ask them to express the importance of using data to leaders throughout the company.
9. Incorrect AI results
There is a tremendous push to use AI within companies and HR teams. AI can sort through data quickly to find patterns, evaluate text entered in an HR system, and recommend charts and graphs to display results.
There are many benefits to using AI in HR, but it's important that the HR data analyst review the results to validate that the information is accurate and meets the business requirements. This is especially important for cases where AI is analyzing data, such as performing sentiment analysis. Negative-sounding words, for example, can have a positive meaning depending on how they are used in a sentence. The sentence "This project could not have been any better" might appear to be a negative sentiment to an AI tool, but when a human reads it, it's clear that it's positive.
Over time, AI applications will improve and likely eliminate these concerns, but in the short term, it's important to validate any results that AI tools produce.
Eric St-Jean is an independent consultant with a particular focus on HR technology, project management, and Microsoft Excel training and automation. He writes about numerous business and technology areas.
