Half of All Companies Do Not Trust Their TA Data Sources

As companies look to the future of talent acquisition, they must re-examine talent analytics. Organizations need a more robust approach that harnesses the power of provable, scientific data. Yet according to a new study by Aptitude Research, only 1 in 3 companies are using performance data in their recruitment decisions, and 1 in 2 companies do not trust their data sources.

TA managers are familiar with collecting data, tracking metrics and key performance indicators (KPIs), but they often fail to turn that data into actionable insights. Analytics is the practice of using metrics to make better decisions. If metrics help answer the “what?” then analytics answers the “so what?”

To shift from a “what” to a “so what” TA strategy, companies must better understand the data they are using, the quality of that data, and how that data is sourced.

The new study, Redefining Success: Talent Analytics for the Future, looks at how companies are rethinking their approach to analytics. Here are some of the top findings from the report:

Companies are not satisfied with the quality of their data. Over the past few years, many TA teams have struggled to manage disparate systems and an influx of data. The primary challenge that they face is not necessarily the quantity of data but rather the accuracy and consistency of that data. We found that less than one-third of companies are very satisfied with their data’s accuracy, quality, and integrity.

Companies are not always starting with the right data. They often start with resume or social-profile information, ignoring certain candidates and including biases. In fact, 55% of companies rely solely on resume data to make talent decisions. This information is not necessarily an indicator of performance or quality of hire. Consequently, by relying solely on the resume to make hiring decisions, employers can erode candidate trust and confidence in the hiring process.

Ethical AI plays a role in analytics. Companies must consider ethical AI when they evaluate providers today. Solutions should be transparent and backed by explanations, describe their methodology, and frequently publish their data. Considering the role of ethics in AI builds confidence with employers and candidates who want to understand how their data is being used. Fifty-two percent of companies stated that ethical AI would mean that data must be transparent and 48% stated that AI would include explanations.

Transparency leads to trust. Data transparency is the ability to easily access and work with data regardless of location and having confidence that the data is accurate, all of which is critical because it creates trust. If a vendor cannot explain its AI, its methods may not be legally defensible. The right vendor should explain how its AI works and validate that its tool is not replicating human biases or otherwise having an adverse impact on protected groups. We found that when data is transparent, it increases the trust in TA leaders, hiring managers, and executives.

Collect unbiased assessment data to inform decisions better. Assessments provide the science and insights to inform decisions better. Companies that leverage assessment providers are three times more likely to improve quality of hire. The data used to drive decisions must incorporate performance and post-hire data. Organizations looking to define quality of hire must look at their existing employees. This data (which includes first-year retention, performance, and productivity) should help inform TA processes.

Talent acquisition has become more complex over the past few years. Companies face new pressures with attracting and engaging talent, and they need quality data to make the best decisions. Talent analytics requires a deeper understanding of what drives success for both employer and candidates. Employers need to consider a framework that will turn metrics into insights and actions. Ultimately, they must be able to transform their approach to analytics by thinking more strategically about the quality of their data.