COMPANY CONFIDENTIAL. DO NOT DISCLOSE.
HR by the numbers
The PNC Financial Services Group brought automation and analytics to Human Resources. JMP® is helping them evaluate internal success, increase equity, improve performance and shrink employee turnover.
_____________________________________________________________________________________
Challenge: In an effort to reduce subjectivity in personnel decisions, PNC is taking an untraditional approach to Human Resources (HR): embracing analytics and automation. When early benchmarking suggested the true potential of HR analytics, PNC’s leadership sought a more sustainable, cost-efficient solution to take the place of the third-party consulting firms to which they had previously outsourced analytics. But bringing analysis in-house required a tool more powerful than Excel.
Solution: JMP® statistical discovery software offers wide-ranging capabilities with a complete analytics workflow that helps HR analysts at PNC process and derive insights from internal labor market data. In addition to the software’s data management solutions, its interactive data visualization platform Graph Builder has been especially helpful for finding patterns in the data and communicating findings that inform strategic personnel decisions. With JMP, HR analysts are ensuring that pay is based on performance and risk management ratings, not confounding factors such as gender, race or age.
Results: PNC’s analytical approach to HR has helped the company to improve performance metrics, increase equity and reduce employee turnover. And, says Jay Wilkinson, PNC Director of HR Metrics and Analytics and Senior Vice President, JMP has been a key factor in these successes – in addition to reducing the team’s outsourcing budget: “Consulting firms were charging us $250,000 to do the type of analysis that we can now do better, or equally as well, with JMP. We brought [the work] in-house because we knew JMP could do it – and it doesn't take a lot of time.”
_____________________________________________________________________________________
Human Resources (HR) and analytics may not be a traditional pairing, but as technology becomes more tightly integrated into all aspects of business, the two fields are merging toward a productive if unexpected marriage. “We live in a world with a lot of complexity, and HR is no exception,” says Jay Wilkinson, Senior Vice President and HR Metrics and Analytics Director at the PNC Financial Services Group. While HR departments have long collected data on employee performance and incentive structures, he explains, they have largely failed to see this data for the asset it is. But PNC is taking an alternative approach.
“We’re working in the language of our core business which, as a bank, is financial metrics – data,” Wilkinson says. “As a result, we are helping to deliver phenomenal talent success that includes growing representation.” Looking to the data to surface insights rather than relying on subjective performance indicators, he says, is making the entire HR enterprise more objective and ultimately boosting equity and job satisfaction across the board. Data-driven approaches to personnel decision making now help determine employee base salary, rectify historical discrepancies in pay, reward exceptional performance and identify, understand and even predict personnel retention issues.
The HR division’s efforts to embrace analytics have not gone unnoticed, even at a financial services behemoth like PNC, now one of the largest banks in the United States. Statistical analysis, Wilkinson says, affords his team credibility when reporting to the company’s executive leadership using facts and evidence to have talent-based discussions. That confidence, and the quantifiable successes the team has introduced, translate into accelerated HR growth.
Exploratory analysis in Graph Builder helps identify and resolve data accuracy issues
While automation and analytics in HR are not altogether unprecedented, most of the best examples to date come from academic studies of internal labor markets. So when Wilkinson first embarked on his crusade to infuse PNC’s HR division with analytical thinking, he had, at best, few prior models to learn from. With the expectation that his team would be looking at categorical, continuous and text data, Wilkinson says, the group needed a platform that was both more statistically adept and more visual than the tools on hand. So they turned to JMP® statistical discovery software.
“JMP supports our analysis of all those types of data – from performance review text to diversity emails to salary data,” he explains. “And what's also nice is that in JMP, we can link HR data to business data like sales. JMP is just a really great tool, and it’s easy to use as well.”
Designed to contain an entire end-to-end analytics workflow – within a code-free or code-light platform – JMP resolves a range of data challenges from access and preparation to analysis and reporting. That variety is critical in an HR setting where, Wilkinson says, managing messy data sets is perhaps the biggest hurdle.
When HR data is added or modified by individual contributors, analysts may not be able to disentangle meaningful changes from unintentional data entry errors. Wilkinson’s team must, therefore, investigate any variation or bias before beginning robust analysis. With its interactive visualization platform, Graph Builder, JMP helps break down the complexity of data issues through data exploration and outlier analysis.
“If you consider all the data mining steps, the first one is always to grab your data and think of it visually,” Wilkinson explains. “Make sure that you understand the distributions. Are there outliers? What does the actual data look like? Don't ever, ever skip that step.” When outliers surface, he adds, they can be investigated and corrected before they skew results.
Once the data has been cleaned and prepared for analysis, Graph Builder enables Wilkinson and his team to experiment and explore, generating ever more nuanced insights into key HR issues. For example, he explains, the team can ask not only whether there is a turnover problem, but which employee demographic is turning over, what kind of turnover they’re observing and how turnover trends have changed in the last weeks or months. The drag-and-drop simplicity of Graph Builder encourages creativity and looking at a question from different angles. “The nice thing about JMP is that we can use pictures, graphs and tables to break down the data and tell a story,” Wilkinson says.
Text Explorer unlocks key business insights from previously underutilized performance review text data
HR divisions collect incredible amounts of unstructured data throughout the year in the form of performance reviews. At many companies, this information has historically gone unused, as it is far too time-consuming and labor-intensive to exploit meaningful business value from text data at scale. Additionally, an unguided manual approach could introduce bias to performance data, which would render any insight useless. But JMP changes that calculation.
The software’s Text Explorer platform, for example, creates frequency tables and word clouds that help evaluate unstructured data – an ideal tool for analyzing free response text from performance reviews. Wilkinson explains how his team has begun using Text Explorer to extract insights in a more reproducible, statistically sound way. And the value is clear: “When we help improve productivity or reduce turnover – those are significant multiples on our investments,” he says. “We use JMP on productivity and retention analysis, and we’ve started to also link it into some of the social network analysis we do. Those all have very nice ROI for the firm.”
After working closely with JMP technical staff to identify and validate new ways to apply text analytics in the HR setting, Wilkinson says he plans to expand the company’s use of unstructured data in the future. To accelerate the impact, he recently made the decision to upgrade to JMP Pro for its more robust text analytics capabilities.
A data-driven strategy aligns incentive plans with performance and risk management ratings
But text mining is just the tip of the iceberg. Wilkinson says analyzing continuous salary and incentive data helps to define a strategic approach to one of the company’s most complex HR challenges: determining employee compensation and ensuring that salaries are distributed equitably. Discretionary annual incentive plans are designed to increase productivity and, he explains, should in theory be aligned with performance and risk management ratings, though the relationship between the two has historically been challenging to measure in a quantitative way.
Once Wilkinson and his colleagues began using JMP, however, they were able to visualize salary distributions for each performance level to ensure that the means were different. They also segmented data by a variety of performance and demographic variables, like job title and career level, to ensure the same relationship held true in each category. Overwhelmingly, he says, the team found that higher performance levels did in fact lead to higher pay – an indicator that can be monitored to validate fairness in HR practices over time.
These outputs are also critical for the documentation the department is required to provide to regulatory authorities like the Equal Employment Opportunity Commission (EEOC), which mandates data review and validation of equitable pay structures. Wilkinson explains how, with JMP outputs, his team was able to easily report on PNC’s compliance to regulators like the EEOC, which places a high value on sound statistical approaches. “It's a very evidence-based approach to have a concise deliverable with a really good map and visceral visuals to show the regulators [the pay/performance] relationship,” he adds.
With JMP, PNC has eliminated the middleman and reduced costs by bringing analytics in-house
In addition to bringing more intangible value – like employee retention and more representation – to PNC, Wilkinson says that providing JMP to his HR team has also helped to reduce the division’s spend on external contracts. Whereas previously the HR division outsourced much of its analytics to external consulting firms, he explains, JMP makes it possible to instead bring analytics in-house. This decision has paid dividends, cutting costs and strengthening HR’s focus on data-driven solutions.
“Consulting firms were charging us $250,000 to do the type of analysis that we can now do equally as well, or better, with JMP,” he says. “We brought [the work] in-house because we knew JMP could do it – and it doesn't take a lot of time.”
The software’s user-friendly interface, coupled with widely available learning resources, make it an easy alternative for a team that lacked prior statistical expertise. Even users who don’t use JMP on a regular basis benefit from the dynamic visualizations it affords. Wilkinson and his team use this visual story to present results to stakeholders. “[Executives] aren’t as interested in getting into the numbers – they want a picture,” Wilkinson says, adding that PNC benefits from a tool that turns raw data into digestible information. JMP levels the playing field for everyone in the room to make data-driven decisions together, and that speed and visibility can be very powerful for bridging the gap between data analysts and management.
“At PNC, our HR group is all about working in service to grow talent and help our businesses achieve their business outcomes,” Wilkinson concludes. “Whether it's about talent, diversity, recruiting – all of that is possible with evidence and facts. That’s what we're trying to do at PNC.”
